![710 zyxwvutsrqpon
Chapter 9 z
Figure 9.1 Schematic of riser concepts [Courtesy of Clausen and D’Souza, Subsea’llKBR zy
(ZOOl)]](https://image.slidesharecdn.com/handbookofoffshoreengineeringvolume-231227165401-1b6778ee/85/HANDBOOK_OF_OFFSHORE_ENGINEERING_Volume-pdf-63-320.jpg)
![Drilling and Production zyxwvutsrqp
Risers zyxwvutsrq
71 z
1 zy
Figure 9.2 Vertical tensioned drilling riser [Note: balljoint (or flex joint) is also located just below
drill floor]
of 1853 m in Brazil for the Roncador Seillean FPSO. Drilling risers are in use in greater
than 3000 m depth.
A top tensioned riser is a long slender vertical cylindrical pipe placed at or near the sea
surface and extending to the ocean floor (see fig. 9.2). These risers are, sometimes, referred
to as “rigid risers” or “direct vertical access” risers.
The development of different types of riser with the riser size (diameter in inches) and
water depth up to 2000 m is shown in fig. 9.3. The envelopes for the different riser types
are given in the figure. The installed SCRs for the floaters are identified in the figure.
The technical challenges and the associated costs of the riser system increase significantly
with water depths [Clausen and D’Souza, 20011. The cost of a riser system for a deepwater
drilling and production platform compares with that of the hull and mooring system.
The risers connecting a floating vessel and the seafloor are used to drill or produce
individual wells located beneath the floating vessel or for import and export of well stream
products. They are connected to a subsea wellhead, which in turn is attached to the
supporting sub-mudline casing. The drilling riser is attached via an external tieback
connector, while the production riser can be attached via either an external or internal
tieback connector [Finn, 19991.The first joint of the riser above the tieback connector is a](https://image.slidesharecdn.com/handbookofoffshoreengineeringvolume-231227165401-1b6778ee/85/HANDBOOK_OF_OFFSHORE_ENGINEERING_Volume-pdf-64-320.jpg)
![194 zyxwvutsrqp
100-yr.winter storm 1.2
and shut-in with PNO
surface tree and
completion tubulars PNO 100-yr. loop current 1.2 zy
Chapter z
9 z
;:Pi l[ormal production
P-N5
supported from
the top of riser
P-N4 zyxwvuts
Table 9.14 Design load cases, Matterhorn TLP [Jordan, et a1 (2004)]
PNO, PSI zyxw
1 1-yr. winter storm 11
PNO Cold core eddy 1.2
PSI 100-yr.hurricanc 1.2
PSI 1000-yr.hurricane 1.5
description
P-K3
p - ~ 4
P-L1
p-L2(2)
p-L3‘3’
p-~4(4)
Design environment
contents(’)
increase factor
completion tubulars
supported from the
top of the riser.
Shut-in with the
surface tree and
completion tubulars
supported at the
top of the riser
PK
PK
100-yr. hurricane 1.2
1000-yr. hurricane 1.5
IPK Il-yr. winter storm 11 I
IPK 1100-yr.loop current 1 1.2 I
~ PSLNO 195% non-exceedance 1 1.2 I
1PSL 1 l-yr. winter storm 11.2 I
1PSL 1 100-yr. loop current I1.5 I
IPSL 1100-vr. hurricane 11.5 I
,Veilkilled with 1
; ~ 10-yr. winter storm ~ lli
1
p-C3 supported from the 10-yr. hurricane
p-c4 sap of the riser PK 100-yr. hurricane 1.5
BOP stack and
completion tubulars
10-yr. loop current
1P-TD 1Tensioner damage(4) 1PSI 1 100-yr. hurricane I1.5 I
Notes:
The follouing arc definitions of the “Pressure & Contents“ column abbreviations:
PNO
PSI
PK
PSL
PSLNO Normal operating surface shut-in tubing pressure with a tubing leak. To overcome this situation and
replace the leaking tubing. a”Bu1lhead Pressure” must be imposed at the surface wellhead that is greater than the
shut-in tubing. “Bullhead Pressures” will not be present during a hurricane when the platform is abandoned
Add a 15 kip snubbing unit BOP, a 41.5 kips snubbing unit, and 25 kips of work string to the top of tree
Operator can remove everything but the 15 kip snubbing unit BOP off the tree before a hurricane or a severe loop
current situation
Tensioner damage is defined as the loss of one tensioner element without any adjustment to the remaining
elements
Normal surface operating pressure under normal flowing conditions
Shut-in tubing pressure with no tubing leak
Well killed with heavy liquid in the tubing and the annulus
Surface pressure shut-in tubing pressure with a tubing leak](https://image.slidesharecdn.com/handbookofoffshoreengineeringvolume-231227165401-1b6778ee/85/HANDBOOK_OF_OFFSHORE_ENGINEERING_Volume-pdf-147-320.jpg)
![Drilling and Production zyxwvutsrqp
Risers I91 z
The analysis may be frequency domain or time domain, coupled or uncoupled. Often,
global vessel motions will be computed by one group and riser dynamics by another. This
paradigm is very risky. For example, vessel motion programs often use different coordinate
systems than riser programs. Translating motions from the origin of a vessel to the riser
hangoff point using a different coordinate system can and often does lead to errors. It is
best to perform a quality check of the procedure by analysing a simple case, e.g. a
monochromatic sine wave, and performing some hand calculations prior to doing the bulk
of the analysis with random wave input.
Another issue with frequency domain analysis is that it can neglect important non-linear
effects such as slowly varying motions and damping. Coupled analysis is the simultaneous
solution of the vessel and riser motions. In deep water the riser loads may actually reduce
the vessel responses [see, e.g. Prislin, et a1 (1999) Halkyard, et a1 (2004) for comparison of
full-scale data with calculations].
The above discussion focuses on strength design. The riser may also fail from fatigue,
hydrostatic collapse, buckling and thermal effects.
Fatigue analysis needs to consider fatigue for vessel motions, wave loadings on the riser
and vortex induced motions. The analysis is similar to that for the drilling riser discussed
above; however, the difference is that production risers are designed to remain in place for
the life of the field, whereas drilling risers are routinely retrieved and inspected. API
RP2RD requires the fatigue life of the riser to be:
Three times the service life (usually the life of the field) for areas accessible for
inspection (or, where safety and pollution risk are low), or,
Ten times the service life for areas not accessible for inspection (or, where safety or
pollution risk are high).
In practice, deep water risers are invariably designed for ten times the service life.
Figure 9.50 shows the DNV fatigue curves used for the analysis of riser components. These
are identical to the DOE curves discussed in Chapter 7 [DNV C N 30.2, HSE, 19951. Riser
connectors without welds use the “B” curve with an appropriate stress concentration
factor. Preliminary fatigue analysis is often performed using the “B” curve to determine an
allowable SCF to achieve the required design life. If this allowable SCF is five or greater, it
is likely that threaded couplings may be used for the connectors. New couplings with lower
SCFs are becoming available as was mentioned above.
The difference between the C, D and E curves, which apply to welded connections, depends
on the quality of the weld. Special considerations need to be given to the presence of H2S
and corrosion. These fatigue curves assume cathodic protection.
The reader is encouraged to review the literature on riser sizing and analysis for deepwater
floaters [e.g. Wald, et a1 2002; 0’Sullivan, et a1 2002; Jordan, et a1 2004; and Bates, et a1
20021,](https://image.slidesharecdn.com/handbookofoffshoreengineeringvolume-231227165401-1b6778ee/85/HANDBOOK_OF_OFFSHORE_ENGINEERING_Volume-pdf-150-320.jpg)
![Drilling and Production Risers zyxwvutsrq
Riser Linepipe
Det Norske Veritas, DNV 1981 0.5* 0.72*
821 z
Table 9.22 Usage factors for internal pressure zyxw
1Design code IUsage factor zyx
(nh) I
lANSI/ASME 31.4 & 49 CFR195 10.6 10.72 I
/ANSI/ASME 31.8 & 49 CFR192 Ios# 10.72# 1
1British Standard 8010, Part 3 10.6 10.72 I
* DNV (1981) specifies the riser (zone zyxwvut
2) as the part of the pipeline less than 500 m from
any platform or building and the pipeline (zone 1) as the part of the pipeline greater than
500 m from any platform building
=ANSI ASME 31.8 specifies the riser zone as the part of the pipeline. which is less than 5
pipe outside diameters from the platform and the pipeline zone as the part of the pipeline.
which is more than 5 pipe diameters from the platform
can be used. It must be observed that DNV and ASME codes specifically refer to
a nominal wall thickness, while IP6 and BS 8010 (1973) refer to a minimum wall
thickness. If a minimum wall thickness is specified, the nominal wall thickness
may then be calculated using a corrosion allowance and a manufacturer’s tolerance (see
Section 9.3.4.1).
In table 9.21 t,,, = corrosion allowance, t,,, = minimum wall thickness, t,,, = nominal
wall thickness, zyxwvu
D
,
,
, = nominal outside diameter, zyxw
oL= specified minimum yield stress
(SMYS), nh = usage factor or fraction of yield stress, P,= internal design pressure, Po =
external design pressure.
The usage factor nh, which is to be applied in the hoop stress formulae, is specified by the
applicable design code and the zone or classification of the pipeline. For submarine
pipelines and risers the code requirements governing design usage factors are summarised
in table 9.22.
Temperature de-rating shall be taken into consideration for the risers and pipelines
operating at high temperatures (typical > 120’C). The pressure containment check
(bursting) should be performed according to DNV 2000 DNV-OS-F101 Section zy
5 D400
[equation (5.14)].
9.3.4.4 Resistance To External Pressure (Collapse)
Two load cases need to be considered with respect to resistance to external pressure:
In-service collapse
Collapse during installation
Failure due to external pressure or hydrostatic collapse is caused by elastic instability of the
pipe wall. For wall thickness determination the external pressure is calculated from the
Previous Page](https://image.slidesharecdn.com/handbookofoffshoreengineeringvolume-231227165401-1b6778ee/85/HANDBOOK_OF_OFFSHORE_ENGINEERING_Volume-pdf-174-320.jpg)
![Drilling and Production Risers zyxwvutsr
823 z
Table 9.23 Out of roundness formulas zyxw
1Out of roundness or initial ovalisation
1Code or standard ITolerance 1
x IUU and x IUU
D n o m
1 D n o m
I
IAPI 5~
D n o m - Dmin
~ ''Oy0 ~
x 100 and
'Dmax - D n o m
1 D n o m D n o m
1 DNV
&ax - Dmin
Dmax +Dmin
BS 8010 (1973) 2.5% 1
l l
The combined loading criterion is to be performed according to DNV 2000 DNV-OS-F101
Section 5 D500 [equations (5.22)-(5.26)]. If the riser is in compression between the supports
the global buckling shall be checked according to Section 5 D600.
9.3.4.6 Resistance to Propagation Buckling
Two load cases need to be considered with respect to resistance to internal pressure:
in-service and during installation. The required buckle propagation wall thickness is
the wall thickness below which a buckle, if initiated, will propagate along the pipeline
or riser, until a larger wall thickness or a reduced external pressure is reached. The wall
thickness required to resist buckle propagation can be calculated from the formulas in
table 9.24.
Generally, the DNV formula is marginally more conservative than the Shell Development
Corporation formula and they are both considerably more conservative than the Battelle
formula. The degree of conservatism required depends on the installation technique in
terms of risk, the length and cost of the line and the water depth in terms of how easily a
repair can be made. In practical terms: changing of any criteria will change the required
wall thickness. However, since designers are normally limited to selecting from API pipe
sizes, there is quite often no actual change in the pipe specified.
If during design, a pipeline is found to be governed by the buckle propagation criteria, then
there are two options open to the designer; the first option is to make the wall thickness of
the pipe sufficient so that a buckle once initiated will not propagate. The second option is
to make the wall thickness of the pipe sufficient to only withstand external pressure
(hydrostatic collapse) and to use buckle arresters. Buckle arresters consist of thick sections
of pipe or welded fittings, which a buckle cannot propagate through. If buckle arresters are
fitted, the damage will be limited to length of the pipeline between arresters, should a
buckle initiate.](https://image.slidesharecdn.com/handbookofoffshoreengineeringvolume-231227165401-1b6778ee/85/HANDBOOK_OF_OFFSHORE_ENGINEERING_Volume-pdf-176-320.jpg)
![824
Table 9.24 Wall thickness to resist buckle zyxw
Chapter z
9 z
Code or standard zyxwvu
1Formula 1 Remarks
DNV 1981 and DNV 1976 Conservative
tnom =-
kDnom
l + k = F
1.15~~0,
Battelle Columbus Laboratories
[Johns, et a1 19761
Shell [Langner, 19751 11.0%
BS 8010 Part 3 (1973)
Note: The propagating buckling check should be performed according to DNV 2000 DNV-OS-F101 Section 5
D500 [equation (5.27)]
This risk, however, is considerably reduced after installation. The choice between the two is
determined by considering the potential cost saving in wall thickness and possibly
installation benefits due to the reduced submerged weights. This is paid off against the risk
of having to replace a relatively large section of pipe or riser possibly in deep water.
9.3.4.7 Hydrotest Pressure
The BS 8010 (1973) design code gives criteria to calculate the minimum hydrostatic test
pressure for a pipeline. The test pressure required to qualify for a MAOP (Maximum
Allowable Operating Pressure) equal to the specified design pressure is either the lower of
150% of the internal design pressure, or the pressure that will result in a hoop stress (based
on specified minimum wall thickness) equal to 90% of the specified minimum yield stress.
The test pressure should be referenced to the Lowest Astronomical Tide (LAT) and due
allowance made for the elevation of the pressure measurement point and parts of the
system above LAT.
For definition of design pressure see DNV 2000 DNV-OS-F101 Section 1. Hydrotest
criteria are discussed in Section 5 B200.
9.3.5 SCR Maturity and Feasibility
Three views of SCR maturity and feasibility are given in figs. 9.56 and 9.57. Figure 9.56
shows the existing SCRs against diameter and water depth. The choice of diameter
and depth as axes is largely motivated by collapse considerations, although installation
capabilities are also relevant. Figure 9.3 shown earlier (similar to fig. 9.56) puts more
emphasis on water depth and diameter than on waves and vessel-types. It also shows recent](https://image.slidesharecdn.com/handbookofoffshoreengineeringvolume-231227165401-1b6778ee/85/HANDBOOK_OF_OFFSHORE_ENGINEERING_Volume-pdf-177-320.jpg)
![Drilling zyxwvutsrqpon
and Production Risers zyxwvutsr
825 z
Figure 9.56 Existing SCRs and technology stretch
Figure 9.57 Estimated feasibility [from Spolton and Trim (2000)l and maturity of SCRs [Note: feasibility
colour scheme developed for 10-in. HPHT oil production SCRs (left shading zyx
= steel, middle light
shading =steel-titanium, right = unproven)] zyxw
chronological development of flexible-pipe riser regimes, indicating that feasibility for
typical production sizes is now approaching 2000 m water depth.
In fig. 9.57 [Spolton and Trim, 20001, titanium parts were progressively substituted for steel
parts as environments and vessel motions considered became worse. Thus, in the left and](https://image.slidesharecdn.com/handbookofoffshoreengineeringvolume-231227165401-1b6778ee/85/HANDBOOK_OF_OFFSHORE_ENGINEERING_Volume-pdf-178-320.jpg)
![826 zyxwvutsrqpon
Chapter z
5
Extreme
Survival
the lower light shading parts of the chart, relatively small amounts of titanium are used; e.g.
in TSJs and TDZs, whereas the top right region represents SCRs composed mostly or
completely of titanium.
These figures are guidelines only but give a reasonable first impression of what has
been and what can be achieved, and what “technology stretch” is required for harsher
conditions.
Soil-structure interaction and VIV effects (separately and together) are major uncertainties
remaining in the SCR design and analysis, although considerable progress has been made
in both areas through the STRIDE and CARISIMA JIPs (among others). Findings
continue to be extended and assimilated into mainstream practice and software codes.
Trenching of SCRs in the TDZ has been observed and may represent an additional critical
case for careful examination during detailed design.
100 yr. Associated Associated Product Design pressure
Associated 100 yr Associated Product Design pressure
Associated Associated 100 yr. Product Design pressure
100 yr. Associated Associated Product Failed mooring
100 yr. Associated Associated Variable Accidental**
1000 yr. Associated Associated Product zyx
- zy
9.3.6 In-Service Load Combinations
The in-service design cases of table 9.25 should be assessed for the most severe loading
direction, which may vary according to the response quantity of interest. Allowable stresses
are given in table 9.26.Latest versions of the API and the DNV riser codes, such as API RP
2RD, and DNV-OS-F201 are commonly used.
Von Mises stress is defined in accordance with API RP 2RD (1998) as:
von Mises stress = [(S- h)’+(S - ~ ) ~ + ( h
- zy
Y
)
’
]
/; (9.33)
Table 9.25 In-servicedesign cases
1Design case 1Wave 1Current 1Wind IContents 1Other* I
Normal To be agreed Associated Product Design pressure
operating (typically 1 yr.)
IPressure test ~ 1 yr. 1Associated Associated 1Water 1Test pressure 1
*Use associated pressure for survival. Maximum pressures are given; i.e. lower pressures should also satisfy checks
**Accidental conditions are discussed further in Section 9.3.7](https://image.slidesharecdn.com/handbookofoffshoreengineeringvolume-231227165401-1b6778ee/85/HANDBOOK_OF_OFFSHORE_ENGINEERING_Volume-pdf-179-320.jpg)
![DriNing and Production Risers zyxwvuts
851 zy
inaccurate. Moreover, thin shells with outer to inner radii greater than 0.8 need curvature
correction factors [refer to PD6493 (1991) for guidance]. zyxw
9.8.2 Paris Law Fatigue Analysis
The so-called Paris Law for fatigue is described using
daldN zyxwvut
= zyxwvu
A(AK)” (9.44)
where da:dN = crack growth rate of crack of depth a vs. the number of applied stress
cycles N, AK = stress intensity factor range, while A and m are material specific constants.
BS7910 (1999) provides recommended values for the Paris Law, which should be suitable
for the fatigue analysis. Material specific data obtained from tests are relatively inexpensive
and may be used in-lieu of codified data.
If idealised stress intensity factor solutions are utilised (e.g. smooth plate solutions) in lieu
of the finite element fracture mechanics analysis of the actual geometry, then relevant stress
concentration factors should be applied to the stress range bins to account for increased
applied stress due to local weld geometry, pipe mismatch, etc.
9.8.3 Acceptance Criteria
The industry has typically followed an approach similar to the schematic in fig. 9.72. The
approach has been to develop curves showing an envelope of elliptical cracks (edge or
embedded), which may grow to the limiting flaw size (see above) in a specified time. The
“specified time” is usually established as a safety factor multiplied by the design life.
Deciding the safety factor is subjective, but must take into account the type of inspection
used during weld fabrication. Additionally, the safety factor should reflect uncertainties in
predicted loads (see, also, Section 9.7 on safety factors).
9.8.4 Other Factors To Consider
Some of the other factors are listed below:
Internal Contents: crack growth may be accelerated in H2S or other corrosive
conditions
Cathodic Protection: crack growth is dependent on the level of corrosion potential
protection
Hydrogen embrittlement from welding
Plastic straining (for reeled risers)
Internal pressure effects on crack growth
9.9 Reliability-Based Design
Reliability-based design is becoming more common in pipeline engineering and other
areas of the offshore industry. Its application to risers is limited at this time. Particularly,](https://image.slidesharecdn.com/handbookofoffshoreengineeringvolume-231227165401-1b6778ee/85/HANDBOOK_OF_OFFSHORE_ENGINEERING_Volume-pdf-204-320.jpg)
![Topside Facilities Layout Development zyxwvutsr
879 z
Figure 10.5 Deck installation scheme on piled jacket structures or floating barge
Figure 10.7 shows the layout of a deck installed on an FPSO. In this case the helideck and
the personnel quarters are separated from the process equipment by a sufficient distance.
Such a separation is possible because the ship-shaped structure provides ample space on its
deck. The flare is placed far away from the personnel and fuel sources.
10.6 Floatover Deck Installation
Deck installation using the floatover deck concept [Salama, et a1 1999 and Section 6.1.1 of
this handbook] in lieu of the traditional crane vessel lifting is a well-accepted method. This
method provides an attractively cost effective way to install decks, especially when the deck
weight exceeds available crane lifting capacity. The floatover method eliminates the use of](https://image.slidesharecdn.com/handbookofoffshoreengineeringvolume-231227165401-1b6778ee/85/HANDBOOK_OF_OFFSHORE_ENGINEERING_Volume-pdf-231-320.jpg)
![Chapter z
11 z
892 zyxwvutsrqp
11.2 Design Basis
The first step in offshore pipeline design is establishing a concise design basis document
(DBD). For consistency, every project requires, in its early phase, the establishment of the
DBD. This is to be used as a reference by the design team for the different aspects of
offshore pipeline design. The DBD provides basic project-dependent information, and
enables consistency and correctness of project calculations, reports, bid specifications,
contract documents, installation procedures, etc, with respect to the fundamental
parameters of the project.
For a major project, a DBD should include the following sections: zyx
Development overview: describes the project location and basic layout.
Reservoir and well information: provides reservoir characteristics, fluid rheology and
production rates.
Environmental defines geotechnical properties along the proposed route (shear
strength, weight, etc.), meta-ocean data (waves and currents), and seawater
temperatures/chemistry zyxwvu
.
Flow assurance: provides information on flowline parameters, e.g. operating pressures,
temperatures and velocities; identify and address flow hazards such as hydrates, wax,
scale, corrosion, slugging.
Wellbore, drilling and completion information: provides safety valve philosophy,
downhole chemical injection, completion design, downhole monitoring, general rig
description, well servicing and intervention process, etc.
Equipment design philosophy:describes the design and selection approach, standardisa-
tion of components and interfaces, equipment design life, quality programme.
Subsea trees and flowline]pipeline sleds: describes subsea trees, flowlineipipeline sled
characteristics, tie-in jumpers, completion/workover system, equipment marking,
corrosion protection.
Production control system and umbilicals: establishes codes and standards, system
overview, subsea instrumentation, redundancy, emergency shut-down valve (ESV)
requirements, surface equipment, subsea equipment, power and hydraulic umbilical,
methanol distribution umbilical (if required), intelligent well completions, metering.
Pipelines: provides general characteristics (grade, size, water depth), route selection,
applicable codes and regulation, system design requirements (design life, cathodic pro-
tection system, etc.), risers and tie-ins, maximum shut-in pressure, corrosion allowance.
Host facilities: gives general description of the host, process design, major equipment
list, interface definitions.
Operation and maintenance: outlines normal production parameters, start-up and
shutdown procedures, routine testing requirements, pigging, system maintenance,
abandonment philosophy.
In general, the DBD is a living document that goes through several revisions during the
course of a project. However, after the front end engineering design (FEED) phase of
a project, the DBD must define the majority of the project requirements. After FEED,](https://image.slidesharecdn.com/handbookofoffshoreengineeringvolume-231227165401-1b6778ee/85/HANDBOOK_OF_OFFSHORE_ENGINEERING_Volume-pdf-243-320.jpg)
![Design zyxwvutsrqponm
and Construction of Offshore Pipelines zyxwvuts
250 or less
897
1.000 zyxwv
Table 11.3 Temperature de-rating factor, T,for steel pipe zyx
1Temperature (“F) 1Temperature De-rating Factor, T 1
,
1350 I0.933 I
1400 10.900
I450 10.867 1
F =construction design factor of 0.72 for the submerged component and 0.60 for the
T =temperature de-rating factor obtained from Table 841.1C of ANSI B31.8,
E =longitudinal joint factor. Obtained from Table 841.1B of ANSI B31.8 (see also
According to 30 CFR 250, all pipelines should be hydrostatically tested with water at a
stabilised pressure of at least 1.25 times the maximum allowable operating pressure
(MAOP) for at least 8 h. The test pressure should not produce a stress in the pipeline in
excess of 95% of the specified minimum-yield strength of the pipeline.
The relationship between the maximum hydrotest pressure and the (nominal) wall thickness
is similar to equation (11.2), and is given by:
riser component,
see table 11.3.
Section 811.253(d)) - see table 11.4.
(11.3)
where Pm&,yd =maximum hydrostatic test pressure, and F= construction design factor,
0.95 for hydrotest.
11.5.2 Collapse Due to External Pressure
During installation, offshore pipelines are typically subjected to conditions where the
external pressure exceeds the internal pressure. The differential pressure acting on the pipe
wall due to hydrostatic head may cause collapse of the pipe. Several design codes present
formulation addressing the design against this failure mode. Amongst these codes, the most
prominent are API RP 1111 (1999) and DNV OS-F101 (2000).
The elastic collapse pressure [equation (11.5b)l and the plastic collapse pressure [equation
(11.5c)l bound the problem. Timoshenko and Gere (1961) proposed a bi-linear transition
between the two equations [see Timoshenko and Gere, 1961, figs. 7-91, which adequately
bridges the two equations. API RP 1111 (1999) adopts a transition between the two
equations [equation (11.5a)],which is simpler than the cubic interaction equation proposed
by DNV. A comparison between the API and DNV collapse pressures, P,, normalised by
the plastic collapse pressure, P,, is given in fig. 11.2. P, as calculated by equation (5.18) in
the DNV OS-F101 (2000). uses ovalisation parameterf, = 1%, as defined in DNV OS-F101](https://image.slidesharecdn.com/handbookofoffshoreengineeringvolume-231227165401-1b6778ee/85/HANDBOOK_OF_OFFSHORE_ENGINEERING_Volume-pdf-248-320.jpg)
![Design zyxwvutsrqpon
and Construction zyxwvutsrq
of Offshore Pipelines
Water depth (ft)
0 lo zyxwv
1Po(psig)
909
Ppgas
(psig) Preq (psig) i
778 5622 1
I500 1222 1729 15449 1
18000 13556 ' 0 12844 1
Table 11.7gives the required design pressure, per equation (11.23), for the significant water
depths assuming a seawater weight of 64 pcf. These values will be used to check that the
required design pressure does not exceed the maximum allowable operating pressure
(MAOP) of the pipeline system.
Based on the numbers given on table 11.7, the minimum hydrostatic test pressure equals
1.25times the maximum required design pressure, or 1.25 x 5622 = 7028 psig. This assumes
that the entire pipeline system will be subject to a single hydrostatic test. The nominal
hydrostatic test pressure is set at a slightly higher value, as follows:
The wall thickness must be selected such that equation (11.2) [given as a design inequality
by equation (11.24)] and equation (11.3) [given as a design inequality by equation (1 1.25)]
are satisfied at every point along the pipeline and riser system, as follows:
Nominal hydrostatic test pressure, Pnom.hyd =7100 psig @ mean water level (MWL)
Leading to 80% of hydrostatic test pressure = 5680 psig @ MWL
(11.24)
(11.25)
Recall that F= construction design factor =0.6 for riser component, =0.72 for submerged
component, and =0.95for hydrotest, E =longitudinaljoint factor, E = 1for API 5L seamless
line pipe, and T= temperature de-rating factor, T= 1for maximum temperature of 200'F.
Also recall that 30 CFR 250 defines that the maximum allowable operating pressure
(MAOP) to be the least of the following:
internal design pressure of the pipeline, valves, flanges and fittings,
80% of hydrostatic test pressure,
MAOP of the receiving pipeline.
The preliminary nominal wall thickness for each pipeline section is calculated to satisfy
equations (11.24) and (11.25). The API 5L (2000) wall thickness selection is shown in
table 11.8 (see also fig. 11.8). To verify that the design equations are satisfied, table 11.8
presentsthe numericalvalues ateachpertinent location alongthepipeline elevation.Note that:
For each pipeline segment the controlling (higher) required design pressure, Preqr
is at
the shallower water depth.](https://image.slidesharecdn.com/handbookofoffshoreengineeringvolume-231227165401-1b6778ee/85/HANDBOOK_OF_OFFSHORE_ENGINEERING_Volume-pdf-260-320.jpg)
![910 zyxwvutsrqpon
Water depth (ft) zyxwvuts
Chapter z
11
Wall Preq zyxwvu
Pldl Pmax.hyd2 MAOP3
thickness (in.) ,(psig) (psig) (psig) (PSk)
500 10.719 15449 15716
3000 10.719 (4581 15716
lo 10.875 1.5622 15895 110052 15680 1
8260 5680
8260 5680
I500 10.875 I5449 15895 110052 15680 1
18000 10.625 12844 14897 17180 14897 Iz
I p l d must be greater than preq
[equation (11.24)]
'P,,,.hyd must be greater than Pno,,.hyd =7100 psig in this example [equation (11.25)]
'MAOP is the least of (a)0.8 x Pnom-hyd, (b) Psdr(c) internal design pressure of valves, flanges and fittings: which in
this example will be assumed greater than the values shown and (d) MAOP of receiving pipelines.
An API wall thickness thinner than 0.625 in., which is 0.562 in., would not be adequate
since it would not satisfy equation (11.25),even though it would satisfy equation (11.24)
for water depths greater than 3670ft.
There is room to revise the wall thickness break to shallower water depth, thus making
the 0.625 in. segment longer, which would lead to project savings. To calculate the
water depth, X. at the wall thickness break, use equation (11.23) with Preq
=Pid.In this
example such an equation is: 6400 - X(64/144) - (8000 - X)(14/144) = 4897, which
leads to X = 2088 ft. Of course the prudent offshore pipeline engineer will always give
some allowance for installation tolerances, and the revised water depth for the WT
break would be set at 2200 ft. However, for purposes of the ensuing numerical
examples, the water depth of 3000 ft will be maintained.
11.7.2 Collapse Due to External Pressure
Equations (11.5a-11.5~)yield the collapse pressure. The factored collapse pressure f o . z
P,
must exceed the net external pressure everywhere along the pipeline, as shown in equation
(11.4). The collapse reduction factor in this example for seamless pipe isfo =0.7. The net
external pressure can be determined as the hydrostatic water pressure at maximum water
depth of each pipeline section assuming zero internal pressure. The net external pressures
are all within the allowable collapse pressure as summarised in table 11.9.
In this example the nominal wall thickness of each pipeline section is used. The authors
judge that for collapse as well as local buckling, it is over-conservative to deduct the
corrosion allowance in such calculations. While burst will occur at the maximum stress
which occurs at the thinnest wall thickness (justifying the corrosion allowance deduc-
tion on burst limit state check), collapse and local buckling involves the formation of
a four-hinge collapse mechanism, with maximum bending moments at four hinges
90" apart around the pipeline cross section [see Nogueira and Lanan 2001; Timoshenko
and Gere, 1961, Section 7.51. Given that the pipe mill's average wall thickness is](https://image.slidesharecdn.com/handbookofoffshoreengineeringvolume-231227165401-1b6778ee/85/HANDBOOK_OF_OFFSHORE_ENGINEERING_Volume-pdf-261-320.jpg)
![Design and Construction zyxwvuts
o
f Offshore Pipelines zyxwvuts
Water depth Wall Collapse f o zyxw
.P, (psig) External
thickness (in.) pressure, P, (psig) pressure, Po (psig)
(ft)
911
, O-500 ,0375 zyxwvu
Table 11.9 Collapse pressure vs. external pressure
10113 7079 ,222
500-3000 0.719 7911 5538 1333
Table 11.10 Factored strain vs. limiting bending strain
3000-8000
Water
depth
(ft)
0-500
0.625 6471 4530 3556
Wall 1 Factored 1 Factored
thickness installation in-place bending
(in.) bending strain strain
f l E l zyxwvuts
(%I f i E 2 (%I
1 0.022 1 3.30 1
Critical 1
[equation (11.7)]
Po - z
P
i
- bending strain
E (Yo)
0.719 ~ 0.4 ~ 0.4 1 0.168 ~ 2.22 1
0.875
0.625 1 0.4 1 0.4 1 0.550 1 0.82 1
0.4 0.4
typically 10% greater than the nominal wall thickness, around the line pipe circum-
ference, the nominal wall thickness is recommended to be used in collapse and local
buckling. Of course, this is a project decision that should be clearly stated within the
project DBD.
11.7.3 Local Buckling Due to Bending and External Pressure
Equation 11.6 gives the limiting bending strain, to avoid the local buckling limit state.
By rearranging it, equation (11.7) is obtained, which yields the maximum bending strain.
Therefore, the installation and in-place bending strains shall be limited per equations
(11.8a) and (11.8b). In this design example, factors of safety adopted arefl =fi = 2.0. The
ovality is conservatively set at 6 = 1%, which leads to g(6) = 0.833. The critical bending
strains per equation (11.7) are shown in the right hand column of table 11.10 (with P,= 0)
and are all greater than the factored bending strains.
11.7.4 Buckle Propagation
The pipeline propagation pressure value, per API 1111 (1999), is given by equations (11.21)
and (11.22). Assuming zero internal pressure the results are shown in table 11.11.
From the results presented in table 11.11, buckle arrestors are required along the 0.625 in.
WT segment, when the external water pressure is greater than 1352 psi. This corresponds to
3042 ft and greater water depth. Buckle arrestor design guidelines can be found in Park and
Kyriakides (1997) and Langner (1999).](https://image.slidesharecdn.com/handbookofoffshoreengineeringvolume-231227165401-1b6778ee/85/HANDBOOK_OF_OFFSHORE_ENGINEERING_Volume-pdf-262-320.jpg)
![Design and Construction of zyxwvutsrq
Offshore Pipelines zyxwvut
921 z
11.11 Pipeline Crossing Design
Pipeline crossing design basically involved protecting the crossed pipeline using articulated
concrete mattresses. Typically, in the US Gulf of Mexico (GOM), two 9 in. thick
articulated concrete mats are used, for a 18 in. separation between the pipelines. The single
lift equations (11.41)-(11.43) below [see Troitsky, 1982; Section 11.6.51 can be used to
calculate the crossing loads, which were relatively small, so that the crossed pipeline could
transfer such forces to the underlying seafloor. Typically, no intermediate supports for the
crossing pipeline were needed.
However, recent GOM Federal regulations require that crossing pipelines be covered
by mats from touchdown to touchdown, for water depths less than 500ft. This leads to a
crossing arrangement depicted in figs. 11.9 and 11.10, which shows with relative scale a
12 in. pipeline crossed by a 24 in. pipeline and 9 in. thick concrete mats. The capping mats
impose a load on top of the crossing pipeline, which is transferred to the crossed pipeline
as a concentrated load on a short ring of pipe (shown with a length zy
L in fig. 11.10).
Finally, the crossed pipeline transfers this load to the seafloor. Thus, crossing design needs
to be evaluated as follows.
The first step in the crossing analysis is to estimate the crossing load. Concrete mattresses'
submerged weight is approximately 6000 Ib for a 9 in. thick mat with 8 ft by 20 ft dimensions,
which leads to a submerged load of about zyxw
M
.
'zyxwvuts
=38 psf. A 4.5 in. thick mat of same dimensions
has a 3600 lb total submerged weight, for a load of about zyx
M, =23 psf. The load imposed by
the capping mattress is estimated by assuming an average drape angle of 30" (fig. 11.lo), and
the corresponding maximum linear load, q, at the crown of the crossing is given by:
4 = W[2zyxwvu
X 1.16(d zyxwvu
4
-ODTOP)
+oD~op] (11.40)
where 1.16 l/cos30, zyxwvu
d is the crossing pipe prop height, which equals distance from
adjacent seafloor (mudline) to bottom of crossing pipe, ODTop is the diameter of the
crossing pipe.
Given an existing pipeline with ODBOT=20.00in., WTBoT=0.812, embedded 3 in, two 9 in.
separation mattresses, a crossing pipeline with 0 D ~ o p
= 12.75 in., WTTop=0.750, and a
9 in. capping mattress; then: d=35 in. and q-391.2 plf. The crossing pipeline has water
filled submerged weight of 83.7plf (= 96.2plf steel weight in air, 44.2plf water contents
at 64pcf, minus 56.7plf buoyancy. When the mattress load is added, the total maximum
pipeline load is qr= 475 plf, or 5.7 times that of the crossing pipeline during hydrotest.
A pipeline on a prop with height d from the adjacent seafloor (see fig. 11.1l), with Young's
modulus E, moment of inertia I and total submerged load q, will have a distance from the
prop point to touchdown 1 given by:
72EId
4
1 =- (11.41)
The total prop force Fp and maximum bending moment at the centre of the span, Mp,
are given by:
F p = 44113 (11.42)
Mp = 4l2/6 (11.43)
Previous Page](https://image.slidesharecdn.com/handbookofoffshoreengineeringvolume-231227165401-1b6778ee/85/HANDBOOK_OF_OFFSHORE_ENGINEERING_Volume-pdf-272-320.jpg)
![14
118zyxwvutsrqpon
1Nominal nct anode weight IProject specified I130 IIbs
15
Weight required
16
Eq. 11.35: C9*C13*
X760/(C10*C14)
17
Anode specificweight
Anode thickness
Average anode diamcter
Estimated anode length
Environmcntal resistivity
Anode Area
Anode resistance
-
-~
180 lb/ft3
-~
Vendor specified
Project defined I .5 in. zyx
C1+c22 12.25 in.
1728*C19/(C2 zyxwvu
1*7c*C23*C2) 22 in.
Project specified 30 ohm-cm
n*C2I *c22 847 in.2
Equation (1 1.37): 0.1278 ohm
-~
- ~
~~ .
. .
0.315*(C24/2.54)/C250.~
IINT(CyC16) +1
1.956
408
Anode current
- --+rrent available
Minimum anode weight ] C15/c17
A
A
480
209
122.5 Ibs
19
20
.
21
22
23
24
25
126 IAnodc potential 1Project specified 1-1.05 IV
127 ICathode ootential IProiect wecificd 1-0.80 Iv
W
N
w](https://image.slidesharecdn.com/handbookofoffshoreengineeringvolume-231227165401-1b6778ee/85/HANDBOOK_OF_OFFSHORE_ENGINEERING_Volume-pdf-274-320.jpg)
![Design and Construction of zyxwvuts
Offshore zyxwvutsrqp
Pipelines zyxwvuts
921 z
M A = F 0 D / 4 (similar to a centred point load in a simply-supported beam). The total
plastic moment capacity, Mc, is the summation of the plastic capacity of the hinges at
3 o’clock and 12 o’clock, which is Mc zyxwv
=L&Syt2/4 [see Nogueira and Lanan, 2001 for
derivation details]. The effective ring length is assumed to be equal to the pipeline diameter,
L =OD; while this assumption is adequate for pipes with D/t < 25, it needs to be validated
for thinner pipes. The inequality MA5 Mc leads to the local ring collapse, or denting load,
FD, on a pipeline, as given in equation (11.44):
Fo = 1.73S,t2 (11.44)
The data of the example above leads to, for both pipes, FD-TOP = 1.73(65)(0.752)=63.3 kip
and FD-BOT= 1.73(65)(0.8122)=74.1 kip. Therefore the crossing pipeline has a factor of
safety, FS =63.3/50.4 = 1.26 and does not meet the safety criteria against local collapse.
The crossed pipeline has a factor of safety, FS =74.1/50.4= 1.47,which also does not meet
the safety criteria. Regarding the ring collapse limit state, changing the capping mats to
4.5 in. thick (which leads to a crossing load of FL=38.2 kip), would lead to a crossing
design within the safety guidelines suggested herein.
In this example, extra supports adjacent to the crossed pipeline are needed for the safety
criteria suggested herein to be achieved. Note that both pipelines must be checked, since the
free body shown in fig. 11.13 also applies to the crossing pipeline; except that the point load
would be inverted: the higher force would be applied at 6 o’clock (the crossed pipeline
reaction) and the smaller force at the 12 o’clock is due to the pipeline self-weight and
capping mats.
The above equations neglect the effects of external pressure. While this is an adequate
assumption for water depths less than 500ft, such effects can be readily addressed by using
the rational model for pipeline collapse. This way, crossing capacity in deep water can be
more precisely estimated. Nogueira and Lanan (2001) showed that external pressure has
the effect of adding to the total applied moments and also decreasing the ring collapse
hinge capacity due to additional compressive hoop stress around the pipe. Thus, advantage
can be taken of the terms given in equation (24) of Nogueira and Lanan (2001) to add the
effects of external water pressure.
The example above illustrates the fact that a new US GOM regulation, while with the goal
of decreasing the risk of pipeline being dragged or damaged by fishing gear, has the effect
of increasing substantially the stresses on pipeline crossings. Therefore, the offshore
pipeline engineer needs to be aware that what used to be a traditionally trivial design
matter, now requires renewed attention. Actually, any change on status quo in any area of
engineering always needs to be carefully considered by knowledgeable and careful
engineers, to assess all implications. zyxwv
11.12 Construction Feasibility
Pipelines are installed on the seafloor by one of the four typical installation methods: J-lay,
S-lay, Reel-lay and Tow. The J-lay and the S-lay method are shown schematically in
figs. 11.14 and 11.15 (the shape each pipe assumes justifies the corresponding name). The
reel-lay method includes one or more pipe spools on board the vessel, and the pipeline is](https://image.slidesharecdn.com/handbookofoffshoreengineeringvolume-231227165401-1b6778ee/85/HANDBOOK_OF_OFFSHORE_ENGINEERING_Volume-pdf-278-320.jpg)
![Design zyxwvutsr
for Reliabiiify: Human and Organisational Factors zyxwvu
941 z
It is becoming painfully clear that our engineering design guidelines for the creation of
sufficient robustness zyxwvu
- damage - defect tolerance in offshore structure systems is not
sufficient. Our thinking about sufficient damage stability and damage tolerance needs
rethinking. Our thinking about designing for the “maximum incredible” events needs more
development. While two offshore structures can both be designed to “resist the 100-yr
conditions” with exactly the same probabilities of failure, the two structures can have very
different robustness or damage tolerance during the life cycle of the structures. The
“minimum” CapEx offshore structure may not have a configuration, excess capacity or
ductility to allow it to weather the inevitable defects and the damage that should be
expected to develop during its life. Sufficient damage tolerance almost invariably results in
increases in CapEx; the expectation and the frequent reality are that OpEx will be lowered.
But, one must have a “long-term” view for this to be apparent.
Recent work has clearly shown that the foregoing statements about structure and hardware
robustness apply equally well to organisations and operating teams. Proper configuration,
excess capacity and ductility play out in organisations and teams in the same way that they
do in the structure and hardware [Bea, 2000a. b]. It is when the organisation or an
operating team encounters defects and damage - and is under serious stress, that the
benefits of robustness become evident. A robust organisation or an operating team is not a
repeatedly downsized (lean and mean?), out-sourced and financially strangled organisation.
A robust organisation is a Higher Reliability Organisation (HRO). zyx
12.2.4 Procedure and Software Malfunctions
Based on the study of procedure and software-related problems that have resulted in
failures of offshore structures, table 12.5 summarises a classification system for procedure
or software malfunctions. These malfunctions can be embedded in engineering design
guidelines and computer programs, construction specifications and operations manuals.
They can be embedded in contracts (formal and informal) and subcontracts. They can be
embedded in how people are taught to do things. With the advent of computers and their
integration into many aspects of the design, construction, and operation of oil and gas
structures, software errors are of particular concern because the computer is the ultimate
fool [Knoll, 1986; Rochllin, 19971.
Software errors in which incorrect and inaccurate algorithms were coded into computer
programs have been the root cause of several recent failures of offshore structures
Table 12.5 Taxonomy of procedure and software malfunctions
1
I Incorrect - faulty I
Inaccurate - untrue
Incomplete - lacking the necessary parts
1Excessive complexity - unnecessary intricacy 1
1Poor organisation - dysfunctional structure I
Poor documentation - ineffective information transmission](https://image.slidesharecdn.com/handbookofoffshoreengineeringvolume-231227165401-1b6778ee/85/HANDBOOK_OF_OFFSHORE_ENGINEERING_Volume-pdf-297-320.jpg)
![948 zyxwvutsrqpo
Chapter zy
12
[Bea, 2000a, b]. Guidelines have been developed to address the quality of computer
software for the performance of finite element analyses. Extensive software testing is
required to assure that the software performs as it should and that the documentation is
sufficient. Of particular importance is the provision of independent checking procedures
that can be used to validate the results from analyses. High-quality procedures need to be
verifiable based on first principles, results from testing and field experience. This has
particular importance in the quality assurance and quality control (QAIQC) in design.
Given the rapid pace at which significant industrial and technical developments have been
taking place, there has been a tendency to make design guidelines, construction
specifications and operating manuals more and more complex. Such a tendency can be
seen in many current guidelines used for the design of offshore structures. In many cases,
poor organisation and documentation of software and procedures has exacerbated the
tendencies for humans to make errors [Rochlin, 19971. Simplicity, clarity, completeness,
accuracy and good organisation are desirable attributes in procedures developed for the
design, construction, maintenance and operation of offshore structures. zyx
12.2.5 Environmental Influences
Environmental influences can have important effects on the performance characteristics of
individuals, organisations, hardware and software. These include:
All three of these environmental influences can have extremely important effects on human,
operating team and organisational malfunctions.
External (e.g. wind, temperature, rain, fog, time of day),
Internal (lighting, ventilation, noise, motions) and
Sociological factors (e.g. values, beliefs and morays).
12.3 Design Objectives: Life Cycle Quality, Reliability and Minimum Costs
12.3.1 Quality
In this development, the term "quality" is defined as freedom from unanticipated defects
in offshore structures. Quality is fitness for purpose. Quality is meeting the requirements
of those that own, operate, design, construct and regulate offshore structures. These
requirements include those of serviceability, safety, compatibility and durability
[Matousek, zyxwvut
19901. Quality is freedom from unanticipated defects in the serviceability,
safety, durability and compatibility of the offshore structure system.
Serviceability is suitability for the proposed purposes, i.e. functionality. Serviceability is
intended to guarantee the use of the system for the agreed purpose and under the agreed
conditions of use. Safety is the freedom from excessive danger to human life, the
environment and property damage. Safety is the state of being free of undesirable and
hazardous situations. The capacity of a structure to perform acceptably during extreme
demands and other hazards is directly related to and most often associated with safety.
Compatibility assures that the structure does not have unnecessary or excessive negative](https://image.slidesharecdn.com/handbookofoffshoreengineeringvolume-231227165401-1b6778ee/85/HANDBOOK_OF_OFFSHORE_ENGINEERING_Volume-pdf-298-320.jpg)
![Design for Reliability: Human and Organisational Factors zyxwvu
949 z
impacts on the environment and society during its life cycle. Compatibility also is the
ability of the structure to meet economic, time, political, business and environmental
requirements.
Durability assures that serviceability, safety and environmental compatibility are
maintained during the intended life of the structure. Durability is freedom from
unanticipated maintenance problems and costs. Experience with offshore structures has
shown that durability is one of the most important characteristics that must be achieved;
if insufficient durability is developed, then there are unanticipated and often undetected
degradations in the other quality characteristics, and many times these degradations have
disastrous results.
This is a holistic definition of the key objective of engineering design processes; to achieve
adequate and acceptable quality [Hessami, 19991. In recent years, a wide variety of
processes, procedures and philosophies intended to improve and achieve adequate quality
in goods and services have been developed and implemented including Total Quality
Management [Demming, 19821, QA/QC and the International Standards Organization
Quality Standards [ISO, 1994a, 1994b, 1994~1.
These components are the building blocks of
a quality management system. Engineers have learned that it is important to recognise that
these processes, procedures and philosophies are all related to the same objective; they
represent complementary parts of activities that are intended to help achieve adequate and
acceptable quality in offshore structures. The challenge has been to learn how to use these
processes, procedures and philosophies wisely, effectively and efficiently. Also, it is
important to note that the traditional “business” objectives (e.g. serviceability, compati-
bility) have been merged with the traditional “safety” objectives; quality can be good for
business and vice versa. zyxwvu
12.3.2 Reliability
Reliability is defined as the probability (likelihood) that a given level of quality will be
achieved during the design, construction and operating life cycle phases of an offshore
structure. Reliability is the likelihood that the structure system will perform in an
acceptable manner. Acceptable performance means that the structure system has desirable
serviceability, safety, compatibility and durability. The reliability, zy
Ps (likelihood of
success), can be expressed analytically as:
Ps = P[C D] (12.1)
where P[] is read as the likelihood that [ 1. D is the demandjs imposed on the system, and
C is the capacity/ies of the system to successfully withstand the imposed demand/s.
The complement of reliability is the likelihood or probability of unacceptable quality;
the probability of failure, PJ
Ps = P[D2 zyxwv
q = 1 - Ps (12.2)
This definition has linked the concepts of probability, uncertainty and reliability with the
holistic definition of quality to reflect upon the likelihoods of achieving acceptable quality
in offshore structures.](https://image.slidesharecdn.com/handbookofoffshoreengineeringvolume-231227165401-1b6778ee/85/HANDBOOK_OF_OFFSHORE_ENGINEERING_Volume-pdf-299-320.jpg)
![954 zyxwvutsrqpo
Chapter 12 z
v)
K zyxwv
0
K
K
w
c
0
w
K
K
0
0 zyxw
z
v)
c
v)
0
0
Concept Preliminaly Detailed Construction zyxwv
8 In Service
Design Design Design Commision
LIFE CYCLE PHASE zyx
Figure 12.3 Life cycle costs to correct errors
are there large direct future costs associated with fixing errors, but also large indirect costs
associated with loss of business and loss of image.
The present value of the total life cycle cost, C,associated with the performance of a system
can be expressed as:
c zyxwv
= c o +( C s t Cr+ CM+ C R ) = co +CF (12.12)
where the subscript 0 refers to the initial cost, zyxw
S refers to loss of serviceability cost, z
Zrefers
to inspection cost, M refers to structural maintenance costs, R refers to structural repair
costs and F to the total future costs associated with the maintenance of the system.
Assuming a continuous discounting, each of the individual costs can be expressed as:
cX= CC, exp(-r T,) = C,(PVF,) (12.13)
where the uppercase subscript X refers to a type of cost, the lowercase subscript zy
x refers to
the specific cost, the summation is taken over the occasions or time for the category of cost,
Y is the net discount rate, Tis the time that the expense is incurred, and PVF is the resultant
net present value function.
All of these categories of costs are variable and uncertain. Likelihoods (or probabilities)
can be entered into the process in several ways. A traditional approach has been to focus
on expected costs in which the estimated cost is multiplied by the probability, P. of
experiencing that cost:
E [CX]= CXPX (12.14)](https://image.slidesharecdn.com/handbookofoffshoreengineeringvolume-231227165401-1b6778ee/85/HANDBOOK_OF_OFFSHORE_ENGINEERING_Volume-pdf-304-320.jpg)
![956 zyxwvutsrqpo
Chapter 12 z
Co is the initial cost versus PFintercept, ACo is the slope of the initial cost curve, and z
Po is
the probability that the estimated initial cost will be realised. Given that the inspection,
maintenance, and repair costs do not vary significantly with PF,differentiating the sum of
initial and future cost with respect to PF to find the point of zero slope gives the PF that
produces the lowest total cost (Pro):
0.435
Pf zyxwvut
-- zyx
O - R,PVF
(12.20)
R,(cost ratio) is the ratio of the present valued future cost, C,, to the expected cost needed
to decrease PF by a factor of 10, ACo:
CF
- ACo
R -- (12.21)
PVF is a present value discount function. Based on continuous discounting and
replacement:
PVF = [l - (1 + zyxwv
Y)-"]/Y (12.22)
For a continuous discount function and long-life system (life > 10 yr), PVF zy
x Y-' where
r is the monetary net discount rate (investment rate minus inflation rate). For short-life
systems (life 5 5 yr), PVF x L, where L is the life in years.
As shown in fig. 12.4, as the costs associated with the development of insufficient quality
increases, the reliability must increase. As the initial costs to achieve quality increases, the
optimum reliability decreases. The optimum reliability is based on the quality that will
develop the lowest total initial and future costs. The marginal probability of insufficient
quality is double the optimum quality probability. It is the quality in which the incremental
investment to achieve quality equals the incremental future benefit (cost/benefit = 1 z
.O).
Reliability of a system element, component and system is a function of its criticality
expressed by the product of the cost ratio and present value function.
Figure 12.4 The economics and likelihood of insufficient quality](https://image.slidesharecdn.com/handbookofoffshoreengineeringvolume-231227165401-1b6778ee/85/HANDBOOK_OF_OFFSHORE_ENGINEERING_Volume-pdf-306-320.jpg)
![Design for Reliability: Human and Organisational Factors zyxwvu
959
defective. This means that some elements must be intentionally “over-designed’’ relative to
the normal loading patterns and distributions so that these elements can carry the loadings
that are transferred to them when other members or elements are damaged, defective or
fail. Appropriate correlation refers to the dependence between the strengths of paired
elements. In systems comprised of parallel elements, independence is desirable. In systems
comprised of elements in series, dependence or high correlation is desirable. This is fail-safe
or intrinsically safe design. Robust systems are not created by over zealous value
improvement programs (VIP), excessive downsizing and outsourcing and excessive initial
cost cutting (reduced CAPEX at the expense of future OPEX).
A Recent work with an HRO has clearly shown that development of robustness in
engineering organisations is a very desirable proactive measure. Such organisations can
tolerate defects and damage and still perform acceptably. This work also has shown that it
takes the same three fundamental attributes: configuration, ductility and excess capacity.
Such organisations are not downsized, out-sourced or cost-cut to the point that the
organisation cannot tolerate daily and abnormal demands. Some organisation “fat” is a
good thing when it allows the organisation to perform well when distressed.
The author has been an active protagonist and practitioner of the proactive reliability
analysis-based approach to help improve the quality of offshore structures for more than
three decades [Bea, 1974, 1975, 2000a; Marshall and Bea, 19761. He believed that this
approach provided an ability to forecast how systems could go bad. Very sophisticated
analytical models could be developed to help foster this belief. Results from these
analyses seemed to have value and to enhance his abilities to address some types of
variability and uncertainty. This approach was workable as long as he dealt with systems
in which the interactions of people with the systems were minimal or minimised.
However, the problem changed radically when people began to exert major influences
on the quality of the systems and in many cases on the physical aspects of the systems
[Bea, 1996a, b]. In this case, his lack of knowledge of the physics and mechanics of
the complex behaviours of people that in the future would design, construct, operate
and maintain the system defined an “unpredictable”, or certainly one with very limited
predictability. The author’s analytical models addressed systems that were essentially
static and mechanical. Yet the real systems were dynamic, constantly changing, and more
organic than mechanical. The analytical models generally failed to capture the complex
interactions between people and the systems that they designed, constructed, operated
and maintained.
The author found most data on the reliability of humans in performing tasks to be very
limited [Kirwan, 1994; Gertman and Blackman, 1994; Dougherty and Fragola, 1986;
Center for Chemical Process Safety, 19941. Existing databases failed to capture or
adequately characterise influences that had major effects on human reliability [Wu, et a1
1989; Haber, et a1 19911. Yet, when the numbers were supplied to the very complex
analytical models and the numbers were produced, the results were often mistaken for
“reality”. There was no way to verify the numbers. If the results indicated that the system
was “acceptable”: then nothing was done. If the results indicated that the system was “not
acceptable”, then generally the equipment and the hardware fixes were studied in an
attempt to define a fix or fixes that would make the system acceptable or ALARP zy
(As Low
As Reasonably Practicable) [Melchers, 19931. When the author went to the field to](https://image.slidesharecdn.com/handbookofoffshoreengineeringvolume-231227165401-1b6778ee/85/HANDBOOK_OF_OFFSHORE_ENGINEERING_Volume-pdf-309-320.jpg)
![960 zyxwvutsrqpo
Chapter z
12
compare his analytical models with what was really there, he found little resemblance
between his models and what was in the field [Bea, 1996bl.
The author does not advocate discarding the analytical-quantitative proactive approach.
He advocates using different types of proactive approaches to gain insights into how
systems might fail and what might be done to keep them from failing [Weick, 2000;
Bea, 2000a, b]. The marked limitations of the analytical models and the quantitative
methods must be recognised or major damage can be done to the cause of the quality
and reliability of offshore structures. The potential for engineers to be “hyper rational”
and attempt to extend the applicability of SRA/PRA/QRA methods beyond their
limitations must be recognised and countered. On the other hand, qualitative methods
(e.g. HazOp, FMEA), in the hands of qualified and properly motivated assessors (both
internal and external) can do much to help the causes of quality and reliability [Center
for Chemical Process Safety, 1989, 19941. Experience, judgement and intuition of the
assessors needs to be properly recognised, respected and fully integrated into the proactive
qualitative and quantitative approaches. Much headway has been made recently in
combining the powers of qualitative methods with quantitative Risk Assessment and
Management (RAM) methods [Bea, 2000a, b]. The qualitative methods are able to capture
more fully the dynamic, changing, organic, complex interactions that cannot be analysed
[Weick, 2000; Haber, et a1 1991; Groeneweg, 19941. Given an input from the qualitative
methods, the quantitative methods are able to provide numbers that can be used to
assist the development of judgements about when, where and how to better achieve
quality and reliability in offshore structures. But, even at this level of development, the
proactive RAM methods are very limited in their abilities to truly provide quality and
reliability in offshore structures. Other methods (e.g. interactive RAM) must be used to
address the unknowable and the unimaginable hazards.
It is the author’s experience in working with and on offshore structure systems for more
than four decades, that many if not most of the important proactive developments in the
quality and reliability of these systems were originated in a cooperative, trust-based venture
of knowledgeable “facilitators” working with seasoned veterans that have daily
responsibilities for the quality of these systems. This cooperative venture includes design,
construction/decommissioning, operations and maintenance/inspection personnel. Yet, it is
also the author’s experience, that many engineering and many well-meaning reliability z
-
risk analysis “experts” are not developing a cooperative environment. This is very
disturbing. The conduct of each operation during the life cycle of an engineered system
should be regarded as the operations of “families”. Knowledgeable, trained, experienced
and sensitive outsiders can help, encourage and assist “families” to become “better”. But,
they cannot make the families better. Families can only be changed from within by the
family members. Proactive measures based on casual or superficial knowledge of a system
or of an operation of that system should be regarded as tinkering. And, tinkering can have
some very undesirable effects and results [Wenk, 1986; Woods, 1990; Weick, 1995;
Bea, 1996, 2001al.
The crux of the problem with the proactive analytical approaches is with the severe
limitations of such approaches in their abilities to reasonably characterise human and
organisational factors and their effects on the performance of a system [Center for
Chemical Process Safety, 1994;Reason, 1997;Groeneweg, 1994;Haber, et a1 1991;Wu, et a1](https://image.slidesharecdn.com/handbookofoffshoreengineeringvolume-231227165401-1b6778ee/85/HANDBOOK_OF_OFFSHORE_ENGINEERING_Volume-pdf-310-320.jpg)
![Chapter 12 z
962
analyses [Xu, et a1 1999]? These elements are frequently referred to as being founded
in “gross errors” [Marine Technology Directorate, 1989; Bea, 19921. Experience has
adequately demonstrated that a very large amount, if not the majority of the defects and
damages we encounter in offshore structures are not in any reasonable or practical sense
“predictable” [Marine Technology Directorate, 1994; Winkworth and Fisher, 1992;
Bucknell, 2000; De Leon and Heredia-Zavoni, 20011. Other approaches (e.g. inductive
information based) must be used to address the unknowable zyx
- unpredictable aspecrs that
still must be managed in the operations of offshore structures.
Studies of the HROs (Higher Reliability Organisations) have shed some light on the factors
that contribute to errors made by organisations and risk mitigation in HRO. The HROs
are those organisations that have operated nearly error-free over long periods of time. z
A wide variety of HROs have been studied over long periods of time. The HRO research
has been directed to define what these organisations do to reduce the probabilities of
serious errors. The work has shown that the reduction in error occurrence is accomplished
by the following [Roberts, 1989, 1993; Weick, 1995; Weick, et a1 19991: (1) command by
exception or negation, (2) redundancy (robustness - defect and damage tolerance),
(3) procedures and rules, (4) selection and training, zyxw
(5) appropriate rewards and
punishment and (6) ability of management to “see the big picture”.
Command by exception (management by exception) refers to the management activity in
which the authority is pushed to the lower levels of the organisation by managers who
constantly monitor the behaviour of their subordinates. Decision-making responsibility is
allowed to migrate to the persons with the most expertise to make the decision when
unfamiliar situations arise (employee empowerment).
Redundancy involves people, procedures and hardware. It involves numerous individuals
who serve as redundant decision-makers. There are multiple hardware components that
will permit the system to function when one of the components fails. The term redundancy
is directed towards the identification of the need for organisational “robustness” - damage
and defect tolerance that can be developed, given proper configuration (deployment),
ductility - ability and willingness to shift demands, excess capacity (ability to carry
temporary overloads) and appropriate correlation (low for parallel elements. high for series
elements).
Procedures that are correct, accurate, complete, well organised, well documented and are
not excessively complex are an important part of an HRO. Adherence to the rules is
emphasised as a way to prevent errors, unless the rules themselves contribute to error.
The HROs develop constant and high quality programs of personnel selection and training.
Personnel selection is intended to select people that have natural talents for performing the
tasks that have to be performed. Training in the conduct of normal and abnormal activities
is mandatory to avoid errors. Training in how to handle unpredictable and unimaginable
unraveling of systems is also needed. Establishment of appropriate rewards and
punishment that are consistent with the organisational goals is critical; incentives are a
key to performance.
An HRO’s organisational structure is defined as one that allows the key decision-makers to
understand the big picture. These decision-makers with the big picture perceive the](https://image.slidesharecdn.com/handbookofoffshoreengineeringvolume-231227165401-1b6778ee/85/HANDBOOK_OF_OFFSHORE_ENGINEERING_Volume-pdf-312-320.jpg)
![Design for Reliabilitj,’ Human and Organisational Factors zyxwvu
913 z
12.5 Instruments to Help Achieve Design Success zyxw
Two instruments will be discussed in the remainder of this chapter that have been
developed recently to help promote more effective application of proactive, reactive and
interactive processes during the life cycle of offshore structures. A development of these
two instruments have been concentrated on taking full advantage of the progress cited in
this chapter while addressing some of the major limitations that have been recognised.
The first instrument (computer program, application protocol) is identified as a Quality
Management Assessment System (QMASO);this is fundamentally a qualitative process to
help guide assessment teams to examine the important parts of offshore structure systems
at different times during their life cycle. These assessment teams include members of the
offshore structure design engineering team system being assessed. The instrument has been
designed to elicit the insights and information that only these people can have.
The second instrument (computer program, application protocol) is a System Risk
Analysis System (SYRASO); this is a PRAiQRA/SRA instrument to help develop
quantitative results that are often required by engineers and managers. Traditional
event tree and fault tree analysis methods have been used in SYRAS. The analytical
templates in SYRAS enable the analyses of each of the life cycles of an offshore structure
and address each of the quality attributes.
A “link” has been developed between the results from the QMAS and the input required
for the SYRAS instrument. This link is based on translating the “grades” developed from
application of QMAS to performance shaping factors (PSF) that are used to modify
normal rates of human,/operator team malfunctions. The link has been developed, verified
and calibrated from the QMAS-SYRAS analyses of failures and successes of offshore
structure systems during their different life cycle phases [Bea, 2000a, b].
12.5.1 Quality Management Assessment System
The QMAS is a method that is intended to provide a level of detail between the qualitative/
less-detailed methods (e.g. HazOps, FMEA) and the highly quantitativeivery detailed
methods (PRA, QRA). The QMAS encompasses two levels of safety assessment: coarse
and detailed qualitative. The objective of the QMAS is with the least effort possible, to
identify those factors that are not of concern relative to quality and reliability, to identify
those mitigation measures that need to be implemented to improve quality and reliability
and to identify those factors that are of concern that should be relegated to more detailed
quantitative evaluations and analyses.
Components
The QMAS system is comprised of three primary components: (1) a laptop computer
program and documentation that is used to help guide platform assessments and record
their results, zyxwvuts
(2) an assessor qualification protocol and training program, and (3) a three-
stage assessment process that is started with information gathering and identification on
factors of concern (FOC), then proceeds to observe operations, and is concluded with a
final assessment and set of recommendations.](https://image.slidesharecdn.com/handbookofoffshoreengineeringvolume-231227165401-1b6778ee/85/HANDBOOK_OF_OFFSHORE_ENGINEERING_Volume-pdf-323-320.jpg)
![1006 zyxwvutsrqpo
Table 13.1 Common dimensionlessquantities in offshore engineering zyx
Chapter z
13
1zyxwvutsrqponmlkjihgfedcbaZYXWVUTSR
Symbol IDimensionless number IForce ratio IDefinition I
1Fr 1Froude Number /Inertia/Gravity ]zyxwvutsrq
u2/gD 1
1Re ~ Reynolds Number IInertia/Viscous
iEu 1Euler Number
,
IInertia/Pressure
I P I P 2 I
1Ch Cauchy Number
I
I UTID ~
1KC Keulegan-Carpenter Number Drag/Inertia
ISt ~ Strouhal Number
In table 13.1,D =member diameter, T = wave period, g = gravity, u = kinematic viscosity,
p = pressure, E = modulus of elasticity and zyxw
fe = vortex (eddy) shedding frequency. The
Froude number applies to gravity waves. The Reynolds number is related to the drag force
in the structure. The Euler number is not as important as these quantities, except for
vertically loaded long, slender structures. The Cauchy number plays an important role for
an elastic structure, such as, compliant towers, risers and tendons. The Keulegan-
Carpenter number is very important for small structural members where forces are com-
puted based on hydrodynamic inertia and drag coefficients (see Chapter 4). The Strouhal
number is the non-dimensional vortex shedding frequency and should be considered for
a moving structure when the flow past a structuralmember separates and produces vortices
past the structure.
The typical current or wave-structure interaction problem involves Froude number,
Reynolds number and Keulegan-Carpenter number. For structures that are subject to
deformation, Cauchy number should additionally be considered. For structures vibrating
in fluid medium, the Strouhal number is also included.
The frequency of vortex shedding, fe, from a stationary circular cylinder of diameter D in
a fluid stream of velocity u has been shown to be a linear function of Reynolds number Re
over a wide range. A relationship between the Strouhal number St and Reynolds number
Re exists in steady flows. It is generally accepted that St =0.2 in the range 2.5 x lo2 < Re <
2.5 x 10’. Beyond this range, St increases up to about 0.3 and then, with further increase in
Re, the regular periodic behaviour of u in the wake behind the cylinder disappears. Some
variation in this trend has been observed in experiments by several investigators,
particularly, outside the constant range of St.
In the offshore structure problem, the most common among the dimensionless scaling laws
presented in table 13.1is Froude’s law. The Reynolds number is also equally important in](https://image.slidesharecdn.com/handbookofoffshoreengineeringvolume-231227165401-1b6778ee/85/HANDBOOK_OF_OFFSHORE_ENGINEERING_Volume-pdf-355-320.jpg)
![Physical Modelling of OSfshore Structures zyxwvuts
1023 z
13.4.6 Modelling of Wind
The wind loads on the structure may be particularly important in the design of such
structures as a floating moored structure. However, it should be emphasised that these
loads in the model system are limited by the associated scaling problems. Wind loads are
functions of Reynolds number and Re is (an order of magnitude) smaller in the model
compared to the prototype Re. Therefore, it is possible that the prototype wind effect falls
in the turbulent region while the corresponding model wind effect is in a laminar region. In
this case, the model test results may be considered conservative. This is why oftentimes, it is
the properly scaled mean wind load that is simulated rather than the wind speed. In other
words, wind speed in the model is adjusted so that the mean model wind load is matched.
Wind may be generated with blowers positioned strategically in front facing the model.
In this case, the model superstructure must be accurately modelled. While wind velocity is
often taken as a steady value, the wind spectrum may be important in some applications.
The frequency range of a wind spectrum is quite broad-banded, often covering a range
from 0.005 to 1 Hz (see Chapter 3 for the description of wind spectrum model).
While a bank of fans is most commonly used in generating wind over a model super-
structure, there are several other methods of simulating wind effect on offshore structures.
The earliest method of simulating a steady wind load on an offshore structure is still
in use. It is achieved with a weight hanging in the direction of wind from the model
with the help of pulleys and strings. A fan mounted on the deck of a floating model has
also been applied to simulate steady and oscillating wind load. The pros and cons of these
three wind simulation methods in a model scale are compared in table 13.7. Each method
has its place in model testing.
To demonstrate the capabilities of the model mounted fan, we present an example of a
wind condition that was considered in a model test of a floating platform. Table 13.8
summarises the parameters that were used in wind spectral simulation.
There are two separate data files needed for the test. The first file is the time history of the
signal that is used to control the pitch angle of the generator blades. The second one is a
time history of the load computed from the wind drag formula [equation (4.13)]. This load
is considered appropriate for the generation of wind load on the model. The measured load
during the test is compared with the computed one. Adjustments are made in the control
signal to obtain a satisfactory match. For the above example, the desired and computed
wind load spectra are shown in fig. 13.5. The comparison between the two in the load
spectra is considered acceptable in this case.
13.4.7 Modelling of Current
In many deepwater locations considerable current prevails. In the design of structures in
these locations, the effect of current may be extremely important. For such structures,
simulation of current in the wave basin is essential. While current may be simulated well
with towing the model (already discussed), this method may not be adequate for many
offshore components.
There are several options in generating local current in the general area of the model
in the test basin [Chakrabarti, 19941. The current generator should, in general, have the](https://image.slidesharecdn.com/handbookofoffshoreengineeringvolume-231227165401-1b6778ee/85/HANDBOOK_OF_OFFSHORE_ENGINEERING_Volume-pdf-372-320.jpg)
![Ofshore InsiaNarion zyxwvutsrq
Table 14.5 Dynamic amplificationfactors zyxw
1107 z
Weight (ton)
Offshore Inshore
-~ ~
1100-1000 11.20
11000-2500 11.15
~ 2500 11.10 11.05 1
Z,(t] is the vessel heave.
M
Mu
niWire
C
K
is the structure dry weight.
is the structure added mass.
is the dry weight of the hoisting wires.
is the vertical drag coefficient of the structure.
is the stiffness of the hoisting wires.
The dynamic amplification factor due to the heave motions is given by:
(14.13)
K
[((M+Mu +( ~ i r e / 2 . 7 1 8 ) ) ~ ~
- K)2+(C~)2]o'5
DAF =
where w is the natural frequency of the spring (rigging) and the mass (structure) system.
Load in Rigging
For lifts that are statically indeterminate. such as a deck lifted by four slings connected to a
single hook, the load distribution of the individual sling is affected by variations in the
lengths of the slings as well as the relative distance of their attachment points from the
centre of gravity (COG). This mis-match causes a skew load which is accounted for using
one of the several methods available in the industry:
Assume that one diagonally opposite pair of slings carries 75% of the weight, while
the other carries 25%. A similar approach is to apply a skew load factor equal to 1.25
to the worst loaded sling. For a symmetrical lifting arrangement with a COG located in
the centre, this approach is equivalent to a 62.5-37.5% distribution, which is less
onerous than the first one [Noble Denton Guidelines 00271.
Provided the mis-match does not exceed 1.5 in. or 0.25% of the length of any one sling
or 3 in. or 0.5% difference between the shortest and the longest slings, it can be ignored
as recommended by the API RP2A. Mis-matches of greater magnitude need to be
accounted for analytically.
The skew load factor can be calculated analytically as recommended by the Det Norske
Veritas Rules for Marine Operations (1996). For a 4-point lift with double symmetrical
sling arrangement, the skew load factor, SKF, is given by equation (14.14).
(14.14)
E
O
SKF= 1 +- zyxw
E](https://image.slidesharecdn.com/handbookofoffshoreengineeringvolume-231227165401-1b6778ee/85/HANDBOOK_OF_OFFSHORE_ENGINEERING_Volume-pdf-456-320.jpg)
HANDBOOK_OF_OFFSHORE_ENGINEERING_Volume.pdf
- 1. HANDBOOK OF OFFSHORE ENGINEERING SUBRATA K. CHAKRABARTI Offshore Structure Analysis, Inc. Plainfield, Illinois, USA Volume I1 2005 Amsterdam zyxwvut - Boston - Heidelberg - London - New York - Oxford Paris - San Diego - San Francisco - Singapore - Sydney - Tokyo
- 2. Elsevier The Boulevard Langford Lane, Kidlington, Oxford OX5 lGB, UK Radarweg 29, PO Box 211, 1000 zyxwvu AE Amsterdam, The Netherlands First edition 2005 Reprinted 2005, 2006 Copyright zyxwvuts Q 2005 Elsevier Ltd. All rights reserved No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means electronic, mechanical, photocopying, recording or otherwise without the prior written permission of the publisher Permissions may be sought directly from Elsevier’s Science & Technology Rights Department in Oxford UK: phone (+44) (0) 1865 843830; fax (+44) (0) 1865 853333; email: permissions@elsevier.com. Alternatively you can submit your request online by visiting the Elsevier web site at zyxwv http://elsevier.comllocate/permissions, and selecting zy Obtainingpermission to use Elsevier material Notice No responsibility is assumed by the publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein. Because of rapid advances in the medical sciences, in particular, independent verification of diagnoses and drug dosages should be made British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library Library of Congress Cataloging-in-PublicationData A catalog record for this book is available from the Library of Congress ISBN-13: 978-0-08-044568-7 (v01 1) ISBN-10: 0-08-044568-3 (VO~ 1) ISBN-13: 978-0-08-044569-4 (v012) ISBN-10: 0-08-044569-1 (VO~ 2) ISBN-1 3: 978-0-08-044381-2 (set) ISBN-10: 0-08-044381-8 (set) For information on all Elsevier publications i visit our website at books.elsevier.com Printed and bound in Great Britain 06 07 08 09 10 10 9 8 zyxwvut 7 6 5 4 3
- 3. V PREFACE zyxwv Due to the rapid growth of the offshore field, particularly in the exploration and develop- ment of offshore oil and gas fields in deep waters of the oceans, the science and engineering in this area is seeing a phenomenal advancement. This advanced knowledge is not readily available for use by the practitioners in the field in a single reference. Tremendous strides have been made in the last decades in the advancement of offshore exploration and production of minerals. This has given rise to developments of new concepts and structures and material for application in the deep oceans. This has generated an obvious need of a reference book providing the state-of-the art in offshore engineering. This handbook is an attempt to fill this gap. It covers the important aspects of offshore structure design, installation and operation. The book covers the basic background material and its application in offshore engineering. Particular emphasis is placed in the application of the theory to practical problems. It includes the practical aspects of the offshore structures with handy design guides, simple description of the various components of the offshore engineering and their functions. One of the unique strengths of the book is the impressive and encompassing presen- tation of current functional and operational offshore development for all those involved with offshore structures. It is tailored as a reference book for the practicing engineers, and should serve as a handy reference book for the design engineers and consultant involved with offshore engineering and the design of offshore structures. This book emphasizes the practical aspects rather than the theoretical treatments needed in the research in the field of offshore engineering. In particular, it describes the dos and don’ts of all aspects of offshore structures. Much hands-on experience has been incorporated in the write up and contents of the book. Simple formulas and guidelines are provided throughout the book. Detailed design calculations, discussion of software development, and the background mathematics has been purposely left out. The book is not intended to provide detailed design methods, which should be used in conjunction with the knowledge and guidelines included in the book. This does not mean that they are not necessary for the design of offshore structures. Typically, the advanced formulations are handled by specialized software. The primary purpose of the book is to provide the important practical aspects of offshore engineering without going into the nitty gritty of the actual detailed design. Long derivations or mathematical treatments are avoided. Where necessary, formulas are stated in simple terms for easy calculations. Illustrations are provided in these cases. Information is provided in handy reference tables and design charts. Examples are provided to show how the theory outlined in the book is applied in the design of structures. Many examples are borrowed from the deep-water offshore structures of interest today including their components, and material that completes the system.
- 4. vi Contents of the handbook include the following chapters: Historical Development of Offshore Structures Novel and Marginal Field Offshore Structures Ocean Environment Loads and Responses Probabilistic Design of Offshore Structure Fixed Offshore Platform Design Floating Offshore Platform Design Mooring Systems Drilling and Production Risers Topside Facilities Layout Development Design and Construction of Offshore Pipelines Design for Reliability: Human and Organisational Factors Physical Modelling of Offshore Structures Offshore Installation Materials for Offshore Applications Geophysical and Geotechnical Design The book is a collective effort of many technical specialists. Each chapter is written by one or more invited world-renowned experts on the basis of their long-time practical experience in the offshore field. The sixteen chapters, contributed by internationally recognized offshore experts provide invaluable insights on the recent advances and present state-of-knowledge on offshore developments. Attempts were made to choose the people, who have been in the trenches, to write these chapters. They know what it takes to get a structure from the drawing board to the site doing its job for which it is designed. They work everyday on these structures with the design engineers, operations engineers and construction people and make sure that the job is done right. Chapter 1 introduces the historical development of offshore structures in the exploration and production of petroleum reservoirs below the seafloor. It covers both the earlier offshore structures that have been installed in shallow and intermediate water depths as well as those for deep-water development and proposed as ultra-deep water structures. A short description of these structures and their applications are discussed. Chapter 2 describes novel structures and their process of development to meet certain requirements of an offshore field. Several examples given for these structures are operating in offshore fields today. A few others are concepts in various stages of their developments. The main purpose of this chapter is to lay down a logical step that one should follow in developing a structural concept for a particular need and a set of prescribed requirements. The ocean environment is the subject of chapter 3. It describes the environment that may be expected in various parts of the world and their properties. Formulas in describing their magnitudes are provided where appropriate so that the effect of these environments on the structure may be evaluated. The magnitudes of environment in various parts of the world are discussed. They should help the designer in choosing the appropriate metocean conditions that should be used for the structure development.
- 5. vii Chapter 4 provides a generic description of how to compute loads on an offshore struc- ture and how the structure responds to these loads. Basic formulas have been stated for easy references whenever specific needs arise throughout this handbook. Therefore, this chapter may be consulted during the review of specific structures covered in the handbook. References are made regarding the design guidelines of various certifying agencies. Chapter zyxwvutsr 5 deals with a statistical design approach incorporating the random nature of environment. Three design approaches are described that include the design wave, design storm and long-term design. Several examples have been given to explain these approaches. The design of fixed offshore structures is described in Chapter 6. The procedure follows a design cycle for the fixed structure and include different types of structure design including tubular joints and fatigue design. Chapter zyxwvutsr 7 discusses the design of floating structures, in particular those used in offshore oil drilling and production. Both permanent and mobile platforms have been discussed. The design areas of floaters include weight control and stability and dynamic loads on as well as fatigue for equipment, risers, mooring and the hull itself. The effect of large currents in the deepwater Gulf of Mexico, high seas and strong currents in the North Atlantic, and long period swells in West Africa are considered in the design development. Installation of the platforms, mooring and decks in deep water present new challenges. Floating offshore vessels have fit-for-purpose mooring systems. The mooring system selection, and design are the subject of Chapter 8. The mooring system consists of freely hanging lines connecting the surface platform to anchors, or piles, on the seabed, positioned some distance from the platform. Chapter 9 provides a description of the analysis procedures used to support the operation of drilling and production risers in floating vessels. The offshore industry depends on these procedures to assure the integrity of drilling and production risers. The description, selection and design of these risers are described in the chapter. The specific considerations that should be given in the design of a deck structure is described in Chapter 10. The areas and equipment required for deck and the spacing are discussed. The effect of the environment on the deck design is addressed. The control and safety requirements, including fuel and ignition sources, firewall and fire equipment are given. The objective of chapter 11 is to guide the offshore pipeline engineer during the design process. The aspects of offshore pipeline design that are discussed include a design basis, route selection, sizing the pipe diameter, and wall thickness, on-bottom pipeline stability, bottom roughness analysis, external corrosion protection, crossing design and construction feasibility. Chapter 12 is focused on people and their organizations and how to design offshore structures to achieve desirable reliability in these aspects. The objective of this chapter is to provide engineers design-oriented guidelines to help develop success in design of offshore structures. Application of these guidelines are illustrated with a couple of practical examples. The scale model testing is the subject of Chapter 13. This chapter describes the need, the modeling background and the method of physical testing of offshore structures in a
- 6. ... zyxwvutsrq Vlll zyxwvutsrqponmlkj small-scale model. The physical modeling involves design and construction of scale model, generation of environment in an appropriate facility, measuring responses of the model subjected to the scaled environment and scaling up of the measured responses to the design values. These aspects are discussed here. Installation, foundation, load-out and transportation are covered in Chapter 14. Installa- tion methods of the following sub-structures are covered: Jackets; Jack-ups; Compliant towers and Gravity base structures. Different types of foundations and their unique methods of installation are discussed. The phase of transferring the completed structure onto the deck of a cargo vessel and its journey to the site, referred to as the load-out and transportation operation, and their types are described. Chapter 15 reviews the important materials for offshore application and their corrosion issues. It discusses the key factors that affect materials selection and design. The chapter includes performance data and specifications for materials commonly used for offshore developments. These materials include carbon steel, corrosion resistant alloys, elastomers and composites. In addition the chapter discusses key design issues such as fracture, fatigue, corrosion control and welding. Chapter 16 provides an overview of the geophysical and geotechnical techniques and solutions available for investigating the soils and rocks that lay beneath the seabed. A project’s successful outcome depends on securing the services of highly competent contractors and technical advisors. What is achievable is governed by a combination of factors, such as geology, water depth, environment and vessel capabilities. The discussions are transcribed without recourse to complex science, mathematics or lengthy descriptions of complicated procedures. Because of the practical nature of the examples used in the handbook, many of which came from past experiences in different offshore locations of the world, it was not possible to use a consistent set of engineering units. Therefore, the English and metric units are interchangeably used throughout the book. Dual units are included as far as practical, especially in the beginning chapters. A conversion table is included in the handbook for those who are more familiar with and prefer to use one or the other unit system. This handbook should have wide applications in offshore engineering. People in the follow- ing disciplines will be benefited from this book: Offshore Structure designers and fabricators; Offshore Field Engineers; Operators of rigs and offshore structures; Consulting Engineers; Undergraduate & Graduate Students; Faculty Members in Ocean/Offshore Eng. & Naval Architectural Depts.; University libraries; Offshore industry personnel; Design firm personnel. zyxwvut Subrata Chakrabarti Technical Editor
- 7. TABLE OF CONTENTS zyxw Preface zyxwvuts ........ v Abbreviations ................................................................................................................ ix Conversion Factors List of Contributors................................................................ Chapter 8. lMooring Systems....................................................................................... 663 8.1 Introduction ........................................................................................................................ 8.2 Requirements ...................................................................................................................... 8.3 Fundamentals ..................................................................................................................... 8.3.1 Catenary Lines ............................ 8.3.2 Synthetic Lines.............................................................. 8.3.3 Single Catenary Line Performance Characteristics ................................................. 8.4 Loading Mechanisms .......................................................................................................... 8.5 Mooring System Design 8.5.1 Static Design............................................................................................................ 8.5.3 Dynamic Design ................................................................ 8.5.5 Effective Water Depth ............................................................................................. 8.5.7 Uncertainty in Line Hydrodynamic Coefficients ......... 8.5.8 Uncertainty in Line Damping and Tension Prediction ........................................... 8.6 Mooring Hardware Components ........................................................................................ 8.6.1 Chain ....................................................................................................................... 8.6.2 Wire Rope ............................................................................................................... 8.6.3 Properties of Chain and Wire Rope .............................................................. 8.6.4 Moorings ............................................................................................ 8.6.5 Connectors ............................................................................................................... 8.6.6 Shipboard Equipment .............................................................................................. 8.6.7 Anchors ................................................ 8.6.8 Turrets .......................................................................... Industry Standards and Classification Rules...................................................................... 8.7.1 Certification ............................................................................................................. 8.7.2 Environmental Conditions and Loads .................................................................... 8.7.4 Thruster-Assisted Mooring ................................................. 8.7.5 Mooring Equipment ................................................................................................ 8.7.6 Tests......................................................................................................................... 8.5.2 Quasi-Static Design ................................................ 8.5.4 Synthetic Lines......................................................................................................... 8.5.6 Mooring Spreads ................................................ 8.7 8.7.3 Mooring System Analysis .............................................. 663 665 665 665 669 670 671 675 675 676 677 680 680 680 681 684 687 687 688 689 689 689 693 693 694 696 697 697 699 704 705 706
- 8. XVI Chapter 9 zyxwvut . Drilling and Production Risers...................................................................z 9.1 9.2 9.3 9.4 9.5 9.6 9.7 9.8 9.9 9.10 9.11 Introduction ........................................................................................................................ 9.2.1 Design Background ............................................................... 9.2.2 Influence of Metocean Conditions .......................................................................... 9.2.3 Pipe Cross-Sect ................................................................ 9.2.4 Configuration ( .............. ........ 9.2.5 Vortex-Induced ............................................................................. 9.2.6 Disconnected Riser .................................................................................................. 9.2.7 Connected Riser............ .................... 9.2.8 Emergency Disconnect Sequence (EDS)!Drift-Off An 9.2.9 Riser Recoil after EDS .................................................................................... Production Risers ............................................................................................................... 9.3.1 Design Philosophy and Background ....................................................................... 9.3.2 Top Tension Risers.................................................................................................. 9.3.3 Steel Catenary Risers (Portions contributed by Thanos Moros & Howard Cook, BP America, Houston, TX) ................................................... 9.3.4 Diameter and Wall Thickness ................................................................................. 9.3.5 9.3.6 In-Service Load Combinations ................................................................................ 9.3.7 Accidental and Temporary Design Cases................................................................ Vortex Induced Vibration of Risers 9.4.1 VIV Parameters ............................................................................................... 9.4.2 Simplified VIV Analysis .......................................................................................... 9.4.3 Examples of VIV Analysis....................................................................................... 9.4.4 Available Codes ............ VIV Suppression Devices............................................................................................ Riser Clashing.......................... 9.6.1 Fatigue Analysis ................................................................................................................. 9.7.1 9.7.2 Fatigue Due to Riser VIV .... 9.7.3 Fatigue Acceptance Criteria ............................................................................ Fracture Mechanics Assessment ......................................................................................... 9.8.1 Engineering Critical Assessment .............................................................................. 9.8.2 Paris Law Fatigue Analysis ........ ................................................. 9.8.3 Acceptance Criteria ..... .......... Reliability-Based Design ..................................................................................................... Design Verification ........ ........................................................................................... Design Codes ................... ................................................. Drilling Risers ...................... .................................................. SCR Maturity and Feasibility ............................... ............. Clearance, Interference and Layout Considerations ....................................... First and Second Order Fatigue .............................................................................. 9.8.4 Other Factors to Consi ......................................... Chapter 10. Topside Facilities Layout Development.................................................... 709 709 714 715 715 715 718 726 730 744 757 166 768 769 779 802 817 824 826 828 828 828 829 832 832 832 836 836 838 842 845 848 849 850 851 851 851 851 853 854 861 10.1 Introduction ........................................................................................................................ 861 10.2 General Layout Considerations .......................................................................................... 862
- 9. 10.2.1 General Requirements zyxwvu ........................................................................................... 10.2.2 Deepwater Facility Considerations........................................................................ 10.2.3 Prevailing Wind Direction ......................................... 10.2.4 Fuel and Ignition Sources...................................................................................... 10.2.5 Control and Safety Systems................................................................................... 10.2.6 Firewalls, Barrier Walls and Blast Walls............................................................... 10.2.7 Fire Fighting Equipment ......................................... 10.2.8 Process Flow........................................................................................... 10.2.9 Maintenance of Equipment ................................................................................... 10.2.10 Safe Work Areas and Operations ............................ 10.2.11 Storage.......................................... 10.2.12 Ventilation ............................................................................................................. 10.2.13 Escape Routes ....................................................................................................... 10.3 Areas and Equipment .................................................................. 10.3.1 Wellhead Areas ................................................. 10.3.2 Unfired Process Areas ............................................................... 10.3.3 Hydrocarbon Storage Tanks ................................................................................. 10.3.4 Fired Process Equipment ....................................................................................... 10.3.5 Machinery Areas ...................................................... 10.3.6 Quarters and Utility Buildings ............................................................... ...... 10.3.7 Pipelines ................................................................................................................. 10.3.8 Flares and Vents......... ............................ Deck Placement and Configuration ................................................................................... Horizontal Placement of Equipment on Deck ...................................................... Vertical Placement of Equipment .............................................. 10.4 Deck Impact Loads ............................................................................................. 10.5 10.5.1 10.5.2 10.5.3 Installation Considerations .................................................................................... 10.5.4 Deck Installation Schemes..................................................................................... 10.6 Floatover Deck Installation ............................................................................................... 10.7 Helideck ........... ............................................................................................ 10.8 Platform Crane ............................................................................................ 10.9 Practical Limit Analysis of Two Example Layouts .................................................................................... 10.10 10.11 Example North Sea Britannia Topside Facility ................................................................. zy Chapter 11. Design and Construction of Offshore Pipelines........................................ 11.1 Introduction 11.2 Design Basis......................................... 11.3 Route Selection and Marine Survey................................................................................... 11.4 Diameter Selection.............................................................................................................. 11.4.1 Sizing Gas Lines ................................................................................................... 11.4.2 Sizing Oil Lines ............... 11.5 Wall Thickness and Grade ................................................................................................. 11.5.1 Internal Pressure Containment (Burst )................................................................. xvii 864 865 866 867 869 869 869 869 870 870 870 871 872 872 872 872 873 873 873 874 874 874 875 876 876 876 877 877 879 881 883 883 883 887 891 891 892 893 893 893 895 895 896 11.5.2 Collapse Due to External Pressure........................................................................ 897
- 10. xviii 11.5.3 Local Buckling Due to Bending and External Pressure zyxw ....................................... 11.5.4 Rational Model for Collapse of Deepwater Pipelines .......................................... 11.6 Buckle Propagation 11.7 Design Example ............................................................................. 11.7.1 Preliminary Wall Thickness for Internal Pressure Containment (Burst) .............................................................................................. 11.7.2 Collapse Due to External Pressure ...................................... 11.7.3 Local Buckling Due to Bending and External Pressure ....................................... 11.7.4 Buckle Propagation ................................................................................................ 11.8.1 Soil Friction Factor .................................... 11.8.2 Hydrodynamic Coefficient Selection .................................................................... 11.8.3 Hydrodynamic Force Calculation ......................................................................... 11.8.4 Stability Criteria .................................................................................................... 11.9.1 11.9.2 Design Example ..................................................................................................... 11.IO External Corrosion Protection .......................................................................................... 11.10.1 Current Demand Calculations ............................................................................. 11.10.2 Selection of Anode Type and Dimensions ............................................................ 11.10.3 Anode Mass Calculations...................................................................................... 11.10.4 Calculation of Number of Anodes 11.10.5 Design Example ................................................................................................... 11.11 Pipeline Crossing Design.................................................................................................... 11.8 On-Bottom Stability ......................................................... 11.9 Bottom Roughness Analysis ........................ Allowable Span Length on Current-Dominated Oscillations 11.12 Construction Feasibility ...................... 11.12.1 J -lay Installatio .......................................................... 11.12.3 Reel-lay ............... 11.12.4 Towed Pipelines ..................................................................................................... 11.12.2 S-lay....................................................................................................................... zy Chapter 12. Design for Reliability: Human and Organisational Factors ..................................................................................................... 12.1 Introduction ........................................................................................................................ 12.2.1 Operator Malfunctions .......................................................................................... 12.2.2 Organisational Malfunctions ................................................................................. 12.2.3 Structure, Hardware, Equipment Malfunctions .................................................... 12.2.4 Procedure and Software Malfunctions ................................................ 12.2.5 Environmental Influences ........................... 12.3.1 Quality ................................................................................................................... 12.3.2 Reliability............................................................................................................... 12.3.3 Minimum Costs ..................................................................................................... Approaches to Achieve Successful Designs 12.4.1 Proactive Approaches ..................................................................... 12.2 Recent Experiences of Designs Gone Bad ....... 12.3 Design Objectives: Life Cycle Quality, Reliability a ........................... 12.4 899 900 905 907 908 910 911 911 912 913 913 914 914 914 916 917 917 918 919 919 920 920 921 927 929 932 933 933 939 939 939 942 944 946 947 948 948 948 949 952 957 958
- 11. XIX zy 12.4.2 Reactive Approaches zyxwvu ..................... ...................... ........ 12.4.3 Interactive Approaches .......................................................................................... Instruments to Help Achieve Design Success....... 12.5.1 Quality Management Assessment System 12.5.2 12.6.1 Minimum Structures ....... 12.6.2 Deepwater Structure roject.................................. ..................... . Summary and Conclusions .........._ .....,,,,,,,,,......_. ..................._. ._. __. ..................................... 12.5 System Risk Assessment System............................................................................ 12.6 Example Applications........................................................................................................ 12.7 965 968 973 973 919 984 984 990 992 z Chapter 13. Physical Modelling of Offshore Structures............................................. 1001 13.1 13.2 13.3 13.4 13.5 13.6 13.7 13.8 Introduction........,.,,,................................... 13.1.1 History of Model Testing ........ 13.1.2 Purpose of Physical Modelling Modelling and Similarity Laws .................. 13.2.1 Geometric Similitude ............................................................................................. 1005 13.2.2 Kinematic Similitude ....... 13.2.3 Hydrodynamic Similitude .............. 13.2.4 Froude Model ........................................................................................................ 1007 13.2.5 Reynolds Model..................................................................................................... 1007 13.2.6 Cauchy Model ................. ....................................................................... 1014 Model Test Facilities ..... 1015 13.3.1 Physical Dimensions ............................................................................................... 1016 13.3.2 Generation of Waves, Wind and Current ............................................................. 1019 Modelling of Environment ........... .......................................... 1019 13.4.1 Modelling of Waves ......... 13.4.2 Unidirectional Random Waves 13.4.3 ........................ 1020 13.4.4 White Noise Seas .................................................................................................. 1021 13.4.5 Wave Grouping ...................................................................................................... 1022 13.4.6 Modelling of Wind ....... 13.4.7 ...................... 1023 Model Calibration .............................................................................................................. 1026 13.5.1 Measurement of Mass Properties ........._._..................................................,,,,.,..,,,. 1027 Field and Laboratory Instrumentation ...... 13.6.1 Type of Measurements ...................................................... .._._.............1030 13.6.2 Calibration of Instruments ............ Pre-Tests with Model zyxwvu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1033 13.7.1 Static Draft. Trim and Heel 13.7.2 Inclining Test ....................... ........................................................................ 1033 13.7.3 Mooring Stiffness Test.....,,,,,,,,,,,,,..................,,,,.........,..,,,................................,,,. 1034 13.7.4 Free Oscillation Test.............................................................................................. 1034 13.7.5 Towing Resistance Test ....................................................................................... 1035 Moored Model Tests in Waves and Current ..................................................................... 1035 13.8.1 Regular Wave Tests ...... . 1035 13.8.2 White Noise Test ............................................................... ................ 1036 ....................... ........................ Multi-directional Random Waves ..................................... Modelling of Current ......................................................... ......................
- 12. xx zyxwvutsrqponm 13.8.3 Irregular Wave Tests zyxwvu ................ .......................................................... 1036 13.8.4 Second-Order Slow Drift Tests.............................. ..................... 1036 13.9.1 Density Effects........ .......................................................... 1037 13.9.2 Cable Modelling ........................................ ........................................ 1037 13.9 Distorted Model Testing............................................................................................... 13.9.3 Modelling of Mooring Lines, Risers and Tendons ............................................................................................ 1038 ........................................................... 1042 ....................................... 1044 13.10 Ultra-deepwater Model Testing ... 13.10.1 Ultra Small-scale Testing ........................................................... 1043 13.10.2 Field Testing................................................ 13.10.3 Truncated Model Testing ............................................... 13.10.4 Hybrid Testi .................................................................................... 1046 13.11.1 Data Acquisi em .............................................. .................... 1050 13.11.3 Data Analysis................... ........................................................... 1051 13.11 Data Acquisition and ........................................ 1050 13.11.2 Quality Ass zyxwvut Chapter 14. Offshore Installation ................................................................................ 1055 14.1 14.2 14.3 14.4 14.5 14.6 Introduction .................................... ........................................................... 1055 Fixed Platform Substructures .................................................. ..................... 1056 14.2.2 Jackets ................... ............................................................................. 1056 14.2.3 Compliant Towers ............................................................................. 1059 14.2.4 Gravity Base Struc ............................................................................. 1061 Floating Structures ................................................. ........................................ 1063 14.3.1 Types of Floating Structures ......................................... ..................... 1063 14.2.1 Types of Fixed Platform Substructures ................................................................. 1056 14.3.2 Installation of FPSOs ....................................................................... 14.3.5 Spar Installation ............................................................ ...................... 1070 14.4.1 Types....................... ................................................................... 1072 14.4.2 Driven Piles................................. ............................................... 1073 14.4.3 Drilled and Grouted Piles............................ .............................. 1074 14.4.4 Suction Embedded Anchors ............................................................. 14.4.5 Drag Embedded Anchors ...................................................................................... 1078 14.5.1 Template Installation ........................................................................ 1079 14.5.2 Positioning and Monitoring..................................................................... 1080 14.5.3 Rigging Requirements....................................................................... 1081 14.5.4 Existing Subsea Facilities....................................................................................... 1082 Subsea Templates ......................................... ....................................... 1079 14.5.5 Seabed Preparation ............................................................................. 1082 Loadout .................................................... ........................................... 1082 14.6.1 Loadout Methods .......................................................... .................... 1082 14.6.2 Constraints ........................................................................... ... 1085 14.6.3 Structural Analysis .................................................................................... 1086
- 13. xx1 z 14.7 Transportation zyxwvuts ........... ............................ 14.7.1 Configuration ............................................................................ 14.7.2 Barges and H 14.7.4 Transport Route ....................................................................... 14.7.5 Motions and 14.7.6 SeafasteningdTie downs ........................................................................................ 1095 14.7.7 Structural Analysis ......................................................... 1095 .................................................. 1096 14.7.8 Inundation, Slamming ........ 14.8 Platform Installation Methods .................................. ............................. 1097 14.8.2 Launch ..................................................................................... 1098 14.8.3 Mating ...................................... .................................................. 1099 14.8.4 Hook-up to Pre-Installed Mooring Lines................................. 14.7.3 Design Criteria and Meteorological Data ............................. 1090 14.9.2 Heavy Lift.. ..................................................................................... 1106 14.9.3 Launching .......................... ......................................................... 1110 14.9.4 Unpiled Stability .................................................. 14.9.7 Tension Leg Platforms .............................................................................. 1 14.9.8 Spar............... ................................................................ 1 14.9.9 FPSO.................... ................................................................ 1 14.10.2 Methods of Pipeline Installation ........................................................................... 1 13 14 14 16 16 16 14.10.3 Types of Risers .................. .................................................. 1119 14.10.4 Methods of Ris 14.10.5 Vessel and Equ 14.10.6 Analyses Required .............................................................................. 1121 Chapter 15. Materials for Offshore Applications ........................................................ 1127 15.1 Introduction ............................. ......................................................... 1127 15.1.1 Factors Affecting Mat ......................................................... 1127 ............................. 1128 15.1.2 Classification of Materials .............. 15.2 Structural Steel ............................................................................................. 1128 15.3 Topside Materials ............................................................................................................... 1130 15.3.1 Materials Applications .............................................................................. 1131 15.3.2 Materials for Seawater ........................................... 1132 15.3.3 Materials for Process Piping and Equipment........................... 1132 15.4 Material for HPHT Applications ....................................................................................... 1133 15.4.1 Limitations of Materials for HPHT Application .................................................. 1133 15.5 Advanced Composite Materials.......................................................................................... 1134 15.6 Elastomers ........................ ................................................................ 1135
- 14. xxii 15.7 Corrosion Control zyxwvut ................................................................ 1137 15.8 Material Reliability and Monitoring .................................................................................. 1138 15.9 Fracture Control................................................................................................................. 1138 z Chapter 16. Geophysical and Geotechnical Design...................................................... 1145 16.1 Preface ............................................................................................ 1145 16.2 Introdu ............................................................................................ 1146 16.2.2 Desk Studies and Planning .................................... 1148 16.2.3 Specifications ......................................................................................................... 1148 16.2.4 Applications ........................................................................................................... 1149 16.3 Geophysical Techniques . .............................................................................. 1152 16.3.1 General............... ............................................................................................ 1152 16.3.2 High-Resolution Reflection Systems ...... 1154 16.3.3 Sounders .............................................................................. 1156 16.3.4 Side-Scan Sonar ..................................................................................................... 1158 16.3.5 Sub-Bottom Profilers ............................................................................................. 1160 16.3.7 Use of Data. .................................... 1164 16.4 Remote Geophysical Platforms .................................... 1165 16.4.1 Remotely Operated Ve .................................................. 1165 16.4.2 Autonomous Underwa .................................... 1165 Seabed Classification Systems ............................................................................................ 1166 16.2.1 Regulations, Standards and Permits................................................ 1147 16.3.6 Marine Magnetometer ..................................................................... 1163 16.5 16.7 Electrical Resistivity Systems 16.8 Underwater Cameras ............ 16.9 Geotechnical Techniques .................................................................................................... 1172 16.9.1 General.......... 1172 16.9.2 Vessels and Rigs .............................................................................. 1173 16.9.3 Methods of Drilling and Sampling........................................................................ 1179 16.9.4 Shallow Soil Sampling and Rock Coring Systems .......................... 16.9.5 Basic Gravity Corer........ ......................................................... 16.9.6 Kullenberg Device ................................................................................................. 1192 16.9.7 Piston Corer ........................................................................................................... 1193 16.9.8 Abrams Corer ........................................................................................................ 1195 16.9.9 Vibrocorer ...................................................... 16.9.10 High Performance CorerTM.......................... 16.9.11 Box Corers ............................................................................................................ 1199 16.9.12 Push-In Samplers................................................................................................... 1200 16.9.13 Grab Samplers....................................................................................................... 1201 16.10.1 Cone Penetration Testing (CPT) Systems 16.10.2 Minicones ............................................................................................ 1209 16.10.3 The ROV ............................................................................................ 1210 16.10.4 Vane Test............................................................. 16.10.5 T-Bar Test ..................................... 16.6 Seismic Refraction Systems ............................................................ 16.10 In situ Testing Systems .........................................
- 15. xxiii z 16.10.6 Piezoprobe Test zyxwvu .............................................................. 1216 16.10.7 Other In Situ Tests .............................................................................................. 1217 16.11 Operational Considerations................................................................................................ 1218 16.11.2 Water Depth Measuring Procedures ............................... 1219 16.11.3 Borehole Stability ................................................................................................. 1221 16.11.4 Blowout Prevention ............................................................................................. 1221 ..... 1223 16.13.1 General................................................................................................................. 1223 16.13.2 Conventional Laboratory Testing........................................................................ 1224 16.13.3 Advanced Laboratory Testing ............................................................................. 1229 1237 16.14.1 Pile Design ........................................................................................................... 1237 16.11.1 Horizontal Control or Positioning ............................. 1218 16.12 Industry Legislation. Regulations and Guidelines............................................................. 1221 16.13 Laboratory Testing ................................................ 16.14 Offshore Foundation Design ........................................................ 16.14.2 Axial Pile Capacity ............................................ 1238 16.14.3 Axial Pile Response ................................................. ............. 1248 16.14.5 Other Considerations ............................................................................ 1254 16.14.6 16.14.7 Pile Drivability Analyses and Monitoring ....... Supplementary Pile Installation Procedures ........................................... 16.15.3 Shallow Foundation Settlement Analyses ........................................................... 1262 16.16 Spudcan Penetration Predictions ...... 16.17 ASTM Standards .................................... ..................... 1264 Index....................................................................................................
- 16. Handbook of Offshore Engineering zyxwvuts S. Chakrabarti (Ed.) zyxwvutsrq C 2005 Elsevier Ltd. zyxwvutsr All rights reserved 663 Chapter 8 Mooring Systems David T. Brown BPP Technical zyxwvuts Services Ltd., Loizdon, UK 8.1 Introduction It is essential that floating offshore vessels have fit-for-purpose mooring systems. The mooring system consists of freely hanging lines connecting the surface platform to anchors, or piles, on the seabed, positioned at some distance from the platform. The mooring lines are laid out, often symmetrically in plan view, around the vessel. Steel-linked chain and wire rope have conventionally been used for mooring floating platforms. Each of the lines forms a catenary shape, relying on an increase or decrease in line tension as it lifts off or settles on the seabed, to produce a restoring force as the surface platform is displaced by the environment. A spread of mooring lines thus generates a nonlinear restoring force to provide the station-keeping function. The force increases with vessel horizontal offset and balances quasi-steady environmental loads on the surface platform. The equivalent restoring stiffness provided by the mooring is generally too small to influence wave frequency motions of the vessel significantly, although excitation by low-frequency drift forces can induce dynamic magnification in the platform horizontal motions and lead to high peak line tensions. The longitudinal and transverse motions of the mooring lines themselves can also influence the vessel response through line dynamics. With the requirement to operate in increasing water depths, the suspended weight of mooring lines becomes a prohibitive factor. In particular, steel chains become less attrac- tive at great water depths. Recently, advances in taut synthetic fibre rope technology have been achieved offering alternatives for deep-water mooring. Mooring systems using taut fibre ropes have been designed and installed to reduce mooring line length, mean- and low-frequency platform offsets, fairlead tension and thus the total mooring cost. To date however, limited experience has been gained in their extended use offshore when compared to the traditional catenary moorings.
- 17. 664 zyxwvutsrqpo Chapter z 8 Mooring system design is a trade-off between making the system compliant enough to avoid excessive forces on the platform, and making it stiff enough to avoid difficulties, such as damage to drilling or production risers, caused by excessiveoffsets. This is relatively easy to achieve for moderate water depths, but becomes more difficult as the water depth increases. There are also difficulties in shallow water. Increasingly integrated mooring/riser system design methods are being used to optimise the system components to ensure lifetime system integrity. In the past, the majority of moorings for FPS were passive systems. However, more recently, moorings are used for station-keeping in conjunction with the thruster dynamic positioning systems. These help to reduce loads in the mooring by turning the vessel when necessary, or reducing quasi-static offsets. Monohulls and semi-submersibles have traditionally been moored with spread catenary systems, the vessel connections being at various locations on the hull. This results in the heading of the vessel being essentially fixed. In some situations this can result in large loads on the mooring system caused by excessive offsets caused by the environment. To overcome this disadvantage, single-point moorings (SPM) have been developed in that the lines attach to the vessel at a single connection point on the vessel longitudinal centre line. The vessel is then free to weathervane and hence reduce environmental loading caused by wind, current and waves. Since the installation of the first SPM in the Arabian Gulf in 1964,a number of these units are now in use. A typical early facility consisted of a buoy that serves as a mooring terminal. It is attached to the sea floor either by catenary lines, taut mooring lines or a rigid column. The vessel is moored to the buoy either by synthetic hawsers or by a rigid A-frame yoke. Turntable and fluid swivels on the buoy allow the vessel to weathervane, reducing the mooring loads. Although the SPM has a number of good design features, the system involves many complex components and is subjected to a number of limitations. More recently, turret mooring systems for monohull floating production and storage vessels (fig. 8.1) have been developed that are considered to be more economic and reliable than SPMs, and are widely used today. The turret can either be external or internal. zyxw An internal turret is generally located in the forepeak structure of the vessel, though a number of turrets have in the past been positioned nearer amidships. Mooring lines connect the turret to the seabed. In order to further reduce the environmental loading on the mooring system from the surface vessel in extreme conditions, disconnectable turret mooring systems have also been developed. Here the connected system is designed to withstand a less harsh ocean envi- ronment, and to be disconnected whenever the sea state becomes too severe such as in typhoon areas. In this section, the fundamentals of mooring systems are covered, the influence of the relevant combinations of environmental loading is discussed and the mooring system design is considered. Also included is information on mooring hardware, including turrets used on weather-vaning floating production systems, model-testing procedures and in certification issues. There are numerous other sources of information on mooring systems, see for example CMPT (1998).
- 18. Mooring Systems 665 z Figure 8.1 Turret moorings. (a) Disconnectibleand (b) Permanent 8.2 Requirements zyxwvuts Functional requirements for the mooring system include: zyxw 1. offset limitations 2. lifetime before replacement 3. installability 4. positioning ability These requirements are determined by the function of the floater. MODUSare held to less restrictive standards than “permanent” mooring systems, referring to production plat- forms. Table 8.1 lists the principal differences in these requirements. 8.3 Fundamentals It is instructive to review the basic mechanics of a mooring line in order to understand its performance characteristics with respect to station-keeping. The traditional wire or chain catenary lines are considered first, followed by taut moorings of synthetic fibre. 8.3.1 Catenary Lines Figure 8.2 shows a catenary mooring line deployed from point A on the submerged hull of a floating vessel to an anchor at B on the seabed. Note that part of the line between A and
- 19. 666 zyxwvutsrqp ; MODU Design for 50-yr return period event Design for 100-yr return period events Anchors may fail in larger events zyxwv Table 8.1 Comparison of typical MODU and FPS mooring requirements Slack moorings in storm events to reduce line tensions zyxwvut Chapter z 8 Moorings are usually not slacked because of risk to risers, and lack of marine operators on board Line dynamics analysis not required Missing line load case not required 1Risers disconnected in storm 1Risers remain connected in storm I Line dynamics analysis required Missing line load case required 1Fatigue analysis not required 1Fatigue analysis required Izy Sea surface -/- / - - Figure 8.2 Catenary mooring line B is resting on the seabed and that the horizontal dimension, a, is usually 5-20 times larger than the vertical dimension, b. As the line mounting point on the vessel is shifted horizon- tally from point zyxwvuts A I ,through A2,A3,A4,the catenary line laying on the seabed varies from a significant length at Al, to none at A4. From a static point of view, the cable tension in the vicinity of points A is due to the total weight in sea water of the suspended line length. The progressive effect of line lift-off from the seabed due to the horizontal vessel movement from A l to A4 increases line tension in the vicinity of points A. This feature, coupled with the simultaneous decrease in line angle to the horizontal, causes the hori- zontal restoring force on the vessel to increase with vessel offset in a non-linear manner.
- 20. Mooring zyxwvutsrqpo Sjstems zyxwvutsr I'zyxwvf 661 Izyxwvutsrqponmlkjihgfed -4 zyxw n zyx Figure 8.3 Cable line with symbols zyxw This behaviour can be described by the catenary equations that can be used to derive line tensions and shape for any single line of a mooring pattern. The equations are developed using a mooring line as shown in fig. 8.3. In the development that follows, a horizontal seabed is assumed and the bending stiffness effects are ignored. The latter is acceptable for wire with small curvatures and generally a good approximation for chain. It is necessary also to ignore line dynamics at this stage. A single line element is shown in fig. 8.4. The term zyxw w represents the constant submerged line weight per unit length, T is line tension, A the cross-sectional area and E the elastic modulus. The mean hydrodynamic forces on the element are given by zy D and F per unit length. Inspecting fig. 8.4 and considering in-line and transverse forcing gives: dT-pgAdz= w s i n 4 - F - ds [ ( 3 1 Ignoring forces F and D together with elasticity allows simplification of the equations, though it is noted that elastic stretch can be very important and needs to be consi- dered when lines become tight or for a large suspended line weight (large 10 or deep waters).
- 21. 668 zyxwvutsrqponm Chapter zy 8 z Figure 8.4 Forces acting on an element of an anchor line zyx With the above assumptions we can obtain the suspended line length s and vertical dimension h as: zyxwvu s = (2) sinh(g) (8.3) (8.4) giving the tension in the line at the top, written in terms of the catenary length s and depth d as: w(s2 +d2) 2d zyxwvu T = The vertical component of line tension at the top end becomes: zyx T z= zyxwv VS (8.6) The horizontal component of tension is constant along the line and is given by: TH= Tcos~$,, (8.7) It is noted that the above analysis assumes that the line is horizontal at the lower end replicating the case where a gravity anchor with no uplift is used. A typical mooring analysis requires summation of the effects of up to 16 or more lines with the surface vessel position co-ordinates near the water plane introducing three further variables. The complexity of this calculation makes it suitable for implementing within computer software.
- 22. Mooring Systems zyxwvutsrq 669 For mooring lines laying partially on the seabed, the analysis is modified using an iteration procedure, so that additional increments of line are progressively laid on the seabed until the suspended line is in equilibrium. Furthermore, in many situations, multi-element lines made up of varying lengths and physical properties are used to increase the line restoring force. Such lines may be analysed in a similar manner, where the analysis is performed on each cable element, and the imbalance in force at the connection points between elements is used to establish displacements through which these points must be moved to obtain equilibrium. The behaviour of the overall system can be assessed in simple terms by performing a static design of the catenary spread. This is described in Section 8.5.2, but it is noted that this ignores the complicating influence of line dynamics that are described in Section 8.4. The analysis is carried out using the fundamental equations derived above. zy 8.3.2 Synthetic Lines For deep-water applications, synthetic fibre lines can have significant advantages over a catenary chain or wire because they are considerably lighter, very flexible and can absorb imposed dynamic motions through extension without causing an excessive dynamic tension. Additional advantages include the fact that there is reduced line length and seabed footprint, as depicted in fig. 8.5, generally reduced mean- and low-frequency platform offsets, lower line tensions at the fairlead and smaller vertical load on the vessel. This reduction in vertical load can be important as it effectively increases the vessel useful payload. The disadvantages in using synthetics are that their material and mechanical properties are more complex and not as well understood as the traditional rope. This leads to over- conservative designs that strip them of some of their advantages. Furthermore, there is little in-service experience of these lines. In marine applications this has led to synthetic ropes subject to dynamic loads being designed with very large factors of safety. Section 8.5.5 discusses the mooring system design using synthetic lines in more detail. Detailed mathematical models for synthetic lines are not developed here, but are z .'.. .......,.___, , , , ,,..(.....,..... --..._._ zyxw Steel Catenary Mooring PolyesterTaut Mooring Figure 8.5 Taut and catenary mooring spread
- 23. 670 zyxwvutsrqpo Chapter z 8 available within the expanding literature on the subject. In particular, these models must deal with: (i) Stiffness zyxwvut - In a taut mooring system the restoring forces in surge, sway and heave are derived primarily from the line stretch. This mechanism of developing restoring forces differs markedly from the conventional steel catenary systems that develop restoring forces primarily through changes in the line catenary shape. This is made possible by the much lower modulus of elasticity of polyester compared to steel. The stretch characteristics of fibre ropes are such that they can extend from 1.2to 20 times as much as steel, reducing induced wave and drift frequency forces. The stiffness of synthetic line ropes is not constant but varies with the load range and the mean load. Further- more the stiffness varies with age, making the analysis of a taut mooring system more cumbersome. Hysteresis and heat build up - The energy induced by cyclic loading is dissipated (hysteresis) in the form of heat. In addition, the chaffing of rope components against each other also produces heat. Cases are known in which the rope has become so hot that the polyester fibres have melted. This effect is of greater concern with larger diameters or with certain lay types because dissipation of the heat to the environment becomes more difficult. Fatigue - The fatigue behaviour of a rope at its termination is not good. In a termination, the rope is twisted (spliced) or compressed in the radial direction (barrel and spike or resin socket). The main reason for this decreased fatigue life is local axial compression. Although the rope as a whole is under tension, some components may go into compression, resulting in buckling and damage of the fibres. In a slack line this mechanism is more likely to be a problem than in a rope under tension. The phenomenon can appear at any position along the rope. Other relevant issues to consider are that the strength of a polyester rope is about half that of a steel wire rope of equal diameter. Additionally the creep behaviour is good but not negligible (about 1.5% elongation over twenty years). Furthermore, synthetic fibre ropes are sensitive to cutting by sharp objects and there have been reports of damage by fish bite. A number of rope types such as high modulus polyethylene (HMPE) are buoyant in sea water; other types weigh up to 10% of a steel wire rope of equal strength. Synthetic fibre lines used within taut moorings require the use of anchors that are designed to allow uplift at the seabed. These include suction anchors, discussed further in Section 8.6. (ii) (iii) (iv) zyxwvuts 8.3.3 Single Catenary Line Performance Characteristics Figures 8.6a and b present the restoring force characteristics of a single catenary line plotted against offset (non-dimensionalised by water depth) for variations respectively in line weight and initial tension. Both figures emphasise the hardening spring character- istics of the mooring with increasing offset as discussed above. While this is a specific example, several observations may be made regarding design of a catenary system from these results.
- 24. Mooring Systems zyxwvutsrq 671 z 3io zyxwvutsrq HlTUL TLNEIOH. zyxwv LN zyxwv / z 1 zy 0 2 b b 1 1 0 OFFSET ~ X WATER PEPTH (a) Effect of changing line weight -- initial tension = 135 kN (b) Effect of changing initial tension -- weight = 450 kg/m Figure 8.6 Restoring force for a single catenary line (depth = 150 m) Figure 8.6a shows the effect of line weight for a single line in 150 m of water with 135 kN initial tension. Under these conditions, the mooring would be too hard with lines weighing 150 kg/m. A 300 kg/m system is still too hard, but could be softened by adding chain. Additional calculations would be required to determine the precise quantity. The 450 kg/m line appears acceptable with heavier lines being too soft at this water depth and initial tension. The softness can be reduced by increasing the initial tension in a given line for the specified water depth. Figure 8.6b shows that latitude exists in this particular system. The choice of initial tension will be determined by the restoring force required. The hardness of a mooring system also decreases with water depth, assuming constant values for other properties. 8.4 Loading Mechanisms There are various loading mechanisms acting on a moored floating vessel as depicted in fig. 8.7. For a specific weather condition, the excitation forces caused by current are usually assumed temporally constant, with spatial variation depending on the current profile and direction with depth. Wind loading is often taken as constant, at least, in initial design calculations, though gusting can produce slowly varying responses. Wave forces result in time-varying vessel motions in the six rigid body degrees of freedom of surge, sway, heave, roll, pitch and yaw. Wind gust forces can contribute to some of these motions as well.
- 25. 612 zyxwvutsrqp Top zyxwvu end surgemotion zyxwvu 4 Chapter z 8 steadywind with randomfluctuatiom zyx ! ! ! ! I ! ! ! waves and wave drift ~....."...........-.............~."..~.."...II."~......."...__..I...,.,_..(..I.." ......."..." ..........,........... "_.." ._..... ".".".".." ..._._.__ seabed -fiction NB:Environmentalzyxwvuts forces u enot nacrrrarily co-lieu Figure 8.7 Environmental forces acting on a moored vessel in head conditions and transverse motion of catenary mooring lines Relevant FPS responses are associated with first-order motions at wave frequencies. together with drift motions at low frequencies (wave difference frequencies). In particular, motions in the horizontal plane can cause high mooring line loads. This is because the frequency of the drift forces results in translations that usually correspond to the natural frequency of the vessel restrained by the mooring system. Consequently, it is essential to quantify the level of damping in the system, as this quantity controls the resonant motion amplitude. Wave period is of great importance and generally the shortest wave period that can occur for a given significant wave height will produce the highest drift forces at that wave height. Furthermore, on ship-shaped bodies, the forces are greatly increased if the vessel is not head on to the waves. This situation will occur if the wind and waves are not in line and the vessel has a single point mooring. For example, on a 120,000 ton DWT vessel the wave drift forces will be doubled for a vessel heading of approximately 20" to the wave direction, when compared to the forces on the vessel heading directly into the waves. There are a number of contributions to damping forces on a floating vessel and the moorings. These include vessel wind damping caused by the frictional drag between fluid (air) and the vessel, though the effect can be small. This has a steady component allowing linearisation procedures to be used to obtain the damping coefficient. Current in conjunc- tion with the slowly varying motion of the vessel provides a viscous flow damping contri- bution because of the relative motion between the hull and the fluid. This gives rise to lift and drag forces. Both viscous drag and eddy-making forces contribute. The magnitude of the damping increases with large wave height. Wave drift damping on the vessel hull is associated with changes in drift force magnitude caused by the vessel drift velocity. The current velocity is often regarded as the structure slow drift velocity. It can be shown that when a vessel is moving slowly towards the waves, the mean drift force will be larger than
- 26. Mooring Sysiems zyxwvutsrq 613 z top zy end surge zyx motion zy I *P z Figure 8.8 Catenary line motions caused by vessel horizontal translation when it is moving with the waves. The associated energy loss can be thought of as slow drift motion damping. There are a number of contributions to the overall damping from the mooring system. These are: Hydrodynamic drag damping - depending on the water depth, line pre-tension, weight and azimuth angle, a relatively small horizontal translation of the vessel can result in transverse motion over the centre section of the line that can be several times larger than the vessel translation itself as indicated in fig. 8.8. The corresponding transverse drag force represents energy dissipation per oscillation cycle and thus can be used to quantify the line damping. Brown and Mavrakos (1999) quantified levels of line damping for variations in line oscillation amplitude and frequency. Webster (1995) provided a comprehensive parametric study quantifying the influence of line pre- tension, oscillation amplitude and frequency and scope (ratio of mooring length to water depth) on the line damping. Vortex-induced vibration - vortex formation behind bluff bodies placed in a flow gives rise to unsteady forces at a frequency close to the Strouhal frequency. The forces cause line resonant response in a transverse direction to the flow and the vortex formation can become synchronised along the length resulting in the shedding frequency “locking in” to the line natural frequency [Vandiver, 19881.This can give a significant increase to the in-line drag forces. It is generally considered that this effect is important for wire lines, whereas for chains it is assumed negligible. Line internal damping - material damping caused by frictional forces between individual wires or chain links also contributes to the total damping. Only limited work has been performed in this area. Damping caused by seabed interaction ~ soil friction leads to reduced tension fluctuations in the ground portion of line effectively increasing the line stiffness. Work by Thomas and Hearn (1994) has shown that out-of-plane friction and suction effects are negligible in deep-water mooring situations, whereas in-plane effects can significantly influence the peak tension values.
- 27. 674 zyxwvutsrq 1 zyxwvutsrqponm #U zyxwvutsrq 0 . 9zyxwvut - 0.8- 0.7 - zyxwv 0.6- 0.5 - 0.4 - 0.3- 0.2- 0.1 - zyxwvu Chapter 8 z (m) 8.6 16.3 (SI Mooring Waves Viscous 12.7 81 15 4 16.9 84 12 4 ----- wave drift damping ., ,............ -. viscousdamping -mooring line damping zyxwv 0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.W 0.10 Surge arnplitude/waterdepth Figure 8.9 Relative energy dissipation caused by surge damping contributions Table 8.2 Relative % damping contributionsfor a 120,000ton DWT tanker in 200 m water Significant Peak period Damping contribution YO wave height The levels of mooring line damping relative to other contributions can, in some situa- tions, be very high. See, e.g. fig. 8.9 showing energy dissipation as a result of wave drift, hull viscous damping and mooring line damping [Matsumoto, 19911 for a catenary mooring spread restraining a model tanker in 200 m water depth. The increased line damping for higher motion amplitudes is caused by the large transverse motion of the catenary lines. Table 8.2 from Huse and Matsumoto (1989) gives measured results for a similar vessel undergoing combined wave and drift motion. Here, damping from the mooring system provides over 80% of the total with viscous and wave drift giving limited contributions in moderate and high seas. The line damping work is extended in Huse (1991).
- 28. Mooring Systems 615 z 8.5 Mooring System Design zyxwvu In this section the range of available design methods for catenary moorings is considered. Their use with synthetic taut moorings is also outlined. The methods should be read in conjunction with the certification standards outlined in Section 8.7. There then follows some considerations associated with effective water depth, an outline of mooring spreads and a discussion of some uncertainties associated with the design procedures and their input data. 8.5.1 Static Design This is often carried out at the very initial stages of the mooring system concept design and is described for a catenary system. Load/excursion characteristics for a single line and a mooring spread are established ignoring fluid forces on the lines. The analysis is carried out by utilising the algorithms described in Section 8.3.1 to calculate the forces exerted on the vessel from each catenary line, given the line end-point coordi- nates on the surface vessel and seabed together with lengths and elasticity. These forces are then summed for all lines in the mooring spread to yield the resultant horizontal restoring and vertical forces. The restoring force and tension in the most loaded line is then calculated by displacing the vessel through prescribed horizontal distances in each direction from its initial position. The results of a typical analysis are presented in fig. 8.10. The steady component of environmental force from wind, current and wave drift effects is applied to the vertical axis of this diagram to obtain the resultant static component of vessel offset from the horizontal axis. The slope of the force curve at this offset gives an equivalent linear stiffness zy C, of the mooring system in the relevant direction for use in an equation of the form: c,x = FJt) (8.8) where co-ordinate x refers to a horizontal degree of freedom (surge or sway), F, is force, and the stiffness resulting from the vessel hydrostatics is zero. The maximum dynamic offset caused by the wave and drift frequency effects is then estimated. Certifying authority standards give guidance on this. It is necessary to check that line lying on the seabed has no upward component of force at the anchor. If there is insufficient line length, the calculations should be repeated with increased length. The load in the most heavily loaded line is then read off and compared with a pre-set fraction of the breaking strength of the line. If the fraction is too high, it is necessary to adjust the line pre-tension, change material specification for each line, alter the line end co-ordinates or number of lines and repeat the calculations. Once the intact system has been established, the calculations should be performed for the case where the most loaded line is broken and similar checks carried out. The method has the disadvantages that conservative assumptions are made in terms of the uni-directional environment and large safety factors need to be applied to account for uncertainties. Furthermore important features of the dynamics are absent from the methodology.
- 29. 676 zyxwvutsrqpo Chapter 8 z 2000 zyxwvu 1500 5 zyxw 51000 E zyxwvu 0 U 500 0 zyxwvuts ' / Restoring zyxwv force , / Maximum line tension - + l o 15 2 0 I 5 Stat,c zyxwv offset Vessel excursion (m) Figure 8.10 Restoring force and most loaded line tension against vessel excursion for a catenary mooring system (static analysis) 8.5.2 Quasi-Static Design zyxwvu This procedure is the next level of complexity; generally, one of the two types of calculations are carried out: 0 A time-domain simulation that allows for the wave-induced vessel forces and responses at wave and drift frequency, while treating wind and current forces as being steady and using the mooring stiffness curve without considering line dynamics. A frequency response method where the mooring stiffness curve is treated as linear and low-frequency dynamic responses to both wave drift and wind gust effects are calculated as if for a linear single degree of freedom system. The basic differences between the static and quasi-static design are that: 0 the quasi-static analysis is usually non-linear in that the catenary stiffness at each horizontal offset is used within the equations of motion. Note that a stiff catenary or taut mooring may have essentially linear stiffness characteristics; the equations of motion are integrated in the time domain. The influence of, at least, some added mass and damping contributions are included, although these tend to be associated with the vessel rather than accurate values including the influence from the mooring system; 0
- 30. Mooring Systems zyxwvutsrq 611 frequency domain solutions are possible but gross assumptions associated with linear- isation of stiffness and damping need to be made. The analysis solves the equation: (m zyxwvut +A)X + zyxw Bx +B , x l x l+ zyxw C,x = FJt) (8.9) in each degree of freedom to give the motions, zyxw x.Coupling between the motions can also be included. The terms m, A, B and B, refer to vessel mass, added mass, linear and viscous damping respectively with F, representing the time varying external forcing. To give reliable answers, the simulation must cover a minimum of 18 h full-scale behaviour in order to provide sufficient statistical data for the low-frequency responses. zy 8.5.3zyxwvutsr Dynamic Design Full dynamic analysis methods are regularly utilised in design, though there is no universal agreement in the values of mooring line damping. This can influence vessel responses and line loads strongly, particularly in deep water. In outline terms, the methodology is as follows: Usually a static configuration must first be established with non-linear time domain solu- tions developed about this initial shape. Often the line is de-composed into a number of straight elements (bars) with linear shape function except for the distributed mass plus added mass that is lumped at the end nodes. Generally, the motions of the platform are calculated independently of the estimates of line dynamics. However for deep-water moorings, the importance of mutual interactions between the mooring lines and the moored platform has been recognised and coupled platform mooring analysis methods need to be used. In this case, the effect of line dynamics on the platform motion is mutually included in a time-domain solution. Importantly, dynamic methods include the additional loads from the mooring system other than restoring forces, specifically the hydrodynamic damping effects caused by relative motion between the line and fluid. Inertial effects between the line and fluid are also included though the influence is often small. Simulations use lumped mass finite element or finite difference schemes to model small segments of each line whose shape is altered from the static catenary profile by the water resistance. Analysis is performed in the time domain and is computationally intensive. Difficulties are: time steps must be small so that wave-induced line oscillations are included, runs must be long to allow for the vessel drift oscillation period, which in deep water may be of the order of 5 min, for a typical floating vessel mooring system design, the weather is multi-directional and a number of test cases must be considered. Line top-end oscillation must be included, because of vessel motion at combined wave and drift frequencies; otherwise, dynamic tension components may be underestimated, or
- 31. 678 zyxwvutsrqpo Chapter z 8 advantages of line damping contributions neglected. It is noted that line dynamics can, in some cases, result in the doubling of top tension when compared to the static line tension. Furthermore, damping levels vary significantly depending on water depth, line make up, offsets and top-end excitation. Hybrid methods that work in the time domain but make a number of simplistic assumptions about the instantaneous line shape are currently being investigated. There is some potential here, but further work is needed to provide methods usable in the design. More efficient frequency domain methods are also being developed that include line dynamics in an approximate manner. zyxwv At present these do not work well when strong non- linearities, such as those caused by fluid drag forces are present, for example, when large line oscillations occur. Figures 8.11-8.13 show results from a design study for a turret-moored monohull vessel positioned at a northerly North Sea location. Figure 8.11 depicts the drift force energy spectra for the vessel in head seas with 1 and 100-yr return period weather. The energy spectra are very broad banded, providing excitation over a wide frequency range that includes, as is usually the case, the resonant surge frequency of the vessel on its mooring system. zyxwvu XtOb zyxwvutsrqpon 0.0000 0.0600 0.1600 0.2400 0.3200 0 zy Frequency(Hz) Figure 8.11 Mooring line analysis - head sea drift spectra 00
- 32. Mooring zyxwvutsrqp Systems 619 z x103 15.00 zyxwvutsrq 000 zyxwvutsrq - I zyxwvuts I / I l l I I I 6W 6300 zyxwvuts 6.600 6900 7200 7500 zyxwvu 7800 8.100 8400 8700 9. ~103 Time (see) Figure 8.12 Mooring line analysis - Line tension vs. time (intact) 6000 6300 6600 6900 7200 7500 7800 8100 8400 8700 9000 xi03 Time (sec) Figure 8.13 Mooring line analysis - Line tension vs. time (transient motion)
- 33. 680 zyxwvutsrqpo Chapter z 8 Figures 8.12 and 8.13 give the line tension graphs for the intact mooring and transient conditions after line breakage for 1 yr storm conditions. Low amplitude wave and high amplitude drift effects can clearly be seen. zyxw 8.5.4 zyxwvutsr Synthetic Lines Essentially, the design procedures for taut moorings are similar to those described for catenary systems with the exception that three stiffness values are used in the design calculations: Bedding-in stiffness zyxwvu - This is the initial elongation after manufacture and is as a result of fibre extension, which may be partially recovered in some circumstances unless the load is maintained. It is also partly due to a tightening of the rope structure, which is retained unless the rope suffers a major buckling disturbance. The bedding-in elonga- tion becomes negligible after approximately one hundred cycles up to a given load. The response after installation, when the rope has been subjected to a certain load cycling regime, is given by the post-installation stiffness. A minimum estimated value of instal- lation stiffness should be used to calculate offsets in the period after installation. Drift stiffness - Cyclic loading under moderate weather conditions, applicable to the mooring during a high proportion of the time, shows a mean variation of tension and elongation which is represented by the drift stiffness. A minimum estimated value of drift stiffness should be used to calculate offsets under normal mooring conditions. Storm stiffness - Under more extreme conditions, the mean variation of tension and elongation is represented by the storm stiffness, which is higher than the drift stiffness. A maximum estimated value of storm stiffness should be used to calculate peak load. Creep with time may also occur, and analyses need to consider this, with re-tensioning at site required throughout the installation lifetime. Calculations must also be performed to assess hysteresis effects inherent in the fibre properties and caused by friction. This will generate heat. zyx 8.5.5 Effective Water Depth Combinations of tide change plus storm surge, for example, together with alterations in vessel draught, because of ballasting, storage and offloading etc. result in changes in the elevation of the vessel fairleads above the seabed. The example given in fig. 8.14 presents the range of elevation levels for a 120,000 ton dwt floating production unit in a nominal water depth of 136m. This elevation range is likely to be relatively larger in shallow water. LAT represents lowest astronomical tide. A number of elevations must be considered in the mooring design to establish the resulting influence on line tension. 8.5.6 Mooring Spreads Although a symmetric spread of mooring lines is the simplest in terms of design, it may not be the optimum in terms of performance. Criteria needing considerations are: directionality of the weather; in particular if storms approach from a specific weather window, it may be advantageous to bias the mooring towards balancing these forces,
- 34. Mooring Sjstems 68 z 1 z Site conditions: water depth at site, to LAT maximum depth of fairleads below WL (loaded) minimum depth of fairleads below WL (ballasted) maximum tide +tidal surge above LAT minimum tide +tidal surge below LAT Maximum vessel fairlead elevation is: water depth minimum depth at fairleads (ballasted) maximum tide +tidal surge fairlead elevation Minimum vessel fairlead elevation is: water depth maximum depth to fairlead (loaded) minimum tide + tidal surge fairlead elevation 136m 16m 8m 2.5m 0.5m 136m -8m +2.5m 130.5m 136m -16m -0.5m 119.5m zy Mean elevation is thus 125m. Figure 8.14 Effective water depth and fairlead position range subsea spatial layout; seabed equipment and pipelines may restrict the positioning of lines and anchors in this region, riser systems; clashing of risers with mooring lines must be avoided and this may impose limitations on line positions, space restrictions in the turret region; it may be beneficial to cluster lines together to gain further space. Figure 8.15 gives an example of a symmetric spread, while fig. 8.16 depicts an alternative arrangement having wide corridors to accommodate a large number of flexible risers for an extensive offshore development. 8.5.7 Uncertainty in Line Hydrodynamic Coefficients There are many uncertainties associated with mooring system design. These include the uncertainties in input data, the environment, its loading on the vessel and mooring system together with the response, seabed conditions and line physical properties. Because of the large number of “fast track” projects, research and development work cannot keep pace and consequently, mooring systems are less cost-effective. requiring higher safety factors or, in some cases, lower reliability. A specific uncertainty is associated with the choice of chain line drag coefficient, required in the design in order to calculate the maximum line tensions including dynamic effects. Furthermore, line drag is the major contribution towards induced mooring damping as discussed earlier. Figure 8.17 provides drag coefficients plotted against Re for harmonic, sinusoidal oscillations taken from Brown, et a1 (1997). Various Keulegan-Carpenter (KC) values
- 35. 682 zyxwvutsrqpo Chapter z 8 Shuttle zyx tanker Figure 8.15 Plan view of symmetric spread Figure 8.16 Riser corridors between non-symmetric spread
- 36. Mooring zyxwvutsrqpo Sjstems zyxwvutsrqp 683 z z zyxwv 1 0 zyxwv X zyxw NTNF zyxwvu I 1991) data l.OOE+OS l a 1.00E+03 1.00E+04 R e Figure 8.17 Measured drag coefficient for chain in harmonic flow conditions KC number e490 += between 70 and 582 are indicated, and results are for the large-scale stud chain samples. Also plotted are the results from NTNF (1991). These data are based primarily on results from a number of tests with small-scale specimens, cross-flow conditions or harmonic oscillations. It is noted that a drag coefficient of 2.6 for chain without marine growth is commonly used in design, whereas 2.4 is common for studless chain. Mooring lines undergo bi-harmonic motions caused by the combined wave and drift floater response. It is known, however, that simply superimposing the wave and drift effects gives erroneous results. The calculation of drag coefficient for harmonically oscillating flow past a body is based on the drag force term of the Morison equation. When there is bi-harmonic flow (Le. two frequencies of oscillation), the situation is not so simple. In resolving the measured force into drag and inertia components, it is possible to define two drag (and inertia) coefficients, appropriate to either of the two frequencies of oscillation. An additional complication arises as either the wave or drift maximum velocity, or indeed the sum of the two may be used within the Morison formulation. Furthermore, alternative Reynolds numbers and KC values may also be established based on the appropriate oscillation frequency and amplitude. Figure 8.18 examines the variation of in-line drag coefficient under bi-harmonic oscillation conditions with wave oscillations in various directions to drift motion. C, is plotted against wave frequency oscillation direction relative to the drift frequency and direction. Drag coefficients are based on the drift frequency of oscillation as the damping contribution to the drift motion of the vessel is of interest. Velocities used to calculate the drag coefficient are based on the combined wave and drift oscillations. The results show a significant increase in drag for the situation with wave oscillations in the transverse direction to the drift when compared to the in-line wave oscillations. In a sense this can be thought of as a drag amplification effect somewhat similar to that induced by
- 37. 684 zyxwvutsrqpo Chapter 8 z 3.7 3 . 6 3.5 3 zyxwv 3.1- 3.1 3 0 15 30 zyxwv 45 zyxwv 60 75 w Wave oscillatlondirection (deg) Figure 8.18 Measured in-tine drag coefficients for chain in bi-harmonie flow zy vortex-induced vibrations, though here the out-of-plane vibrations are caused by top-end motion in the transverse direction, as opposed to flow-induced loading. Curves are plotted for wave to drift motion amplitude ratios (Aw/Ad) of 0.27 and wave to drift motion frequency ratios (fw/fd) from 4.4 to 13.2. In a realistic sea state, a mooring line will be subjected to motions at wave frequencies both in in-line and transverse directions to the imposed drift motions. Consequently, in order to use the present results in design it is necessary to interpret the vessel surge, sway and yaw motions at wave frequencies to establish the relevant translation angle of the fairlead in the horizontal plane relative to the drift motion. This can then be used in conjunction with the drag coefficient values interpolated from fig. 8.18. It is also necessary to estimate the ratios of wave to drift motion amplitude and wave to drift motion frequency of oscillation. A simple method to establish the latter could be to use the zero-crossing period of the sea state relative to the drift period. Linear and higher-order potential flow analysis methods or model test data can be used to estimate amplitude ratios. In the absence of more refined data, fig. 8.18 provides appropriate results of in-line drag coefficient for use in design. 8.5.8 Uncertainty in Line Damping and Tension Prediction Work initiated by the International Ship and Offshore Structures Congress (ISSC), Committee 1.2 (loads) presents a comparative study on the dynamic analysis of suspended wire and stud chain mooring lines [Brown and Mavrakos, 19991.A total of 15contributions to the study were provided giving analysis results based on dynamic time or frequency domain methods for a single chain mooring line suspended in 82.5 m water depth and a wire line in 500 m depth. Bi-harmonic top-end oscillations representing in-line combined wave- and drift-induced excitation were specified.
- 38. Mooring Systems 685 z 1 2 0 zyxwvuts $10 1 zyxw ---C-T=lOOs Mean zy (M) -+-T=100s zyxw M-S - 4 - - T = 1 0 0 s M + S I --x- -T=100s Expt zyx 9 B O B E, 4 0 m 6 0 a 0 E - 2 0 0 zyxwvut 0 5 10 1 5 2 0 Drift Amp. (m) Figure 8.19 Chain line damping vs. drift induced top-end amplitude (drift period = 100 s) - no wave oscillation, water depth = 82.5 m zyxw The mooring line damping results for chain are compared with the limited available experimental data. The results provided by the participants show a fair agreement despite the complexity of the numerical methods. Predictions of dynamic tension based on time-domain methods show scatter, the estimates of damping giving further discrepancies. Some results were based on frequency-domain methods for which there are even more disagreement. The uncertainty in results is quantified by plotting the mean, mean plus/minus one stan- dard deviation (M+S, M-S) of tension and line damping from the various data provided by contributors. Clear trends in tension and damping with oscillation frequency and amplitude are also revealed. Calculated line damping values are plotted against drift-induced oscillation amplitude for the chain in 82.5 m water depth in fig. 8.19. Here there is no oscillation at wave frequencies. The results indicate that increasing the drift top-end amplitude from 10 to 20 m causes an increase in damping by a factor of approximately 4.5. It is noted that doubling the oscillation period caused the damping to reduce by 50%. Similar trends with drift-induced amplitude were observed for the wire in 500 m water depth. Figures 8.20 and 8.21give dynamic tension components (total tension minus static catenary tension) for the chain (with drift amplitude and period of 10 m and 100 s respectively) and wire (with drift amplitude and period of 30 m and 330 s respectively). It is seen that a number of contributions with the wire results predict total tensions less than the catenary value. A possible reason for this is that the calculation method for catenary tension does not include stretch of the seabed portion and thus may give slightly conservative values. Contributor data may allow stretch of this grounded portion. There is a consistent trend throughout these results in that both the dynamic tension and the mooring line damping increase significantly as the line wave-induced top-end motion increases. There is also large uncertainty in the results; for example, contributor responses given in fig. 8.20 indicate a
- 39. 686 zyxwvutsrqpo Chapter z 8 z 5 0 0 0 zyxwvuts 4 zyxwvutsr 500 4 0 0 0 zyxwvuts 3500 3000 2500 2000 1500 1000 500 0 zyxwvuts - l f = l O s Mean (M) zy -4 - -TI 10s M+S - - f - T = l O S M-S X Tz13s Mean (M) 0 T ~ 1 3 s M+S X T=l& M-S 0 1 2 3 4 5 6 7 8 W s w Amp. (m) Figure 8.20 Chain maximum dynamic tension vs. wave-induced top-end amplitude - with drift oscillation, water depth = 82.5 m 2000 1500 1000 500 0 -500 0 1 2 3 4 5 6 7 a Wave Amp. (m) +T=tOs Mean tk’ - f - f : l O S M-S - 4 - - f = l O ~ M+S X T.13~ Mean (M X fz13s M*S 0 T-13s M+S Figure 8.21 Wire maximum dynamic tension vs. wave induced top end amplitude - with drift oscillation, water depth = 500 m line tension standard deviation at 8 m wave amplitude of over 600 kN about a mean of 4000 kN. The catenary (static) tension not plotted here is approximately 3500 kN. More recently, a number of studies have developed efficient numerical and analytical solution techniques for the evaluation of mooring line dynamics. Aranha and Pinto (2001b)
- 40. Mooring zyxwvutsrq Sjstems zyxwvutsrqpon 687 derived an analytical expression for the dynamic tension variation along the cable’s suspended length, whereas Aranha, et a1 (2001a) followed the same methodology to obtain an analytical expression for the probability density function of the dynamic tension envelope in risers and mooring lines. Gobat and Grosenbaugh (2001a) proposed an empirical model to establish the mooring line dynamic tension caused by its upper end vertical motions. Aranha, et a1(2001a) introduced a time integration of the cable dynamics equations. Chatjigeorgiou and Mavrakos (2000) presented results for the numerical prediction of mooring dynamics, utilising a pseudo-spectral technique and an implicit finite difference formulation. zyxwvu 8.6 Mooring Hardware Components The principle components of a mooring system may consist of Chain, wire or rope or their combination Anchors or piles Fairleads, bending shoes or padeyes Winches, chain jacks or windlasses Power supplies Rigging (e.g. stoppers, blocks, shackles) zyxw 8 . 6 . 1 Chain Chain and wire make up the strength members for the mooring system. There are two primary chain constructions. Stud-link chain (fig. 8.22a) has historically been used for mooring MODUS and FPSOs in relatively shallow water. It has proven strong, reliable and relatively easy to handle. The studs provide stability to the link and facilitate laying down of the chain while handling. Figure 8.22 (a) Stud-link and (b) Studless chain
- 41. 688 zyxwvutsrqpo Chapter z 8 Permanent moorings have recently preferred to use open link, or studless chain (fig. 8.22b). Removing the stud reduces the weight per unit of strength and increases the chain fatigue life, at the expense of making the chain less convenient to handle. Chain size is specified as the nominal diameter of the link, “D” in figs. 8.22a and b.’ The largest mooring chain manufactured to date is the 6.25 in. (159 mm) studless chain for the Schiehallion FPSO in the North Atlantic (West of Shetlands). The specification of chain properties is an important function in any mooring system design. The chain is sold in a variety of grades. Grade 4 zyxw (K4) is the highest grade chain currently available. Drilling contractors have traditionally used the oil rig quality (ORQ) chain, which has detailed specifications in API Specification 2F.2Properties of these chains are presented here. 8.6.2 Wire Rope Wire rope consists of individual wires wound in a helical pattern to form a “strand”. The pitch of the helix determines the flexibility and axial stiffness of the strand. Wire rope used for mooring can be multi-strand or single-strand construction. The princi- ple types used offshore are shown in fig. 8.23. Studlink chain and six-strand wire rope are the most common mooring components for MODUS and other “temporary” moorings. Multi-strand ropes are favoured for these applications because of their ease of handling. Six-strand rope is the most common type of multi-strand rope used offshore. Mooring line ropes typically consist of 12, 24, 37 or more wires per strand. The wires have staggered sizes to achieve higher strength. Common “classes” of multi-strand rope include [Myers, 19691:zyxw 0 6 zyxwvuts x 7 Class: Seven wires per strand, usually used for standing rigging. Poor flexibility and fatigue life, excellent abrasion resistance. Minimum drum diameter/rope diameter (Did) = 42. Figure 8.23 Wire rope construction ’Note that mooring design guidelines require that the chain be oversized to allow for corrosion *API. “Specification of Mooring Chain”, 2F (latest edition).
- 42. Mooring Sjstems zyxwvutsrq 689 6 x 19 Class: 1627 wires per strand. Good flexibility and fatigue life and abrasion resistance. Common in lifting and dredging. Minimum D/d = 26-33. 6 x 37 Class: 27-49 wires per strand. Excellent fatigue life and flexibility, but poor abrasion resistance. Minimum D/d = 16-26. Multi-strand wire ropes may contain either a fibre or a metallic core. The core is important for support of the outer wires, especially on a drum, and in some applications to absorb shock loading. Fibre core (FC) ropes are not generally used for heavy duty marine applications. Metallic core ropes may be one of the two types: independent wire rope core (IWRC) or wire-strand core (WSC). IWRC is the most common core filling for heavy marine applications. Single-strand ropes are more common in large permanent installations. The wires are wound as a helix with each layer wrapped in a different direction. This provides “torque balancing”, preventing the rope from twisting when under load. The spiral strand is more fatigue resistant than the multi-strand rope. Corrosion resistance is enhanced by either sheathing with a polyurethane coating, adding zinc filler wires or using galvanised wires. Sheathing provides the best performance, provided that the handling procedures insure against damage to the sheath. zyxwvu 8.6.3 Properties of Chain and Wire Rope Tables 8.3 and 8.4 are taken from the Det Norske Veritas OS-E301 and show the mechanical properties of common grades of mooring chain in which zyxw d is the nominal diameter in mm. Tables 8.5 and 8.6 show the mechanical properties of the most common types of mooring chain and wire in English units. The quantity “d” is the nominal diameter in inches. The rope and chain properties are constantly being improved. Latest values should be obtained from the manufacturers. 8.6.4 Moorings Figure 8.24 gives a typical line leg for a catenary moored floating production unit in 140m water depth. Lower and upper terminations are of chain to avoid seabed wear and excessive bending associated with handling. In a number of moors, one shot (27.5 m) of chain is used at the line top-end and a spiral wound wire over the centre section that does not contact the seabed. 8.6.5 Connectors Connectors are used to join sections of chain to one another, connecting chain to wire rope, connecting to padeyes on anchors or vessels, etc. The common types of connectors for the stud link chain and studless chain are given in DNV OS-E301. Mooring connectors are designed to take the full breaking strength of the chain or wire rope, but their fatigue properties require special attention. There is very little fatigue data for the standard connec- tors and their use is therefore not recommended for permanent moorings. Links used in permanent moorings should be special purpose designs. An example of a triplate is shown in fig. 8.25. Next Page
- 43. 690 zyxwvutsrqpon Grade NV R3 NV R3S NV R4 zyxwvuts Chapter 8 Minimum yield zyxwvu 1Minimum tensile Minimum Minimum reduction strength strength elongation (%) of area (%) (N/mm 1 410 690 17 50 490 770 15 50 580 860 12 50 ~ ~N/mm2) Grade NVR3 Minimum Charpy V-notch energy zyxw (J) Temperature’ Average Single zyx (“C) Base Weld Base Weld 0 60 50 45 38 -20 40 30 30 23 NVR3S 0 65 53 49 40 -20 40 33 ~ 34 25 zy Table 8.4 Formulas for proof and break test loads (adopted from DNV OS-E301) NV R4 zyxwvut 1 0 -20 70 56 53 42 50 36 38 27 1Stud Chain Links 1NV R3S 10.0180d2(44-0.084 10.0249d2(44-0.084 I Type of chain 1Grade Stud Chain Links NV R3 1Stud Chain Links /NV R4 I0.0216d2(44-0.08d) 10.0274d2(44-0.08d) 1 Proof test Break test load, kN load, kN 0.0156d2(44-0.08d) 0.0223d2(44-0.08d) Studless Chain Links INV R3 Studless Chain Links NV R3S Studless Chain Links NV R4 These links are typically engineered and tested as “fit for purpose” designs for each project. Cable terminations consist of a socket, which is a cast in-place to achieve a strength equivalent of the wire rope. The connecting socket may be either “closed” or “open”, see fig. 8.26. 0.0156d2(44-0.08d) 0.0223d2(44-0.08d) 0.0174d2(44-0.08d) 0.0249d2(44-0.08d) 0.0192d2(44-0.08d) 0.0274d2(44-0.08d) Previous Page
- 44. Mooring Systems zyxwvutsrq Steel area (in,2) zyxwvuts 69 z 1 K 4 Studless chain Spiral strand 6 Strand IWRC 2.64d2 0.58d2 0.54d2 zy Table 8.5 Properties of mooring chain and wire rope Breaking strength (kip) Stiffness (kip) 3.977d2(44-2.0324 126d2 93.2d2 i 10,827d2 13,340d 8640d2 IWeight in water (lbift) I7.83d2 1 1.74d2 1 1.59d2 i 3.75 1 110.1 4 I 125.3 2035 152,250 24.5 1772 1 187,5941 22.4 1311 121,500 2283 173,226 27.8 2016 1213,4401 25.4 1491 138,240 Table 8.6 Tabulated mooring component data 5 5.25 5.5 5.75 1 1K4 studless chain 1Suiral strand 1IWRC wire roue i 195.8 3366 270,666 43.5 3150 333,500 39.8 2330 216,000 215.8 3655 298,409 48.0 3473 367,684 43.8 2569 238,140 236.9 3950 327,506 52.6 3812 403,535 48.1 2819 261,3601 258.9 4251 357,956 57.5 4166 441,054 52.6 3081 285,660 strength strength strength 6.25 6.5 6.75 305.9 4864 422,916 68.0 I 4922 521,094 62.1 3641 337,500 330.8 5176 457,426 73.5 5324 563,615 67.2 938 365,040 356.8 5490 493,289 79.3 5741 607,804 72.4 4246 393,660 14.5 I 158.6 I 2808 1219,2391 35.2 1 2552 1270,135 1 32.2 1 1887 1 174,9601 14.75 i 176.7 i 3083 1244.2761 39.3 I 2843 i 300.9841 35.9 I 2103 I 194.9401 16 1 281.9 ! 4556 1389,759 1 62.6 1 4536 1480,240 1 57.2 1 3355 1311,0401 17 1 383.7 1 5805 1530,505 1 85.3 1 6174 1653,660 1 77.9 1 4567 1423,3601
- 45. 692 zyxwvutsrqp 04 zyxwvut 03 zyxwvut 02 01 zyxwvuts Chapter 8 5" 114 Anchor Chain L = 320m zyx 5" zyxwvutsr 1/2 Anchorchain L = 300m L = zyxw OOmzyxw 6" Anchor Chain L = 27.5m Spiral Strand Rope Figure 8.24 Typical mooring line components (shackles not shown) Link Side Socket Side Figure 8.25 Triplates (DNV OS-E301)
- 46. Mooring zyxwvutsrqpo Systems zyxwvutsrq 693 z Figure 8.26 Wire rope sockets (DNV OS-E301) Figure 8.27 Examples of chain fairleads 8.6.6 Shipboard Equipment Shipboard equipment depends on the type of line (wire rope or chain) connected to the vessel, and whether the mooring is used for positioning or is static. For example, the chain jacking system may be placed on top of a column for a semi-submersible or placed on the platform for an FPSO. A typical fairlead for a chain at the platform end is shown in fig. 8.27. On the left hand-side, a bending shoe-type fairlead is depicted. On the right hand-side the chain is fed through a rotary sheave. 8.6.7 Anchors Anchors are basically of two types, relying either on self-weight or suction forces. The traditional embedment anchors, as shown in fig. 8.28, are not normally designed for vertical force components. Holding power is related to anchor weight and type of seabed.
- 47. 694 zyxwvutsrqpo Chapter z 8 z Figure 8.28 Drag anchor Figure 8.29 Deep water FPSO design using suction anchors Figure 8.29 depicts a deep water floating production vessel moored with a taut station keeping system of fibre rope using suction anchors. These allow vertical anchor loads. The angle at the line lower-end is noted as being 40” to the horizontal. Figure 8.30 shows a typical suction anchor installation sequence. By reversing the suction process, the anchor can be “pushed” from the seabed using over-pressure. Piles can be used as an alternative to anchors. However, they require a large crane installation vessel with piling capability. 8.6.8 Turrets’ The design of monohull turret structures used for single-point moorings in floating production systems must allow for large static and dynamic loading caused by the vessel motions in waves together with forces transmitted by the mooring system. The hull design in the turret region must reflect the fact that the amount of primary steel is reduced here
- 48. Mooring zyxwvutsrqp Sjstems zyxwvutsrqp Figure 8.30 Suction anchor installation sequence zyxw 695 with an appropriate increase in the stress concentration. A comparison of the existing developments using turret-moored vessels in use indicates wide variations of turret position. Indeed some early North Sea designs use a turret placed close to the vessel amidships, whereas a number of Far Eastern applications place a disconnectable turret off the bow. Careful selection of turret position is important because of its influence on: Mooring line tension and riser loading zyxwv - The turret position alters the vessel yaw and hence the surge and sway motions, thus influencing the mooring line tension. This is also affected by the vessel heave and pitch motions. In particular, the pitch contribution to the turret vertical motion is relatively high for the turrets near or off the vessel bow. The combined effects can also result in high loading on the riser system. Vessel yaw - The motion response magnitude in yaw is likely to increase significantly if the turret is placed close to amidships, because the yaw restoring moment causing the vessel to head into weather is reduced. Use of azimuthing thrusters, if fitted, can be employed to control the yaw but with an increased capital and operating cost. Increased yaw results in more wear on the turret bearing, together with higher downtime because of inertial loading from the vessel motions. It can also cause yaw instability of the vessel. The low-frequency yaw about the turret also needs to be restricted in order that hydrocarbon off-loading from the vessel stern can be carried out with high operability levels. Figure 8.31 shows the stern horizontal displacements for two vessels [Brown, et a1 19981 with turrets positioned at 12 and 36.5% of the hull length from the vessel
- 49. 696 zyxwvutsrqpo Chapter z 8 z 0 20zyxwv 30 40 Xd&pl zyxw ai8trm (m) zyxw Figure 8.31 Monohull stern horizontal motion in head seas amidships responding in identical sea states of H, zyxw = 8.7 m and Tp= 11.8 s wind of 60 kt was also simulated at 60” to the wave direction (using fans and a turret-mounted spring mechanism). The results show large increases in stern transverse (Y) motions when the turret is closer to amidships. Rigid body oscillation in the horizontal plane - The natural frequencies and amplitude of oscillation can be affected by the position of the turret. The full low-frequency vibration behaviour of a turret-moored vessels is not well understood. The turret rotates within the vessel hull using a combination of radial and thrust bearings positioned on roller assemblies at deck and within the hull. Transmission of hydrocarbons from the non-rotating components, such as the turret and risers, to the weather-vaning vessel is carried out using either a stacked swivel or “drag chain” type system. This also permits the continuous transfer of hydraulic and electrical control lines. 8.7 Industry Standards and ClassificationRules The specific requirements for design of mooring systems are defined in Classification Rules and Industry Recommended Practices by API RP 2SK, Det Norske Veritas and Bureau Veritas. Additionally Lloyds, NMD, NPD and IACS provide similar rules and design information. Industry Guidance Notes or Recommended Practices are non-binding recommendations, which are sometimes incorporated into design criteria either in whole or
- 50. Mooring Systems zyxwvutsrq 697 in part. Classification Rules or Offshore Standards are invoked, if the owner of a platform elects to have the platform classed. In this case they become binding rules. The specific requirements for floating production systems vary among these various reference^.^ MODU rules do not explicitly cover mooring and leave the specification of safety factors and other conditions to the owner. There is a significant difference in the current mooring criteria between European (mainly North Sea) and the U.S. Gulf of Mexico as reflected in API RP 2SK. zyx As an example, the DNV Offshore Standard for Position Mooring specifies different safety factors for design depending on the criticality of the production. The safety factors are also applied differently. The DNV practice applies a separate safety factor to the computed mean load (FOS zyxwvuts = 1.4)as opposed to the dynamic load (FOS =2.1) (for dry tree applications). On the other hand, the API recommendation is for a single safety factor of 1.67 to be applied to the peak load for all types of mooring systems. The European standards also make allowances for application of the quantitative risk assessment methods for the selection of appropriate design loads. 8.7.1 Certification Representative certification authority rules, such as those issued by DNV (2001) give guidance on relevant issues associated with mooring systems. There is strong emphasis on catenary analysis using chain and wire and, more recently, guidance on taut moorings using fibre ropes. The standards are, in many cases, developed from those for mobile drilling units. The objectives of the standards are to provide: The standards are typically divided into a number of sections as follows: Environmental conditions and loads Mooring system analysis Thruster-assisted mooring Mooring equipment Testing A further description of certification standards is given below for one particular authority. It is necessary to refer to the relevant certification standard for full information. Uniform level of safety to mooring systems, Guideline for designers, suppliers and contractors, Reference document for contractual considerations between suppliers and contractors. 8.7.2 Environmental Conditions and Loads Survival environmental criteria for permanent moorings are usually based on a 100-yr return period event. It is common to use two or three environments including the 100-yr 3Recommendedpractices are subject to continual review and updating. These values should not be considered definitive. The latest documentation should be consulted.
- 51. 698 zyxwvutsrqpo Chapter 8 wave with associated wind and current, and the 100-yr wind with associated wave and current. In high current environments, such as, the Gulf of Mexico deepwater, North Atlantic and certain areas of West Africa and Southeast Asia, the current may be the controlling event and a 100-yr current plus associated wind and waves is also specified. Specification of “associated wind and waves” is somewhat subjective. A more rigorous method for specifying environment is to perform a “response-based analysis”, see for example Standing, et a1 (2002). This method employs a simplified mathematical model of the platform and mooring responses to various environmental conditions. Hindcast environmental data covering many years, including extreme events, is compiled and used as input to this model. This might involve thousands of cases covering, for example, hindcast conditions every 6 h going back 10-20 yr at the specific site. The statistics of the responses are tabulated to determine a “100-yr response” usually defined as that response having a 0.01 chance of exceedance in any year. The environments which generate this response and responses close to this response are chosen for more refined analysis. “Response-based modelling” is not presently required by any rules or recommended practice, but it may be specified by the owner. The DNV Offshore Standards recommend determining a “100-yr response” for design based on a compilation of wave heights, periods covering a span of 100-yr environments and selecting the combination yielding the worst response. In order to calculate the mooring line structural response it is necessary to apply appro- priate environmental loads for the site under consideration. This usually corresponds to the wave and wind conditions having return periods of 100-yr, together with 10-yr return period current conditions. However, if, for example, current and wind are the dominant features, such as Gulf of Mexico conditions with loop currents and hurricanes, then IO-yr sea conditions combined with 100-yr current and wind should be assessed. A number of sea states should be selected along a “contour line” representing the joint probability of significant wave height and peak wave period combinations at the mooring location. The contour represents wave height and period pairs for a specified return period, for example, 100-yr. Guidance notes and standards give examples of contour lines. Wind loads should consist of both steady and time-varying components, the latter being specified in both DNV and API documentation. The weather directions to be considered depend on the vessel mooring arrangement. For vessels that cannot change direction relative to the weather it is necessary to consider waves, wind and current acting from the same directions. These are head, quartering and beam, together along with the mooring line for vessels with the symmetric mooring patterns. For non-symmetric mooring patterns, all directions, with a maximum 45” spacing, should be assessed. For vessels that can weathervane, site data should be used, if available, otherwise collinear weather should be applied at 15“ to the vessel bow, together with a non-collinear condition with bow waves, wind and current acting from the same side at respectively 30 and 45” to the bow. Wind and current loads can be established by model tests and/or calculations, see for example OCIMF (1994). Calculations are based on a drag force formulation, suitable coefficients being established from model tests or computational fluid dynamics. Current forces will increase, if the water depth is typically less than three times the vessel draught, OCIMF providing relevant enhancement factors. Current forces on multiple riser systems should be considered though forces on a system consisting of only a single riser are usually
- 52. Mooring Systems zyxwvutsrq 699 z ignored. Current loads on moorings are only considered, if these are dominant, such as at sites with loop currents. Marine growth on long-term moorings should be included by increasing the line weight and drag coefficient, zyxwvut C , . A marine growth density of 1325 kg/m3is common, and the standards provide equations to calculate the mass of growth depending on the line type and diameter, together with growth thickness and water depth. The line drag coefficient can be assumed to increase linearly with growth thickness. For new lines, the standards indicate the following drag coefficients: Cd= 2.6 for stud chain, Cd= 2.4 for studless chain, Cd = 1.8 for six-strand steel wire rope, Cd = 1.2 for spiral strand with sheathing, Cd = 1.6 for spiral strand without sheathing. Waves provide three loading mechanisms acting on the floating vessel. These result in mean wave drift motions, and responses at wave and low frequency as described in Section 8.3. For catenary moored structures, the restoring stiffness contributions to the wave frequency motions from the mooring and riser system are ignored in deep waters, though must be investigated for water depths below 70 m. For the taut moored structures, the restoring forces from the mooring and riser system must be addressed to establish whether they influence motions at wave frequencies. Shallowwater also influences the horizontal motions of the vessel for depths less than 100m, in that surge and sway motion amplification factors must be included. These can result in a doubling of the deepwater motions for large wave periods in a very shallow water. Low-frequency motions for semi-submersibles and ships should be calculated in the horizontal directions only, that is, surge, sway and yaw. For deep draft floaters, such as spar platforms, vertical responses also need to be assessed. It is important to establish a stable equilibrium position for the vessel, where the steady forces of current, wind and wave drift balance the restoring forces from the station-keeping system. For systems that are free to yaw, vessel rotation should be included when calculating the mean forces. The frequency or time-domain methods may be used to establish the vessel low-frequency response about this stable equilibrium position. Alternatively, the model test results may be used. It is important that the model test or simulation is carried out over a suitable length of time to give appropriate statistical quantities. zyxw A minimum of 3 h full-scale equivalent time is specified, though usually significantly longer time is beneficial. The model testing has been addressed in detail in Chapter 13. 8.7.3 Mooring System Analysis Certification standards give guidance on the methods employed to perform the structural design of wire, chain and fibre mooring systems, including their combinations, used on floating vessels, including deep draft floaters, such as spars. The mooring system is assessed in terms of three limit states based on the following criteria: Ensuring that individual mooring lines have suitable strength when subjected to forces caused by extreme environmental loads - ultimate limit state (ULS).
- 53. 700 zyxwvutsrqpo Chapter zy 8 Ensuring that the mooring system has suitable reserve capacity when one mooring line or one thruster has failed zyxwvu - accidental limit state (ALS). Ensuring that each mooring line has suitable reserve capacity when subject to cyclic loading - fatigue limit state (FLS). Guidance on the structural stiffness characteristics of wire, chain and synthetic fibre is given. For wire, this depends on whether the make-up is six strand or spiral strand; for chain, the stiffness depends on chain diameter. For fibre moors, it is necessary to establish the non-linear force-extension behaviour of the rope. If this is not available, then the vessel excursion should be established using the estimated post-installation line stiffness for both the ULS and ALS. Characteristic line tensions for ULS, ALS and FLS can be found using the storm stiffness. Section 8.4.5 describes these stiffness criteria in more detail. The analysis procedures are divided into those attributable to establishing the platform response, and those associated with calculating the mooring line behaviour. Mooring line analysis must include the influence of line dynamics, if the vessel is to be used for float- ing production or storage, or if operations in depths greater than 200 m are considered. Additionally, vortex-induced vibration needs to be addressed for platforms of deep draft. The platform response is, in many situations, strongly influenced by the damping associated with the low-frequency motions. This depends on sea and current conditions, mooring and riser make-up, together with water depth. Model tests can be used to establish damping, though as described in Section 8.4.9,damping levels associated with the mooring are difficult to quantify. Risers can provide restoring, damping and excitation forces making their influence on floater response more complicated. The mooring analysis should ideally consider line dynamics, i.e. the inertia and drag force contributions acting on the line components, when calculating line loading associated with the platform wave frequency motions. Quasi-static analysis, allowing for submerged weight and elasticity of line, platform motion and seabed reaction/friction forces, is usually appropriate when dealing with platform mean- and low-frequency motions. In establishing the characteristic line tension for either the ULS or ALS, Gaussian stati- stical methods are used, recognising the random nature of the platform response and line tensions under realistic environmental conditions. This allows the maximum wave and low-frequency platform excursions to be found, based on the relevant motion standard deviation and the number of oscillations during a specifiedperiod, usually taken as 3 h. The above excursions are combined, after including the mean offset, by taking the larger of the sum of the significant and maximum excursions. Finally, if line dynamics are considered, the maximum wave frequency line tension is obtained from its standard deviation. This depends on the excursion about which wave frequency motion occurs and the number of associated platform oscillations. Combining this with the mean and quasi-static tension components gives the characteristic dynamic line tension. The mooring analysis must also consider the characteristic capacity or strength for the ULS and ALS, recognising that the line strength is likely to be less than the average strength of its components, whether these be chain links or wire fibres. Thus the charac- teristic capacity includes the influence of the component mean breaking strength and
- 54. Mooring Sjstems zyxwvutsrq 701 z its coefficient of variation. Other connecting links and terminations must be designed with higher strength characteristics than the main line elements, together with improved fatigue lives. The design equations to be used for ULS and ALS are based on the concept of partial safety factors (see Chapter 5). The design equation is of the form: where zyxwvuts S, is the line capacity and T,,,,,,, Tc,dyn are the characteristic mean and dynamic tensions. The partial safety factors, zyxwv Y , , , , and tidy,, are specified in the standards. These take on values of between 1.1 and 2.5 for the ULS, and 1.0 and 1.35 for the ALS. The values depend on the intended operation of the vessel, in that higher factors are imposed where mooring failure could lead to unacceptable situations such as loss of life, collisions, sinking or hydrocarbon release. The safety factors are also higher, if a quasi-static analysis, as opposed to a more rigorous dynamic analysis is carried out. In evaluating the vessel excursions and line tensions, care must be taken not to exceed the permissible vessel offset and line length. For example, horizontal offsets will be influenced by gangway connections to another fixed or floating structure. For rigid riser operations, offsets are limited by the maximum allowable riser angle at the BOP flex joint, and must also allow for heave compensation equipment. Manufacturers’ limitations must be considered for flexible risers and steel catenary risers. Line lengths are influenced by whether anchors can withstand up-lift loads. N o up-lift is allowed for the ULS, but up-lift may be allowed for the ALS, if the vertical loads do not impair the anchor-holding power. The layout of the subsea architecture must also be considered within the context of mooring system analysis. For the ULS and ALS there must be a minimum vertical clearance between lines and all subsea equipment of respectively 10 and 0 m (no contact). z A further safety factor should also be applied for situations where analysis has been performed at the limiting sea state for normal operations, usually corresponding to mild weather. The safety factor applies to the mean and dynamic tension components, that is the last two terms on the left hand-side of equation (8.10). For mooring chains designed to be positioned at the same location for greater than four years, the characteristic capacity of the line must be reduced for the effects of corrosion. This corrosion reduction is larger for components at the seabed and in the surface splash zone. If regular inspection schemes are to be carried out, the required corrosion reductions are smaller. For steel wire rope, the lifetime degradation depends on the construction and level of protection applied. Note however that when addressing the FLS, only 50% of the corrosion allowance need be applied. When considering the mooring FLS, it is necessary to account for the accumulated fatigue damage that occurs from cyclic loading by individual sea states making up the long-term environment. The relevant vessel heading should be allowed for. For each of these sea states, it is necessary to calculate the mooring system response together with the sea state occurrence frequency. In practice, the long-term environment can be discretised into something like 8-12 headings and 10-50 sea states.
- 55. 702 zyxwvutsrqpo Chapter z 8 In an individual sea state the fatigue damage d, is given by: zyx (8.11) where the number of stress cycles, zyxwvu n,, is calculated from the product of the mean up- crossing rate of the stress process (in zyxwv Hz), the probability of occurrence of the sea state, together with the mooring system design lifetime in seconds. The term fsl represents the probability density of peak to trough nominal stress ranges for the individual state. The stress ranges are obtained by dividing the line tension ranges by the nominal cross-sectional area. This is taken as xd2/4 for steel wire rope and 2xd2/4 for chain. The procedure for fibre ropes is described further here. The term n, in equation (8.11) represents a fatigue property of the line, giving the number of stress ranges of magnitude s that would lead to failure. For wire and chain, the capacity against fatigue caused by tension is defined in terms of the number of stress range cycles given by: log(nc(s))= log(aD) - m log(s) (8.12) where zyxwvut s is the stress range double amplitude (MPa) and m, and uD are the slope and intercept on the seawater S-N curves, given in the Standards for various chain and wire rope types. In practice, the integral given in equation (8.11) can be replaced by discrete terms for each sea state zyxwvu i, in terms of the expected value of the nominal stress range. Additionally, if the stress process has negligiblelow-frequency content, then narrow-banded assumptions allow the damage to be established in terms of the stress standard deviation. If, however, there are wave and low-frequency contributions to the stress, then rainflow counting will provide the most accurate estimate. For this situation, two alternatives, the combined spectrum or dual narrow-banded approach, described in the standards can be used. For fibre rope, the capacity against fatigue caused by tension-tension effects is given by: (8.13) where R is the ratio of tension range to characteristic strength and in, UD are given in the standards. The design equation to be used for FLS is similar to that for ULS and ALS, being of the form: where d, is the damage that accumulates as a result of all the individual environmental states over the system design lifetime, and yF is a fatigue safety factor. The following guidance is given on safety factors: yF= 3, for wire and chain line that can regularly be inspected on-shore. zy ~ ~ = 5 , for wire and chain line that cannot regularly be inspected on-shore, and is configured so that the ratio of fatigue damage in two adjacent lines is less than 0.8.
- 56. Mooring zyxwvutsrqpo Systems zyxwvutsrq 703 Common Chain Link Baldt or Kenter Connecting Link yF= 5-8, for wire and chain line that cannot regularly be inspected on-shore, and is configured so that the ratio of fatigue damage in two adjacent lines is greater than 0.8. z 0 yF= 60, for polyester rope. Note that this is much larger compared to steel because of the increased variability in fatigue test results. Fatigue properties of wire and chain are typically defined in terms of T-N relationship derived from tension-tension fatigue tests. Similar to conventional S-N fatigue curves, the design fatigue curve is in the form: 3.36 370 3.36 90 N = K . R - ~ (8.15) where N zyxwvuts = number of cycles, R = ratio of tension range (double amplitude) to nominal breaking strength, M = slope of T-N Curve, and K = intercept of the T-N Curve. M and K are given in table 8.7, where Lm = ratio of mean applied load to the breaking strength of wire rope from the catalogue. The chain fatigue data presented in API RP2SK is for the stud link chain. DNV OS-E301 presents data in the form zyxwvu [API, Chaplin, 19911: 1 Sixjmulti strand rope 14.09 n,(s) = zyxwv aos-" (8.16) where n&) = number of stress ranges, s = stress range (MPa), aD = intercept of the S-N curve, m = slope of the S-N curve. Values of aD and m are given in table 8.8. lo(3.20-2.79Lm) Table 8.7 Fatigue curve parameters for wire rope and chain (from API RP 2SK) Spiral strand rope 15.05 lo(3.25-3.43Lm) IComponent lM IK 1 Spiral strand rope, Lm = 0.3 5.05 166 Type a D m Stud chain ~ 1.2x10" 1Studless chain 1 6 . 0 ~ 10" 1 3.0 1 3.0 1Six strand wire rope i 3 . 4 ~ 1014 1 4.0 1 [Spiral strand wire rope ~ 1.7~10" I 4.8 1
- 57. 704 zyxwvutsrqp Figure 8.32 Chain wire fatigue curves based on stress (DNV OS-E301) zyx Chapter 8 The DNV curves are shown in fig. 8.32. This relationship is similar to the API curve, but it is based on stress rather than tension. In order to convert from tension to stress the nominal steel areas given in a table in the API RP2SK may be used. The fatigue of wire rope and chain running over sheaves and fairleads will generally be lower than pure tension-tension fatigue. Additional stress due to bending may be used to account for this effect. For effects other than tension fatigue, for example chain or wire bending and tension-compression for fibre ropes, further consideration, such as experimental testing, is required. As an alternative to the above procedures, mooring design may be carried out using structural reliability analysis. Standards give guidance on target annual probabilities of failure when performing reliability analysis. 8.7.4 Thruster-Assisted Mooring This section of the standards gives methods and guidance associated with the design of thruster-assisted moorings. Thrusters can be used to reduce the mooring system loads caused by mean environmental forces. provide damping of the low-frequency motions and assist in heading control. For manual and automatic remote control systems respectively 70 and 100% of the net thrust can be used when establishing the ULS or ALS. However, if a failure leads to a thruster stop situation during the ALS then this must be considered equivalent to a line failure. The available (net) thrust can be estimated by calculation at the early design stage based on the propeller thrust at bollard pull. A useful conversion factor is 0.158 kN/kW for nozzle
- 58. Mooring Sjstems 705 z propellers and 0.105 kN/kW for open propellers. These values need correcting to account for in-flow velocity at the propeller, propeller rotation sense and propeller/thrust installation geometry and arrangement zyxwv - see for example API RP 2SK for further guidance and Ekstrom, et a1 (2002) for information on the thruster-thruster interaction. Thrust contributions to station-keeping can be evaluated using the methods of mean load reduction and system dynamic analysis as follows: The mean load reduction method involves subtracting the surge and sway components of allowable thrust from the mean environmental loads for spread-moored vessels. For single-point moored vessels, the standards give guidance for methods to establish the contribution to the yaw moment when thrusters are used to influence vessel heading. A system dynamic analysis generally consists of a surge, sway and yaw simulator. This can produce mean offset and low-frequency vessel responses corresponding to time- domain records of environmental force. Wave frequency forces are not balanced by the system. Thrusters can consist of both fixed and rotating configurations and be of variable pitch and speed. The selection is made based on the requirements of the mooring system, but the appropriate configuration must have an automated power management system. There should be a manual or automatic remote thrust control system. Automatic control systems are more sophisticated than manual and can have features such as monitoring of vessel position and line tension alarms, consequence analysis and simulation capabilities, relevant data logging, self-diagnostics and allow system response to major failures. Further details are given in the standards. zyxwvu 8.7.5 zyxwvutsr Mooring Equipment Standards provide requirements for all mooring equipment and its installation for temporary and emergency mooring, position mooring and towing. Only a brief overview is given here. Information on various anchor types is provided including fluke, plate, piled, gravity and suction anchors. Specifications for anchor construction materials are also discussed. Data on mooring chains and associated connecting links and shackles is also provided. Offshore mooring chain is graded depending on its minimum yield and tensile strength, together with Charpy v-notch energy. For long-term mooring systems, where onshore inspection is not possible, only limited connection elements, such as D shackles or triplates (fig. 8.25), are acceptable. Where mobile offshore units change location frequently, other connections such as Kenter shackles, C links and swivels are allowed in the mooring line make-up. Generally there is a lack of documented fatigue data on these latter connection elements, though API RP 2SK does provide fatigue information on Kenter shackles. Six-strand wire rope (fig. 8.23) is normally used by mobile offshore vessels for anchor and/ or towing lines. This rope is commonly divided into two groups; either 6 by 19, consisting of 6 strands with between 16and 27 wires in each strand; or 6 by 36, consisting of 6 strands with between 27 and 49 wires in each strand. Long-term floating production vessels use spiral strand steel wire ropes as this has improved fatigue and corrosion behaviour.
- 59. 706 zyxwvutsrqpo Chapter 8 Synthetic fibre ropes can be used either as inserts in a catenary mooring layout or as part of a taut leg system. Recognised standards, such as API RP 2SM have been produced that document the use of fibre ropes. The technology is still developing, but fibres being considered for mooring system use include polyester, aramid, high-modulus polyethylene (HMPE) and nylon. Standards specify the relevant load bearing yarn properties and tests to be documented, together with those for the yarn sheathing material. Rope constructions under consideration are parallel strands, parallel yarns and “wire rope constructions”. Braided constructions are not considered because of the concerns over their long- term fatigue behaviour. Guidance is also given on stiffness values for polyester, aramid and HMPE for post-installation, drift and storm conditions for deepwater fibre moorings. Other potential failure modes are also discussed in the standards including: hysteresis heating zyxwvu - lubricants and fillers can be included to reduce hotspots, creep rupture - in particular this is relevant to HMPE yarns, and the risks need careful evaluation, tension ~ tension fatigue - only limited data exist, indications being that fatigue resistance is higher than for steel wire ropes. axial compression fatigue - on leeward lines during storms for example, prevented by maintaining a minimum tension on the rope, particle ingress - causes strength loss by abrasion from water-borne material such as sand, prevented by using a suitable sheath and not allowing contact between the rope and seabed. Fibre rope terminations under consideration included socket and cone, conventional socket and spliced eye, the latter being the only one presently qualified at sizes appropriate to deep-water mooring systems. The standards give design, material requirements and capacity for additional mooring hardware including windlasses, winches, chain stoppers and fairleads together with end attachments. The necessary structural arrangement for the mooring equipment is also specified, together with arrangements and devices for towing purposes and measurement of line tension. Lee, et a1 (1999) describe the ABS approach on synthetic ropes, while Stoner, et a1 (1999) present the contents of an engineer’s design guide for fibre moorings, emphasising the limitations in the available test data. Stoner, et a1 (2002) outline additional work necessary before fibre moorings can be used at harsh weather locations. zy 8.7.6 Tests The standards give comprehensive guidance on tests to be carried out on mooring system hardware including the following: zyxwv 0 Mooring chain and accessories, Steel wire rope, Windlass and winch assemblies, 0 Synthetic fibre ropes. Fluke anchors for mobile/temporary and long-term moorings, Manual and automatic remote thruster systems,
- 60. Mooring Systems zyxwvutsrq 707 More information can be found in the standards. For example, the UK Health zy & Safety Executive (2000) gives a comprehensive discussion of model testing techniques for floating production systems and their mooring systems. zyxw References American Petroleum Institute, “Recommended practice for design and analysis of stationkeeping systems for floating structures”, API RP-2SK (latest edition). American Petroleum Institute (March 2001). “Recommended practice for design, manufacture, installation and maintenance of synthetic fiber ropes for offshore mooring”, API RP 2SM, (Ist ed.). Aranha, J. A. P., Pinto, M. O., and Leite, A.J.P. (2001a). “Dynamic tension of cables in random sea: Analytical approximation for the envelope probability density function”. Applied Ocean Research, Vol. 23, pp. 93-101. Aranha, J. A. P. and Pinto, M. 0. (2001b) “Dynamic tension in risers and mooring lines: An algebraic approximation for harmonic excitation”. Applied Ocean Research, Vol. 23, Brown, D. T. and Liu, F. (1998). “Use of springs to simulate wind induced moments on turret moored vessels”. Journal of Applied Ocean Research, Vol. 20, No. 4, pp. 63-81. pp. 213-224. Brown, D. T., Lyons, G. J., and Lin, H. M. (1997). “Large scale testing of mooring line hydrodynamic famping contributions at combined wave and frift frequencies”, Proc. Boss z 97, zyxwvutsr srhIntl. Con$ on Behaviour of Offshore Struct., Delft, Holland, ISBN 008 0428320, pp. 397406. Brown, D. T. and Mavrakos, S. (1999). “Comparative study of mooring line dynamic loading”. Journal of Marine Struct., Vol. 12, No. 3, pp. 131-151. Chaplin, C. R. (August 1991). Prediction of Wire Rope Endurance for Mooring of Offshore Structures, Working Summary, Joint Industry Project (JIP) Report issued by Noble Denton & Associates, London. Chatjigeorgiou, I. K. and Mavrakos, S. A. (2000). “Comparative evaluation of numerical schemes for 2-D mooring dynamics”. International Journal of Offshore and Polar Engineering, Vol. 10(4), pp. 301-309. CMPT (1998). Floating Structures: A Guide for Design and Analysis. Vol. 2, Ed. Barltrop, N. 101/98. Det Norske Veritas OS-E301 (June 2001). “Position Mooring”. Ekstrom, L. and Brown, D. T. (2002). “Interactions between thrusters attached to a vessel hull”, 21” Offshore Mechanics and Arctic Engineering Intl Conf., American Society of Mechanical Engineers, Paper OMAE02-OFT-28617, Oslo, Norway. Gobat, J. I. and Grosenbaugh, M. A. (2001a). “A simple model for heave-induced dynamic tension in catenary moorings”. Applied Ocean Research, Vol. 23, pp. 159-174.
- 61. 708 zyxwvutsrqpo Chapter z 8 Gobat, zyxwvuts J. I. and Grosenbaugh, M. A. (2001b). “Application of the generalized-cr method to the time integration of the cable dynamics equations”. zyxw Computer Methods in Applied Mechanics and Engineering, Vol. 190, pp. 48174329. Health zyxwvuts & Safety Executive, UK. (2000). “Review of model testing requirements for FPSOs”, Offshore Technology Report 2000/123, ISBN 0 7176 2046 8. Huse, E. and Matsumoto, K. (1989). “Mooring line damping owing to first and second order vessel motion”, Proc. OTC, Paper 6137. Huse, E. (1991).“New developments in prediction of mooring line damping”, Proc. OTC, Paper 6593, Houston, USA. Lee, M., Flory, J., and Yam, R. (1999). “ABS guide for synthetic ropes in offshore mooring applications”, Proc. OTC, Paper 10910, Houston, Texas. Matsumoto, K. (1991). “The influence of mooring line damping on the prediction of low frequency vessels at sea”, Proc. OTC, Paper 6660, Houston, USA. Myers, J. J., ed. (1969). Handbook of Ocean and Underwater Engineering, McGraw-Hill Book Company. NTNF (1991). “FPS 2000 Research Programme - Mooring Line Damping”, Part 1.5, E Huse, Marintek Report. Oil Companies International Marine Forum (OCIMF) (1994). “Prediction of wind and current loads on VLCCs ”, (2nded.) Standing, R. G., Eichaker, R., Lawes, H. D., Campbell, and Corr, R. B. (2002). “Benefits of applying response based analysis methods to deepwater FPSOs”, Proc. OTC, Paper 14232, Houston, USA. Stoner, R. W. P., Trickey, J. C., Parsey, M. R., Banfield, S. J., and Hearle, J. W. (1999). “Development of an engineer’s guide for deep water fiber moorings”, Proc. OTC, 10913, Houston, Texas. Stoner, R. W. P., Ahilan, R. V., and Marthinsen, T. (2002). “Specifying and testing fiber moorings for harsh environment locations”, Proc. OMAE, 28530. Thomas, D. 0.and Hearn G. E. (1994). “Deep water mooring line dynamics with emphasis on sea-bed interaction effects”, Proc. OTC, 7488, Houston, USA. Vandiver, J. K. (1988). “Predicting the response characteristics of long flexible cylinders in ocean currents”, Symposium on Ocean Structures Dynamics, Corvallis, Oregon. Webster, W. (1995). “Mooring induced damping”. Ocean Engineering, Vol. 22, No. 6, pp. 571-591.
- 62. Handbook of Offshore Engineering zyxwvutsr S . Chakrabarti (Ed.) zyxwvuts 02005 Elsevier Ltd. zyxwvutsrq All rights reserved 709 Chapter 9 Drilling and Production Risers James Brekke zyxwvuts GlobalSantaFe Corporation, Houston, TX, USA Subrata Chakrabarti Offshore Structure Analysis, Inc., Plainfield, IL, USA John Halkyard Technip Offshore, Inc., Houston, TX, USA zyxw 9.1 Introduction Risers are used to contain fluids for well control (drilling risers) and to convey hydro- carbons from the seabed to the platform (production risers). Riser systems are a key component for offshore drilling and floating production operations. In this chapter section 9.2 covers drilling risers in floating drilling operations from MODUS and section 9.3 covers production risers (as well as drilling risers) from floating production operations. A riser is a unique common element to many floating offshore structures. Risers connect the floating drilling/production facility with subsea wells and are critical to safe field operations. For deepwater operation, design of risers is one of the biggest challenges. During use in a floating drilling operation, drilling risers are the conduits for operations from the mobile offshore drilling unit (MODU). While connected much of the time, drilling risers undergo repeated deployment and retrieval operations during their lives and are subject to contingencies for emergency disconnect and hang-off in severe weather. Production risers in application today include top tension production risers (TTRs). flexible pipes steel catenary risers (SCRs), and free-standing production risers. More than 50 different riser concepts are under development today for use in deepwater and ultra- deepwater. A few of the most common riser concepts are shown in fig. 9.1. According to Clausen and D’Souza (2001), there are more than 1550 production risers and 150 drilling risers in use today, attached to a variety of floating platforms. About 85% of production risers are flexible. Flexible risers are applied in water depths of up to 1800 m, while a top tension riser and a steel catenary riser are used in depths as much as 1460 m. The deepest production riser in combined drilling and early production is in a water depth
- 63. 710 zyxwvutsrqpon Chapter 9 z Figure 9.1 Schematic of riser concepts [Courtesy of Clausen and D’Souza, Subsea’llKBR zy (ZOOl)]
- 64. Drilling and Production zyxwvutsrqp Risers zyxwvutsrq 71 z 1 zy Figure 9.2 Vertical tensioned drilling riser [Note: balljoint (or flex joint) is also located just below drill floor] of 1853 m in Brazil for the Roncador Seillean FPSO. Drilling risers are in use in greater than 3000 m depth. A top tensioned riser is a long slender vertical cylindrical pipe placed at or near the sea surface and extending to the ocean floor (see fig. 9.2). These risers are, sometimes, referred to as “rigid risers” or “direct vertical access” risers. The development of different types of riser with the riser size (diameter in inches) and water depth up to 2000 m is shown in fig. 9.3. The envelopes for the different riser types are given in the figure. The installed SCRs for the floaters are identified in the figure. The technical challenges and the associated costs of the riser system increase significantly with water depths [Clausen and D’Souza, 20011. The cost of a riser system for a deepwater drilling and production platform compares with that of the hull and mooring system. The risers connecting a floating vessel and the seafloor are used to drill or produce individual wells located beneath the floating vessel or for import and export of well stream products. They are connected to a subsea wellhead, which in turn is attached to the supporting sub-mudline casing. The drilling riser is attached via an external tieback connector, while the production riser can be attached via either an external or internal tieback connector [Finn, 19991.The first joint of the riser above the tieback connector is a
- 65. 712 zyxwvutsrqpo Chapter 9 Figure 9.3 Progress of production riser diameters with water depth [Courtesy of Clausen and D’Souza, Subsea7/KBR (2001)) zyxw special segment called the stress joint that is designed to resist the large bending moments, but flexible enough to accommodate the maximum allowable riser angular displacements. Typically these joints are composed of a forged tapered section of pipe that can be made of either steel or titanium. Newer designs call for the stressjoints to be composed of a series of pipe segments that are butt-welded or a group of concentric pipes welded to a special terminating flange. In lieu of a stress joint, elastomeric flex or ball joint may be used to accommodate bending at the sea floor. The top tension risers are initially held in a desired tension which helps in the bending resistance of the riser under the environmental loads. This tension is provided by a mechanical means, as shown in fig. 9.4a for a drilling riser. The tension may also be provided by syntactic foam or buoyancy cans. A top tension riser designed for the appli- cation with Spar is shown in fig. 9.4b. The Spar riser uses buoyancy tanks for the top tension. The riser entering the keel of the spar is detailed in the figure. Three different riser pipe configurations are illustrated in fig. 9.4b. In the first case the Neptune Spar uses a single 9-5jS in. diameter casing, which encompasses the production tubing and other annulus lines. In the second case a dual casing riser is used with internal tubing. In the third configuration the riser tubing strings are separate, requiring fewer riser pipes and lessexternal buoyancy. It is better suited for deeper waters where large riser weight becomes a problem. The selection of the riser configuration is based on a risk/cost benefit analysis. The general riser dimensions are based on the reservoir information and the anticipated drilling procedures. The size of the tubing is determined from the expected well flow rate. The wall thickness of each riser string is computed from the shut-in pressure and drilling and completion mud weights. The outside dimension of the components that must pass through the pipe, such as subsurface safety value (SSSV), drill bit, or casing connector generally determines the internal diameter of the riser. The hoop stress usually governs the wall
- 66. Drilling zyxwvutsrqpon and Production Risers zyxwvutsr Figure 9.4 Drilling and production riser configuration zyx 713 thickness of the riser pipes. In deeper waters, the wall thickness may depend on the axial stress. The capped-end force generated by the internal pressure should also be considered in computing the axial stress. The bending stress is a determining factor at the upper and lower ball joints of the riser. In these areas thicker riser elements may be required to limit the stresses. The dimensions of the stress joint are more difficult to compute since they must be strong and flexible at the same time. Generally, a finite element program is used that
- 67. 714 zyxwvutsrqpo Chapter z 9 z determines the riser bend to the desired maximum angle at the joints. The dimensions are adjusted until the required strength is achieved. The potential for riser interference is also checked during an early determination of the riser component dimensions. zy 9.2 Drilling Risers This section provides a description of the analysis procedures used to support the operation of drilling risers in floating drilling. The offshore drilling industry depends on these procedures to assure the integrity of drilling risers, with the goal of conducting drilling operations safely, with no environmental impact, and in a cost-effective manner. The main emphasis of this section is on drilling risers in deep water (Le. greater than 900 m or 3000 ft) and some specific coverage is given to drilling from dynamically-positioned drillships in ultradeep water (Le. greater than 1800 m or 6000 ft of water). Besides analytical procedures, some coverage is given to the operational procedures and the equipment that are peripheral to the drilling riser. However, a comprehensive treat- ment of drilling riser operations and equipment is outside the scope of this chapter. References to the industry guidelines given below provide additional details. As the water depths for drilling operations have increased, the importance of the drilling riser has grown in importance. Effectiveanalytical support of the drilling riser and the related operations can substantially reduce the cost and risk of drilling an offshore well. The potential loss of a drilling riser presents high consequences. Currently, the cost of the drilling riser can be tens of millions of dollars; but in addition, the cost of operational downtime for an event involving the loss of a drilling riser can exceed one hundred million dollars. Avoidance of such losses further benefits the entire oil and gas industry through improved safety, reduced environmental impact, and reduced insurance cost. Some of the guidelines for analysis and operation of drilling risers are contained in API Recommended Practice 16Q (1993). As of this writing, this document, API RP 16Q, is being revised for release by the International Standards Organisation (ISO). Another related document, API Bulletin 5C3 is referenced for its collapse and burst formulas used in drilling riser design. This document is entitled “Bulletin on Formulas and Calculations for Casing, Tubing, Drill Pipe, and Line Pipe Properties, API Bulletin 5C3, Sixth Edition, October 1, 1994”. This section will cover some of the important aspects in the procedures for drilling riser analysis. This begins with a discussion of metocean conditions, which are a primary driver in determining the operationallimitations of a drilling riser at a specificsite. This is followed by discussions of the design and configuration of a riser, including the issue of vortex-induced vibration and how the configuration can be modified to help manage it. The remaining sections cover analysis of the drilling riser in various conditions such as disconnected, connected, during emergency disconnect, and as recoil occurs after disconnect. Sample riser analysis results are reported in this chapter for various water depths as deep as 2700 m or 9000 ft. These results are taken from the analyses done for specific sites for which data are available.
- 68. Drilling and Production Risers zyxwvutsr 715 z 9.2.1 Design Philosophy and Background zyxwv To assess whether bending or riser tension dominates, the following non-dimensional number [Moe, 20041 T,L~ h.v,tens zyxwvutsrq = may be used. For equal to 1, the stiffness contribution from the bending and tension stiffness will be about the same, while for larger values the tension stiffness will dominate. Here Torepresents the average tension, L the riser length, E l is the bending stiffness and z y1 the number of half waves. The effects of tension and bending stiffness are both typically included in the riser analysis, and in the water depth of interest, tension dominates the stiffness. 9.2.2 Influence of Metocean Conditions The selection of accurate metocean conditions for a specific site for use in the analysis of a drilling riser is usually difficult, but it can, sometimes make the difference in whether or not a well can be drilled economically. The drilling riser is analysed based on the collection of wind, waves, and the current profile conditions for a specific well site. These metocean conditions can be based on information for a general region or an area near the well site. Whatever the case, a common understanding of the basis for the metocean conditions between the metocean specialist and the riser analyst is an important part of the process. The current profile often drives the analytical results used for determining when drilling operations through a riser should be shut down. The steady current loading over the length of the riser influences the riser deflections, and the top and bottom angles that restrict drilling operations. Furthermore, high currents cause vortex-induced vibrations (VIV) of the riser, which lead to increased drag load and metal fatigue. Current profile data at a future well site can be more difficult to collect than data on winds and seastates due to the large amount of data to be gathered throughout the water depth. Furthermore, current features in many regions of the world tend to be more difficult to analyse due to a lesser understanding of what drives them, particularly in the deeper waters. Winds and waves are important when considering the management of drilling riser operations in storms. Although not as important for determining the shape of the riser, the winds and seastates have a greater bearing on when the drilling riser should be retrieved (pulled) to the surface, Le. when the mooring system will be unable to keep the vessel within an acceptable distance of the well. Drilling risers are operated in conditions all over the world. These include large seastates off the east coast of Canada and the North Sea, the combination of high seastates and high currents west of Shetlands, the high currents offshore Brazil and Trinidad, and the cyclonic events combined with high currents in the Gulf of Mexico and offshore northwestern Australia. Typical metocean conditions for the Gulf of Mexico are listed below in table 9.1. 9.2.3 Pipe Cross-section The sizing of the pipe is important in order to assure the integrity of the riser for burst and collapse considerations. Collapse is generally checked to ensure the riser can withstand
- 69. Table 9.1 Typical design metocean criteria for Gulf zyxwvu of Mexico zyxwv 4 z u .z c z Riser connected/drilling Riser connected/non-drilling knots 87.5 103.2 ft 41.0 Winds Vwind (1 h) Vwind (I min) Seastate Hs TP (s) Mean T (s) Current TT- Surface 492 984 22.0 45.0 53.1 m 12.5 m 5.8 19.0 10.6 19.0 115.0 I I 8.2 6.9 11.6 zyx 7 1 I I _ _ ~ ~- m/s knots m/s knots 0.30 0.59 2.00 3.89 2.00 3.89 0.30 0.59 0.15 0.30 2.00 3.89 1.50 2.92 1.20 2.33 0.80 1.56 0.40 0.78 0.30 0.20 0.39 __ 0.15 ____ mjs 1.00 1. O O ____ 0.20 ___ knots I .94 I.94 ___ 0.39 knots m/s knots 2.72 0.30 0.59 2.72 0.30 0.59 0.15 0.30 -~ 2.72 2.14 1.56 1.17 0.58 0.39 0.15 0.30 ~ _ _ I I - knots 0.59 0.59 0.30 ____ m/s 0.30 0.30 0.15 1.40 1.10 0.80 0.60 0.30 0.15 0.30 0.20 0.20 0.39 Near bottom NOTES: 0 .2 s - z 2 - i ~~ Drilling can be conducted with mud weights of up to 16 ppg mud. Depcnding on the silc-spccific current conditions, drilling could be limited for certain mud weights ~ Somc level of vortcx-induced vibration (VIV) could he expericnccd in the eddy conditions. Depending on the sitc-specificcurrent conditions, vortcx-suppression devices v) could he warranted
- 70. Drilling and Production zyxwvutsrqp Risers zyxwvutsrq 717 exterior pressure due to a specified voided condition in the riser, while burst is checked to ensure that the riser can withstand the interior pressure from the drilling fluid (mud). The bore of the wellhead housing generally dictates the bore (inside diameter) of the riser pipe, and resistance to collapse and burst pressures generally dictates its wall thickness. z 9.2.3.1 Wellhead Housing The oil and gas industry has generally selected a few standard bore sizes for its subsea wellhead housings. These wellhead bore sizes include 18-3/4411,, 16-3/4411. and 13-5p-h The selection of the bore size determines the size of the casing strings that can be run through the wellhead and hung off in the wellhead housing. The most common of these in use today is the 18-3/4411,wellhead. With this wellhead size, the drilling riser inner diameter should be greater than 18-3/4-in., so most risers have a 21-in. (or, in some cases, 22-in.) outer diameter, leaving enough margin for the variable riser wall thickness that may be necessary for deeper waters. 9.2.3.2 Burst Check For the burst check, the water depth, the highest mud weight, the fabrication tole- rances and the yield strength of the pipe are used to determine the minimum wall thick- ness of the riser. API Bulletin 5C3 (1994) is commonly used as the basis for this calculation. 9.2.3.3 Collapse Check The riser must have sufficient collapse resistance to meet the conditions imposed by the operator. For an ultra deep water well, typical conditions call for collapse resistance sufficient to withstand the riser being void over half its length. This requirement usually covers the case of emergency disconnect in which a column of 17-ppg mud falls out of the bottom of the riser and momentarily becomes balanced with the pressure of seawater after the pressure has been equalised. In shallower water (less than 6000 ft), larger lengths of gas-filled riser may be required based on the risk of other events such as gas in the riser or lost returns. A number of design conditions can be considered when engineering the riser to resist collapse. Among others, these can include the following: 1. A gas bubble from the formation enters the well and expands as it enters the riser. The likelihood of a gas bubble filling the riser in a modern drilling operation is remote. However, it did occur once in 1982 [see Erb, et a1 19831.When this incident occurred, the subsea blowout preventer (BOP) was not shut-in when the flow was detected due to concerns about formation integrity. The surface diverter was being used to direct the flow overboard when it malfunctioned, causing loss of the mud column in the riser. In a modern drilling operation, the likelihood of riser collapse is greatly diminished because the shut-in of the BOP is a standard procedure when dealing with a kick. Returns are lost to the well, leaving a void on the top of the riser. The voiding of a large portion of the riser due to lost returns is a remote possibility. A large amount of lost returns would likely be detected. The contents of the riser (mud) are partially lost during an emergency disconnect of the riser. The u-tube that would occur during an emergency disconnect would typically leave no more than about 50% of the riser tube void after the pressure is equalised, if 2. 3.
- 71. 718 zyxwvutsrqpon Figure 9.5 Riser collapse profiles (22 in. x 1.125 in. plus 8% machine tolerance) zy Chapter z 9 z the mud weight were about 17 lb/gallon (twice that of sea water). The lesser mud weights would void less of the riser. API Bulletin 5C3 (1994) is commonly used as the basis for selecting the wall thickness to resist collapse. The calculation depends on the voided depth of riser, the yield strength of the pipe (in some cases) and the fabrication tolerances of the pipe. Collapse calculations using API 5C3 demonstrate that a 22-in. riser with 1-1/8-in. wall thickness resists collapse, if it is completely void in 9000 ft of water. With fabrication tolerances of 8% on wall thickness, the riser resists collapse with the top 8000ft of riser void. Figure 9.5 shows the external pressure resistance of the riser with an 8% fabrication tolerance vs. depth compared to the applied pressure from the hydrostatic head of seawater. The riser’s collapse resistance varies with depth due to a dependence on pipe wall tension. For various wall thicknesses of 21-in. risers and for various pipe wall tensions, calculations of water depth ratings of a voided riser pipe have been done based on the API 5C3. The results are shown in fig. 9.6. These curves are based on a “no margin” for fabrication tolerances. 9.2.4 Configuration (Stack-Up) This section covers the issues considered in determining how the drilling riser is configured, or its “stack-up”. The key issues in the riser stack-up are to assure the riser is heavy enough to be deployed without excessive angles in the currents expected during deployment and to assure the weight of the riser and Blow-Out Preventor (BOP) is within the hook load capacity of the vessel.
- 72. Drilling and Production zyxwvutsrqp Risers zyxwvutsrq 719 8000 7000 6000 zyxwvutsr E n 5000 Q b 4000 zyxwvutsr CI zyxwvutsrq 2 zyxwv 2 p 3000 2000 zyxwvuts c 1000 0 0 200 400 600 800 1000 1200 1400 1600 1800 2000 Pipe Wall Tension (kips) Figure 9.6 Riser collapse ratings (21 in. nominal wall thickness) 9.2.4.1 Vessel Motions and Moonpool Dimensions zyxw The vessel response amplitude operators (RAOs) used in the riser analysis can either be analytical calculations or estimates derived from the model tests. These RAOs are converted into the format required by the riser analysis program. In cases in which the vessel is not in a head seas or beam seas heading, planar riser analysis programs require that the surge and sway motions be combined. Typical vessel dimensions used for an ultradeep water drillship riser model are as follows: Upper Flex Joint Centre above Water Line - 63 ft. Drill floor above Water Line - 85 ft. Vertical Centre of Gravity (VCG) above Baseline of Vessel (Keel) - 47.55 ft. Draft of the Vessel - 29.5 ft. Height of the BOP Stack - 63 ft. Height of the BOP Stack from Wellhead Connector to Centre of the Bottom Flex Joint - 55 ft. The terms used above will be illustrated in figures in the upcoming sections. 9.2.4.2 Connection to Vessel The arrangement of the riser through the moolpool is shown in fig. 9.7. The riser is supported by the vessel through the combination of a tensioned telescopic joint and a top flexjoint in an opening in the vessel called the “moonpool”. The telescopic joint has an inner
- 73. 720 zyxwvutsrqpo Chapter zy 9 z L.I.~~ z Figure 9.7 Vessel moonpool and riser arrangement barrel and an outer barrel that allow vertical motion of the vessel while holding the riser with near-constant tension. The tensioning ring at the top of the “outer barrel” of the telescopic joint provides the connection point for riser tensioner lines, which maintain relatively constant tension through their connection to the compensating tensioner units. Top tension variation is minimised through the use of tensioner units that are based on a hydraulic/pneumatic system with air pressure vessels providing the springs. The tensioner lines wrap over “turn-down’’ sheaves located just under the drill floor. These tensioner lines route back to the tensioner units that are located around the perimeter of the derrick. The upper flex joint is located above the “inner barrel” of the telescopic joint where it provides lateral restraint and reduces rotation through elastomeric stiffness elements. A diverter located just above the upper flex joint and just below the drill floor allows mud with drill cuttings returning from the well through the riser annulus to be dumped to a mud processing system. A closer view of this arrangement is shown in fig. 9.8.
- 74. Drilling and Production zyxwvutsrq Risers zyxwvutsrq 721 z Figure 9.8 Riser upper flex joint, diverter, and turn-down sheaves 9.2.4.3 Riser String The riser string consists of “joints” (segments) of riser pipe connected at the drill floor and “run” (deployed) into the water. Figure 9.9 shows a typical ultra deepwater riser joint that is 75 ft long and has a continuous steel riser pipe down the middle. As shown, this riser joint has five pairs of buoyancy modules strapped on the outside and flange-type connectors at each end. As discussed below, the riser joints carry auxiliary lines, and thus are made up with bolted flange, dog-type or other non-rotating connections. The cross-section of a typical riser joint is shown in fig. 9.10. This figure shows auxiliary lines that are clamped to the riser pipe. These lines include choke and kill lines that provide for well control, a riser boost line that can be used to pump mud into the riser annulus just above the BOP stack to improve return of cuttings, a spare line, and a hydraulic line that controls subsea functions. Buoyancy material is shown strapped on the riser and external slots are provided in the buoyancy for attachment of multiplex (MUX) control cables. Figure 9.9 Typical riser joint
- 75. 122 zyxwvutsrqpo Chapter z 9 z Figure 9.10 Typical riser joint cross-section 9.2.4.3.1 Riser Joint Properties Riser joint properties include their weights in air, in water, with buoyancy, and without buoyancy. These weights can vary as the joints are deployed in deep water due to compression of the buoyancy and water ingress. Other properties include the joint length and the hydrodynamic properties such as drag diameter, drag coefficient, inertial diameter and inertial coefficient. Typical values for the joint properties used in an ultra deep water riser model are shown below in table 9.2. 9.2.4.3.2 Riser Stack-Up The riser stack-up consists of joints with lengths typcially ranging from 50 to 75 ft, depending on the drilling rig. Table 9.3 below shows the weight of each component in a riser string for a typical ultra deepwater drilling rig. Each component listed has its submerged weight listed, with the exception of the tensioner ring, which is expected to be above the water line. The total weight of the riser without the LMRP is used for determining the top tension required to support the string. The total weights of the riser with the LMRP and with the full BOP are used to determine the hanging weight of the string. The considerations in the joint stack-up of a riser string include assuring the riser is heavy enough to be deployed without excessive angles in the currents expected during deployment, and to assure the weight of the riser and BOP is within the hook load capacity of the vessel. This weight is regulated by bare joints or partially-buoyant in the string. The bare joints are often placed at the bottom of the string to get full benefit from the weight as deployment of the string first starts. Due to other considerations, such as VIV due to high currents, the bare joints may be placed in the region of high current often near
- 76. Table 9.2 Typical ultra deepwater riser joint properties zyxwvu In-water weight of bare joint zyxwvutsrq (Ibs)' In-air weight of ioint w/buoyancy ( I ~ s ) ~ Properties 30,975 30,975 57,724 60,199 zyxwvut IIn-air weight of bare joint (Ibs)' 135,644 135,644 IJoint length (ft)' I75 I75 Ih a i r weighl/length of bare joint (lb/ft)2 1475.3 1475.3 Ih a i r weight of buoyancy on joint (Ibs)' 122,080 124,555 INet lift of buoyancy on joint (Ibs)' 130,330 130,565 IIn-water weight of joint w/buoyancy (lbs)' I645 1410 IBuoyancy compensation' 197.92% 198.68% Drag coefficient' 156.5 I1.00 IInertial diameter (inches)' zyxwvuts I55.5 156.5 IInertial coefficient' 12.00 12.00 22 in. x 1.125 in. 22 in. x 1.125 in. 1.125-in. Wall w/59.5 in., Wall w/60 in., Wall bare 7.5 k buoyancy I10 k buoyancy ijoint 35,644 135,644 135,644 75 175 175 475.3 475.3 475.3 35,000 27,920 30,975 30,975 30,975 ~~~ 66,789 640 70,644 zyx tiosi- 30,975 97.93% 190.14% 59.5 160.0 141.3 1.oo 11.00 I1.00 59.5 160.0 137.5 2.00 12.00 12.00 1 ~ information provided 2 ~ h a i r weighi dividcd by joint length zyxwvutsrqponm 3 ~ In-water wcight ol' barc joint cquals 0.869 timcs in-air wcight of bare joint 4 In-air weight of joint w/buoyancy is in-air wcight of buoyancy plus in-air weight of bare joint zyxwvutsrq 5 ~ In-water weight of joint with buoyancy is in-air weight of a bare joint minus net lift ol'buoyancy 6 ~ Buoyancy compensation is (in-watcr wcight of bare joint minus in-water weight of joint with buoyancy) divided by in-water weight of bare joint 4 N w
- 77. Table 9.3 Installed weight zyxwvuts of riser string in 9000 ft of water zyxwvu 45,240 lb 9667 lb v z N z a 45.24 kips 9.67 kips 1 36 28 33 13 7 1 20 ft 20 ft 75 ft 2700 ft 75 ft 2100 ft 75 ft 2475 ft 75 ft 975 ft 75 ft 525 ft 15 ft 15 ft - - 1 40 ft 40 ft w/BOP 9010 ft w/LMRP 671.03 kips w/LMRP w/o LMRP 7540.05 kips 7314.45 kips In-air In-water Unit weight Totdl weight Equipment supported by tensioners Tensioner ring* 55,000 Ib 55.00 kips 39,I50 Ib zyx - - t - 39.I5 kips Slipjoint outer barrel Middle flex joint 10-ft pup joint 15,658 Ib 115.66 kips 13,607Ib 113.61 kips 20-ft pup joint Joint with 3000-ft depth buoyancy 57,724 Ib 12078.06 kips 645 Ib 23.21 kips 410 lb 640 Ib 21.11 kips 3055 lb Joint with 5000-ft depth buoyancy 11685.57 kips 60,199 Ib 66,789 Ib 2204.06 kiss Joint with 7500-ft depth buoyancy Joint with 10000-ft depth buoyancy - 70,644 Ib I918.37 kiss Bare joint with 1.125-in. wall LMRP with one annular BOP 435,108 Ib [435.11 kips w/BOP 8040.75 kips 1w/LMRP 18970 ft Iw/o LMRP 18955 ft Hang-off ratio of in-water weight to in-air weight of string w/LMRP: Top minus bottom pipe wall tension (all in-water weights except in-air weight for riser tube): 1768.62 kips Bottom pipe wall tension in riser string above LMRP (3000k top): 1 2231.38 kip: * h a i r weights used
- 78. Drilling and Production Risers 125 z the top of the riser string in an alternating, “bare-buoyant’’ configuration. VIV and methods for mitigating it will be discussed in Sections 9.4 and 9.5. Another consideration to be discussed later in this chapter is riser recoil. When an emergency disconnect is carried out, the presence of bare joints in the string improves the behaviour of the riser string and thus increases the range of top tensions that allow the riser to meet specified performance criteria. The most important of these criteria are the avoidance of contact between the riser and the rig floor, the avoidance of slacking in the tensioner lines, and the avoidance of subsequent downward movement of the lower marine riser package (LMRP) causing contact with the BOP. These issues will be discussed further in Section 9.2.9. With a specified length of joints making up the riser, the riser string generally has to include one or two shorter joint lengths to make the string length match up with the water depth. For this purpose, shorter joints are employed just below the telescopic joint. Since the lengths of these pup joints get no shorter than 10 or 5 ft, the telescopic joint is generally not exactly at mid-stroke at a specific location. This inexact match-up becomes a consideration in both Section 9.2.8 on emergency disconnect and Section 9.2.9 on riser recoil. zy 9.2.4.4 Connection to BOP Stack At the seabed, the riser connects to the blowout preventer, or “BOP” stack, which provides subsea well control after the well has been drilled to a depth that warrants it. The lowest riser joint connects to a riser adapter on top of the BOP stack. This connects to a lower flex joint located inside the upper portion of the BOP called the lower marine riser package (LMRP). As will be discussed later, the LMRP can be disconnected from the BOP, and this is called a part of the emergency disconnect sequence (EDS). Just above the seabed, the BOP is landed on a wellhead that is connected to the surface casing. The BOP arrangement is shown in fig. 9.11. 9.2.4.4.1 Bottom Flex Joint At the bottom of the riser, a flex joint provides a connection to the BOP stack. This connection provides lateral restraint and resists rotation through elastomeric stiffness elements. The rotational stiffness improves the performance of the riser by reducing the bottom flex joint angle, thus permitting drilling in more severe conditions. 9.2.4.4.2 BOP Stack This discussion of drilling riser analysis procedures includes discussion of the BOP stack due its make-up (LMRP plus lower BOP), weight, height, and connection to the seabed. The weight of the LMRP and lower BOP are important when considering deployment and retrieval of the riser as discussed in Section 9.2.6, and riser recoil as discussed in Section 9.2.9. The height of the BOP determines the elevation of the riser’s bottom flex joint above the seabed. The connectors in the BOP and the loads passed through to the conductor pipe are an important part of the analysis of wellhead and conductor loading discussed in Section 9.2.7. Furthermore, analysis is often conducted to determine the load expected on each of the BOP connectors under a set of defined loading conditions.
- 79. 126 zyxwvutsrqp RISER zyxwvu JOINT WITH NO BUOYANCY FLEX JOINT LOWER MARINE RISER PACKAGE TOP OF WELLHEADzyxwvutsrqp Figure 9.11 BOP arrangement Chapter 9 9.2.4.4.3 WellheadlConductorlSoil zyxwvu The wellhead, conductor, and soil are also part of the drilling riser analysis procedure. Flexibility within these elements alters the behaviour of the riser. For example, soft soils would allow rotation of the BOP stack relative to the mud line. This would reduce the angle of the flex joint (relative angle between the riser and BOP stack) which would permit drilling with larger vessel offsets. The differences could be important, especially in considering limits for drilling or concerns with reaching the limits of the flex joint. As will be discussed in Section 9.2.8, the wellhead and conductor can become the first to exceed their allowable stresses in a drift off scenario associated with an emergency disconnect. In that case, the loads applied from the riser to the wellhead and on into the conductor pipe are calculated as part of the riser analysis methodology. The key properties that are included in this analysis are the rated capacity of the wellhead, the cross-sectional properties of the conductor pipe (typically the inner strings are ignored), and the p-y curves or the shear strength profiles of the soil. 9.2.5 Vortex-Induced Vibration (VIV) This section covers the subject of vortex-induced vibration (VIV) as it relates to a drilling riser. The details of riser VIV are covered later (see Section 9.4). Ocean currents can cause VIV of a drilling riser that can lead to costly downtime in a drilling operation and ultimately fatigue failure of the riser as discussed by Gardner and Cole (1982). Such fatigue failure in a drilling riser could result in detrimental effects such as costly inspection and repairs, loss of well control, and compromise of safety. In this text, VIV mitigation measures are considered to be part of configuring the riser.
- 80. Drilling and Production Risers 121 z 9.2.5.1 Calculation Methods zyxwvu Vortex-induced vibration can be calculated using the hand checks, computational fluid dynamics (CFD), and empirical methods. Each of these methods has their place, depending on the current profile being investigated and the level of rigor required. 9.2.5.1 .I Hand Checks Hand checks for calculating the VIV fatigue damage are most applicable when metocean conditions include currents that are constant with depth. Such conditions can exist in shallow-water locations where the current is driven by tides (e.g. the English Channel) or close to the mouths of rivers. When the current is constant with depth, VIV can be very severe. In these cases, the Strouhal equation can yield a good approximation that can be used to determine the VIV frequency. The amplitude can be estimated as being equal to say, one diameter, or some other value that could be derived from the work of Blevins (1977) or others. Using the mode shape associated with the natural frequency closest to the VIV frequency, the amplitude can be used to determine the curvature of the riser. This curvature can then be used to calculate bending stress which, together with the VIV frequency, can be used to determine a fatigue damage rate and a predicted fatigue life. 9.2.5.1.2 Empirical Methods High-current conditions in deep waters generally have large amounts of shear (i.e. current velocity that varies with depth). Such sheared currents are most important for the VIV riser analysis for locations in the Gulf of Mexico and offshore Brazil, Trinidad, the UK, and other high-current areas. Although uniform currents lead to the most severe vortex-induced vibration (VIV), sheared (change of velocity with depth) currents can also lead to VIV. Analysis techniques to predict VIV frequencies and amplitudes are often considered to be a part of a drilling riser analysis procedure. Although research on riser VIV has been ongoing for decades, predictions of VIV amplitudes in real ocean currents still have uncertainties. Empirical techniques for calculating VIV and the resulting fatigue damage have been developed by Vandiver (1998) and Triantafyllou (1999). Related work has been carried out by Fumes, et a1 (1998). Current profiles that cause the larger VIV amplitudes are those that have nearly uniform current speed and direction over large portions of the water column. If the current profile has a large amount of shear, the likelihood of VIV is reduced. 9.2.5.1.3 Computational Fluid Dynamics Computational fluid dynamics (CFD) is another alternative for calculating the vortex- induced vibrations of a riser. This technique simulates the flow of fluid past the riser, models flow vortices, and predicts the riser motions. CFD techniques are under development with the objective to better model the physics, but the method requires large amounts of computer time to simulate VIV of a full length deepwater riser. A simplified analysis using two-dimensional CFD “strips” to represent fluid-structure interaction has been investigated by Schultz and Meling (2004).
- 81. 728 zyxwvutsrqpon Chapter zy 9 9.2.5.2Detrimental Effects zyxwvu In VIV induced by high currents, a drilling riser vibrates normal to the flow up to an amplitude of about one diameter, or 50-60 in., since the buoyancy outer diameter must be included. For a drilling riser in high currents, the period of the vibration can be in the range of 2 to zyxwvuts 4 seconds, based on the Strouhal equation which shows the frequency (period) linearly dependent on diameter and current speed. The detrimental effects of VIV are two-fold, drag force amplification and fatigue. 9.2.5.2.1Drag Force Arnplijication VIV causes an increase in the drag force on the drilling riser. The effective drag coefficient may be up to twice the value of a riser that is not experiencing VIV. 9.2.5.2.2 Fatigue Due to VIV Due to the vibration of the riser, alternating bending stresses cause an accumulation of fatigue damage. As a general statement, the most fatigue damage in the riser tends to occur near the bottom or near the top depending on the depth of the current profile. High damage occurs at the top due to the proximity of the current profile; and high damage occurs at the bottom because the effective tension in the drilling riser is low, leading to short bending modes with high curvature. The fatigue of risers due to VIV has been addressed later. 9.2.5.3 VIV Suppression/Management The metocean criteria (including current profiles) specified by the operator is used to determine if vortex suppression devices such as fairings might be needed to reduce drag force on the riser and suppress VIV. Because of the uncertainties in predicting VIV, this decision is sometimes made using site-specific analysis conducted by the operator and, at times, independent analysis using different methods. Fairings are an expensive option due to the cost of the fairings themselves and the additional rig time required to install them during riser running. Less expensive alternatives include strakes, alternating bare and buoyant joints [Brooks, 19871, and simply increasing the riser tension. The less expensive alternatives are not as effective, but can be adequate in many instances. 9.2.5.3.1 Stack- Up Adjustments The choice of where fairings are to be installed in the riser stack-up (Le. the description of joint properties along the string) has a large influence on the cost-effectiveness of well drilling operations. Fairings have been shown to be very effective. They can reduce drag force to as low as one-third of its original value and they suppress VIV almost entirely - provided they cover the portion of the riser where the high currents are predicted to be incident, This estimate of where the current is present in the water column may be highly uncertain, As a further complication, once the fairings are installed, removing or rearranging them would involve pulling (retrieving) the riser - a procedure that could take several days. Furthermore, the notion of placing fairings over the full length of a deepwater riser (i.e. greater than 3000-ft)is cost-prohibitive. Generally, at sites with severe currents, operators have chosen to put fairings over the top portion (500 ft or so) of the riser to cover the most likely high current events.
- 82. Drilling and Production Risers 129 z Strakes are external ribs placed on the riser string, most commonly in a helical shape. When compared to the fairings, these devices are less effective, but are still good at VIV suppression. They allow amplitudes of vibration with 10-30% of a diameter. z A disadvantage of strakes is their 30-50% additional drag force when compared to an unsuppressed riser. Typically, strakes can be installed on the riser joints prior to running (installing) the riser, thus minimising the high costs associated with additional rig time. The concept of using alternating bare and buoyant joints in the riser string (staggered joints) has been documented in Brooks (1987) as a means for reducing the VIV amplitude. This technique also provides a slight reduction in drag force. This is a popular technique because it involves no preparation by rig personnel other than to have bare joints available and sequenced properly. One disadvantage is that bare joints are required, usually near the surface, where their weight cannot be used to full benefit in running the riser. Additional discussion on this subject may be found in Sections 9.4 and 9.5. 9.2.5.3.2 Operating Tensioi? Instead of altering the riser stack-up, VIV suppression can be achieved by increasing the operating tension. The concept of this suppression method is to excite lower modes of the riser, which have longer mode lengths. As a result, curvatures and stresses are lower and fatigue damage is reduced. An advantage of this technique is that it helps no matter where the currents are in the water column and it has virtually no effect on the well drilling operation, since the riser does not have to be pulled. However, this technique often has little effectiveness, particularly for a dynamically-positioned vessel requiring emergency disconnect. In these vessels, riser recoil considerations during emergency disconnect usually dictate that maximum riser operating tensions are not significantly higher than the minimum riser operating tensions required to conduct well drilling operations. The margin for increased tension is thus quite small. Suppression devices may not be necessary if an operator can show that the metocean conditions will not involve high current during the drilling of the well. For example, presently low activity of currents could be used to justify a forecast of low activity for the duration of a well; and this could justify use of an unsuppressed riser. However, loop currents and related or unrelated deep ocean currents are still difficult to predict. Currents that are deep in the water column, whether driven by the loop current or other mechanisms, are particularly difficult to predict (or manage VIV suppression) with any certainty. A disconnect of the riser due to VIV in high currents is generally avoided, if at all possible. Such a disconnect event in high currents would result in the riser taking on a large angle and possibly contacting the side of the moonpool. If the bathymetry allows, the vessel could be allowed to drift toward deeper water to manage the riser angle and avoid contacting the seabed. If a disconnect does occur in high currents, it will likely be due to an emergency disconnect or a planned disconnect to protect the integrity of the wellhead connector and the conductor pipe.
- 83. 730 zyxwvutsrqpon Chapter 9 9.2.5.3.3 On-Board VIV Measurements The detection of VIV-induced alternating stresses in the riser pipe wall and the asso- ciated fatigue damage can be done using a variety of systems. The sensors that are used to measure VIV will not be discussed in this text. The two main categories of systems used to gather information on riser VIV are the so-called “real time” system and the so-called “flight recorder” system. zyxwvu As the name suggests, the real-time system gathers, analyses and displays VIV data virtually immediately after the riser undergoes the response. The flight recorder system gathers and stores the data until the riser is pulled, at which time the stored data can be removed for analysis. The real-time system provides data so that, if desired, it can be used to base operational decisions on management of the riser. This system generally involves a more complex measurement system, possibly with cables that need to be installed as the riser is being run. The flight-recorder system provides data only after the riser has been pulled, so that the data cannot be used to support operational decisions; it is intended more for the support of inspection decisions or VIV research. This system involves independent canisters mounted at selected locations along the riser. 9.2.6 Disconnected Riser This section covers the response of the drilling riser when its bottom is in a disconnected condition.This condition can occur during running (deployment or installation) of the riser or during pulling (retrieval) of the riser. Additionally, the riser can be in this condition when the riser has been disconnected for operational reasons. An understanding of the riser’s response in this condition is important to avoid damage to the riser and components on or around the riser that could lead to expensive repairs or ultimately loss of the riser or a compromise in safety. 9.2.6.1 Lateral Loading The lateral force applied to a drilling riser causes it to move into a deflected shape. This shape depends on the distribution of the in-water (submerged) weight of the string, including that of the lower marine riser package (LMRP) or the full blowout preventer (BOP) that are on its bottom. The shape also depends on the current profile being experienced and the lateral velocity of the drilling rig. The effects of weight and drag force plus remedial measures such as “drift running” (to be discussed later in this section) and tilting of vessel determine how well the riser can be deployed in the presence of high lateral loading. 9.2.6.1.I Lateral Response zyxwv During DeploymentiRetrieval Lateral response of the disconnected riser string is based on how close to vertical the riser string is at the critical stages of deployment. At the start of deployment, the motion of the BOP and the angle of the riser are important, as the BOP is being deployed into the waves and current. Riser analysis can be used to determine the likelihood of contact between the BOP and the side of the moonpool. As the riser is lowered further, strong surface currents can cause a large angle of the riser where it passes through the diverter housing (the opening in the drill floor). If the angle
- 84. Drilling and Production Risers zyxwvutsrq 73 z 1 becomes large enough, the riser can contact the side of the diverter housing, causing damage to the buoyancy material or causing the riser to become stuck so that it cannot be further deployed or retrieved. As more of the riser becomes deployed, top angles generally reduce, provided the ocean currents are primarily at the surface. However, currents at mid- depth or near the bottom can cause excessive angles leading to problems similar to those noted above. In addition, these currents can cause problems in landing the full BOP or, in particular, the lighter LMRP. The response discussed above is governed primarily by the drag properties (drag diameter and drag coefficient) of the riser, the riser's distribution of in-water weight, and the bottom weight of the BOP or LMRP. The drag force on the riser can be considered as proportional to the velocity squared according to Morison's Equation [see Krolikowski and Gay, 19801, so that the shape of the riser depends heavily on the current. Considering a minimal current, a riser that is negatively buoyant above the BOP will tend to take on an approximate catenary shape in the absence of current. This will lead to a bottom angle that is larger than the top angle. By contrast, if the riser is positively buoyancy, it will take on an approximate inverse catenary shape with the top angle larger than the bottom angle. The weight on the bottom, either that of the BOP or the LMRP, determines the straightness and the average angle of the riser. The same deployment considerations also apply to retrieval. When unlatching a drilling riser at the seabed, ocean currents can cause the riser to take on a top angle that prevents it from being pulled or run back down. In a planned disconnect, this situation can be avoided by using the riser analysis to predict the response. However, in an emergency disconnect that can occur on a dynamically-positioned drillship, no control exists over the metocean conditions in which the disconnect occurs. In this case, the vessel is generally maneuvered to manage retrieval of the riser. 9.2.6.1.2 Deployment/Retrieval Limits The limits that apply to the deployment and the retrieval process described above depend on the riser and rig equipment. The top angle limits depend on the inner diameter of the diverter housing and the outer diameter of the foam buoyancy on the riser. As a new riser joint is brought in and connected to the top of the string, the weight of the string is transferred to the lifting gear located, say 50-75 ft above the drill floor. zyx As this occurs, the riser deflects about this high pivot point in response to the current. A deflection equal to the undeflected radial gap between the riser and the diverter housing causes contact. Typically, the top angular limit for contact in this configuration is about 0.5". As shown in fig. 9.12, when the riser is landed in the spider at the level of the drill floor, the top angular limit of the riser depends on the radial gap between the riser and the diverter housing. In this figure, the riser is shown contacting the top and bottom sides of the diverter housing with an angle of 6.87". Typically, the riser is centred at the drill floor and the limiting angle for contact at the bottom of the diverter housing is more like 3". When compared to the configuration with the riser suspended from the lifting gear, the angular limit is larger with the riser landed in the spider because the string pivots about a point that is much lower. The contact again occurs against the side of the diverter housing, which is say 15 ft below the drill floor. The riser can be landed in the spider during high currents,
- 85. 132 Chapter zy 9 z - ,# zyxwv 6 0 2 zyxw Figure 9.12 Riser clearance in diverter housing zyxw without the need to run or pull. In this case, the limit might be compressive damage to the buoyancy or overstressing of the pipe. Another limit that applies to the deployment/retrieval process is the geometrical limit associated with the BOP or LMRP contacting the side of the moonpool. Also at the final stage of deployment, the angular limits dictate whether the LMRP can latch up to the lower BOP or whether the BOP can latch up to the wellhead. 9.2.6.1.3Application of Tensioned-Beam Analysis A variety of tensioned-beam analysis programs can be used to estimate the response of a riser during deployment or retrieval. Static, frequency-domain or time-domain riser analysis programs can be used, depending on the amount of detail needed. 9.2.6.1.4 ‘tDriytRunning” Solution In various parts of the world such as the Gulf of Mexico, Trinidad, and Brazil, deepwater drilling operations can be interrupted by lateral loading in high currents, particularly while running the riser. To counter this, a “drifting running” procedure is used for running the drilling riser in high currents. In this procedure, a dynamically-positioned vessel drifts towards the well in the direction of the current as the riser is run. This process allows the riser to be run in higher currents than would otherwise be possible and avoids rig downtime while waiting for the current to subside. Running riser without drifting could lead to riser binding in the diverter housing, and could cause excessivestress in the riser pipe and damage to the foam buoyancy. Figure 9.13 shows the deflection of a riser during deployment with the ship stationary.
- 86. Drilling and Production Risers zyxwvutsr 733 z SHIP STATIONARY zyxw Figure 9.13 Riser deployment in high current zyxw - ship stationary When currents are high during the riser running operations, special equipment or procedures may be warranted to run the riser and land the BOP stack on the wellhead. The terms used in this section will apply only to the riser; however, similar procedures can be used for running casing. A procedure called “drift running” uses controlled down-current motion of the drilling vessel to pass the riser through the rotary and diverter housing. This procedure has been used throughout the industry to successfully land the BOP stack without damaging the riser or the running equipment. Figure 9.14 illustrates the reduced top angle that can be achieved through the use of “drift running”. When the riser string is exposed to high current, it takes on an angle. This angle is a function of the force applied by the current and the weight of the string. If the angle at the top of the string is excessive, the string will see high stresses or bind in the diverter housing, -CURRENT- Figure 9.14 Riser deployment in high current - drift running
- 87. 734 zyxwvutsrqpo Chapter z 9 preventing it from being run. Binding due to excessive side load or high stresses in the riser can occur (1) when the string is hung off in the rotary or (2)when the string is supported by the lifting gear. These two configurations are very different in terms of the forces applied to the riser and the effects of high current. When a riser is landed out in the rotary, an excessive angle can cause the riser to contact the side of the diverter housing. This can lead to high stresses in the riser and possible damage to the buoyancy material. The angle that causes contact with the diverter housing depends on the inner diameter of the diverter housing and the outer diameter of the riser buoyancy. When the riser is hung off in the rotary table, the consequence is excessive bending stress in the riser or damage to the buoyancy material. When a riser string is supported by the lifting gear, an excessive angle can cause binding that could prevent running the string. Passing the riser through the rotary table with an excessive angle could damage the buoyancy material by scraping it against the side of the diverter housing. In a more extreme situation, lateral forces can cause binding in the diverter housing as the force against one side of the diverter housing becomes so large that the riser cannot be run. The top angle of the riser that can lead to contact with the diverter housing is generally quite small. In a typical example, the top angle for contact is less than 0.4" considering a 6-in. gap between a centralised riser and the diverter housing just after bringing in a new 75-ft riserjoint. zyxwv As the lateral force associated with this contact increases, binding becomes more likely. Drift running involves a controlled drift of the vessel down a "track line" in the direction of the current at a speed that minimises the top angle. Ideally, a speed and track are chosen to minimise the top angle of the riser/casing string as it is being run. In practice, the proper speed can be selected through co-ordination between the captain and the crew on the rig floor. By observing the position of the riser string as it passes through the rotary, the crew on the rig floor can provide information to the captain that can be used to correct the speed and direction of the drift. In this manner, the riser string can be run in whatever current is present, provided VIV concerns have been addressed. For optimal efficiency in the drift running procedure, the vessel would need to pass over the wellhead just as the riser string has been fully run so that the BOP can be latched up. This requires an informed estimated starting point. The distance and bearing angle of the start-up location with respect to the wellhead can be calculated using an average current profile based on the best information available for current profiles along the track line. Allowances should be included in this estimated starting point to account for changes in the current profile and bathymetric features. Changes in the current profile can cause overshoot, coming up short, or being off line of the wellhead. In addition, bathymetric features such as escarpments, as shown in fig. 9.14, might require adjustments in the drift running program such as hanging off the riser string in the rotary during certain stages of drifting. As noted earlier, a relatively small angle (less than 0.4") could cause contact, just after bringing in a new riser joint when the riser is being supported and run by the lifting gear. Since no contact occurs up to a top angle of, say, 3.3" when the string is hung off in the
- 88. Drilling and Production zyxwvutsrqp Risers I35 z rotary, this is the configuration in which corrections can be made. When the string is hung off in the rotary, the vessel can slow down or possibly even move up current very slowly without damaging the string, depending on the current conditions. This should be done with care not to overstress the riser pipe or damage the riser buoyancy. This flexibility to slow down or move up current allows the BOP to be latched after corrections are made or when maneuvering over a well near an escarpment. The captain and the drilling superintendent can carry out the riser running operations and the landing of the BOP stack by estimating the starting point. Such an estimate can be developed with the intent of making the starting point estimate based on an initial measurement of the current profile. It is understood that the current profile will change during the operation, so allowances in the estimate are needed. The intent is to keep the riser in the centre of the diverter housing during running. Measurements of current can be obtained using current metres such as the acoustic Doppler current profiler (ADCP). ADCPs can be mounted on the ship and on the remotely operated vehicle (ROV) as it is being run, thereby providing current measurements over the full water depth. Measurements of current speeds and directions at the various depths can be used as guidance for the operations. Joint length and riser running speed (joints per hour) are the other inputs. These quantities are used to calculate the speed of running the riser string and should include testing of the choke and kill lines. An alternative to the drift running procedure is an equipment solution called the moonpool centering device [Gardner and Cole, 19821.The centering device is a movable structure that applies a force at several locations on the riser string as it is being run. Rollers on the centering device are used to allow the riser to pass. The centering device is intended to keep the riser string centred and vertical as it passes through the rotary and the diverter housing. The disadvantage of this concept is that the device tends to be a heavy and cumbersome. 9.2.6.1.5 Case History of “Drift Running” During February of 2001 in Trinidad, the Glornar Jack Ryan drillship experienced a block of submerged high current with a peak speed of 2.6 knots and more than 2 knots over a depth interval of 900 ft. This resulted in the riser/BOP running operation, which normally would require 2-3 days to run, requiring nearly 20 days to run. As part of this experience, the following procedure was developed for running riser in such severe conditions. Commence running the BOP when set up on DP at a location of 30 miles from the drilling location. Continue running the BOP on the DP mode until the drill floor informs the bridge of difficulties due to angle of the riser. When this stage is reached, take the vessel off the DP and drift while running joints. Make attempts to put the vessel back on the DP while making a riser joint connection and revert to drifting while running it. Anticipate a stage in which the vessel would have to be on continuous drift to run the riser. Carry out continuous calculations to ascertain the cut off point for running the riser and using the remaining water depths for recovery.
- 89. 136 zyxwvutsrqpon Chapter z 9 During the entire operation the DPOs will log the times for running each riser joint, the drift distance, the current metre data for the depth of the BOP, and the position of the riser/BOP relative to the moon pool. Based on the current profiles, estimate the depth at which the BOP will be below the high current and the riser angle will decrease. If this depth cannot be reached by the calculated cut off point, then recovery will begin. 9.2.6.2 Vertical Loading If the metocean conditions include high seastates while a riser is disconnected and hung off, vessel heave motion could cause dynamic, vertical loading in the riser. Such vessel heave motion could occur if high seastates occur when the riser is hung off in any of the following configurations: The structural response of a drilling riser that is hung off from a floating drilling vessel is a critical issue for drilling operations in ultra-deep water. A hung-off riser can be exposed to storm conditions prior to its connection to the wellhead or after disconnection. In ultra- deep water, the axial dynamics of the riser are driven by the riser’s increased mass and its increased axial flexibility when compared to a shorter riser. With these effects, vessel heave motion and wave and current forces cause riser tension variation, riser motions, and alternating stresses. If secured in a hang-off configuration, the riser can be put into a “hard” hang-off configurationin which it is rigidly mounted to the vessel or a “soft” hang-off configuration in which the riser is compensated. Brekke, et a1 (1999) describes the advantages and limitations of the “soft” hang-off configuration when compared to the “hard” hang-off configuration as applied to the Glomar Explorer drill ship at a site in 7718 ft of water when subject to winter storms in the Gulf of Mexico. The advantages include: The limitations of the “soft” hang-off configuration are as follows: during deployment or retrieval of the riser while the riser is secured in a hang-off configuration peak hang-off loads are minimised; compression in the riser is avoided; motion of the riser is reduced; riser stress variation is minimised. vessel heave motion does not exceed the stroke limits of the telescopic joint and tensioners on-board personnel are available to monitor/adjust the tensioners’ set point During the deployment or retrieval process, the riser is generally in the hard hang-off condition. 9.2.6.2.1 Peyformance zyxwvu During Hang-off Coizditions Structural analysis of an ultra-deepwater riser will show larger axial (vertical) dynamic response than a shallower water riser due to the influence of the riser’s additional mass and
- 90. Drilling and Production zyxwvutsrqp Risers I 3 1 z increased axial flexibility. Several computer programs are available within the industry for the 3-D time-domain riser analysis required for the combination of axial and lateral dynamic riser analysis. Brekke, et a1 (1999) shows that 3-D random wave riser analysis is needed to determine accurate riser response estimates. This analysis discussed the fundamental contributors to tension variation, including zyxwvu (1) mass of the riser string times the vessel's vertical accele- ration, (2)resonance at the axial natural period, and (3) lateral motions of the riser leading to additional tension variation. Random analysis is more accurate than regular wave analysis because it models the full spectrum of the seastate and thus avoids artificial response peaks near natural periods. In random analysis, a realistic random seastate is generated in preparation for the riser analysis. The typical riser simulation is run for 1000wave cycles, representing about a 3 h storm. In order to determine the results from this analysis, the peak and trough response of each parameter are determined as the maximum and minimum values that occurred during the simulation. If need be, this random analysis approach could be made more accurate by running multiple simulations and averaging the results or using statistical methods to obtain the extreme values. 9.2.6.2.2 Riser Model The riser computer model is based on the riser joint properties and riser stack-up listed earlier in tables 9.2 and 9.3. For the hard hang-off, the riser is connected directly to the vessel so that it heaves and moves laterally the same amount as the vessel, but it is free to rotate at the top flex joint. For the soft hang-off, the vertical motion of the riser is compensated, but it still moves laterally with the vessel. The riser is connected to the vessel through springs whose total stiffness depends on the stiffness of the tensioner system and the Crown-block Motion Compensator (CMC). The stiffness value also depends on the weight supported by the system (Le. whether the LMRP or the BOP is suspended) and how the load is shared between the tensioners and the CMC. No damping is typically assumed for the combination tensioner/CMC system because the tensioner recoil valve is assumed to be inactive. For the hydrodynamic model in the vertical direction, the riser is modelled with a tangential drag coefficient of 0.2 and an inertial coefficient of 0.1 along its length. The BOP'LMRP is modeled according to the dimensions of a horizontal plate consistent with its length and width and a vertical drag coefficient of 1.1. As noted earlier, vessel RAOs for seas approaching 45' off the bow (135" case) typically give the largest heave and lateral motions for this type of analysis. For the hang-off analysis, heave motions have the most significant influence on the results. 9.2.6.2.3 Metocean Conditions for Hang-Off Analysis Differing metocean conditions could be rationalised for analysis of the various riser configurations. For deployment or retrieval conditions, a seastate leading to lesser heave such as, zyxwvut 5 ft maximum vessel heave (this peak-to-trough heave (DA) occurs once during a 3 h seastate) may be consistent with the requirement for running riser as stated in a vessel's operating manual. For the storm hang-off configuration, extreme storm conditions
- 91. 738 zyxwvutsrqpo Chapter 9 (e.g. the 10-yr winter storm) may be required to accommodate the possibility of disconnecting and securing the riser in such conditions. In the deployment or retrieval configuration, as noted above, only the hard hang-off is generally analysed since the riser is either landed in the spider or supported on the traveling block by the lifting gear. For the storm mode, the riser can be analyzed for both the hard and the soft configurations. Hurricane conditions are not generally analysed, since the riser is expected to be retrieved and secured onboard the vessel during such events. 9.2.6.2.4 Design Limits for Hang-Off Analysis Design limits used in a typical analysis are as follows: Maximum top tension during deployment: 1500 kips (rating of the lifting gear). Minimum top tension during deployment: 100 kips (avoid uplift on spider or lifting gear with 100 kips margin). Minimum tension along riser during deployment: no explicit limit since momentary compression in the riser does not represent failure. (The consequences of compression are covered by motion/stress limits.) Maximum top tension in 10-yr storm: 2000 kips (rating of substructure, diverter, upper flex joint, and other components). Minimum riser tension during 10-yr storm: no explicit limit since momentary compression in the riser does not represent failure. (The consequences of compression are covered by the motion/stress limits.) Riser Stress: Per limits in API RP 16Q. Moonpool Contact: Avoid contact between the riser (intermediate flex joint) and the moonpool with a 10% margin based on the nominal riser position. Maintain a sufficiently heavy string to allow deployment and retrieval in a reasonable levels of current without binding in the diverter housing or contacting the moonpool. A heavier string also helps keep the riser from contacting the moonpool after disconnect z during a drift off and controls riser recoil response during emergency disconnect. 9.2.6.2.5 Interpretation of Analysis Results Riser analysis for the 10-yr storm conditions can be used to compare riser response in the “soft” and the “hard” hang-off configurations. For a typical ultra deepwater well, the first axial natural period of a hung off riser could be about 5 s. As noted earlier, the soft hang-off configuration with the LMRP is modelled using a spring that connects the top of the riser to the vessel. According to riser eigenvalue analysis, the soft hang-off configuration could have a first axial natural period in the range of 30-50 s. Riser analysis for the deployment and the storm hang-off conditions was conducted for the Glomar C. R. Luigs in 9000 ft of water to estimate peak loads with axial tension variation. For the riser deployment mode (riser deployment or retrieval), riser analysis is run to determine the tension variation expected with different riser buoyancy configurations.
- 92. Drilling and Production Risers 139 z As noted above, the design limits can be a maximum tension of 1500 kips based on the capacity of the lifting gear and a minimum tension of 100 kips established as a margin above zero tension. This analysis was done for the metocean conditions associated with the 5-ft vessel heave. For storm hang-off conditions, riser analysis shows that the “soft” (compensated) hang-off configuration has much less riser motion and tension variation than the “hard” (rigid) hang-off. Hard hang-off loads are slightly higher than the 2000-kip capacity of the substructure. The soft hang-off is the preferred option as long as the vessel heave does not exceed slip joint stroke limits and on-board personnel are available to monitor/adjust the tensioner set point. Within these limitations, the risk assumed with a soft hang-off is virtually identical to that assumed when the riser is in its connected configuration. For the deployment mode (riser deployment or retrieval), riser analysis was run to determine the tension variation expected with different riser buoyancy configurations. As noted above, the design limits are a maximum tension of 1500 kips based on the capacity of the lifting gear and a minimum tension of 100 kips established as the margin above zero tension. This analysis was done for the metocean conditions associated with the 5 ft vessel heave as previously described. Analyses were run for cases with an LMRP suspended on the bottom of the riser string and for cases with a BOP on the bottom of the riser string. Both of these cases are important because the LMRP case generally gives the lowest minimum tension in the riser and the BOP case generally gives the highest maximum tension in the riser. Riser buoyancy configurations with 2 bare joints, 5 bare joints, and 10 bare joints were run with the LMRP; and buoyancy configurations with 10 bare joints and 15 bare joints were run with the BOP. The results are used to determine the range of configurations that would satisfy the tension limits. The results of the deployment analysis are summarised in fig. 9.15. This figure shows the variation in riser top tension versus the number of bare joints in the riser, with a minimum, mean (riser string weight in water), and maximum tension curve shown for the LMRP cases on the left side and for the BOP cases on the right side. The minimum and the maximum allowable tensions (100-kip and 1500-kip limits defined earlier) are shown as horizontal dashed lines. Based on this figure, a riser buoyancy configuration with seven or less bare joints would satisfy design limits on maximum tension (with the BOP) and minimum tension (with the LMRP). Based on the considerations of in-water weight noted earlier, a number of bare joints less than seven would result in an in-water to in-air weight percentage less than 9%, so that seven bare joints is the optimal value. The riser analysis results shown in fig. 9.16 illustrate that riser tension variation during deployment is much higher at the top of the riser than it is near the bottom. This is mainly due to the dominance of inertial loading caused by the mass below each elevation along the length of the riser. Two pairs of curves are shown in fig. 9.16, with each pair made up of a minimum and a maximum tension curve. Each pair represents an extreme case, with the pair on the left representing the LMRP and two bare joints in the riser string, and the pair on the right representing the full BOP and fifteen bare joints. In both cases, the figure
- 93. 740 zyxwvutsrqpo Chapter z 9 z 2000T I ' I I Izyxwvu I I I I I 1 1 I I t 1800 1600 zyxwvutsr 21400 e1200 zyxwvutsrq szyxwvuts '5 1000 E $ 800 P zyxwvu 8 600 400 200 I I 1 I I I I - - 0 0 1 2 3 4 5 6 7 8 9 1 0 1 1 1 2 1 3 1 4 1 5 Number of Bare Joints in the Riser String *Maximum Tension (kips) --Max Allowable Tension (kips) - - Riser String Wt In Water (kips) - .Min Allowable Tension (kips) - A - Minimum Tension (kips) Figure 9.15 Riser top tension ,ariation and design limits during deployment Figure 9.16 Tension variation during deployment
- 94. DrillinR and Production Risers zyxwvutsr With LMRP 2 bare zyxwv 15 bare zyxw I10 bare 741 With BOP 10 bare 1 15 bare 1 Mean top tension I I kips kips kips kips kips 529 [ 614 755 11190 1327 I 1 1Max. toD tension I963 11042 (1163 11583 11710 1 Min. tension along length 1Min. tor, tension I184 I279 1437 1804 I980 1 15 110 I266 657 I836 shows that the top portion of the riser experiences much more tension variation, and stress variation, than the bottom portion. In this 5-ft heave condition, the hung-off riser with the LMRP and two bare joints comes close to compression in its upper portion and the hung-off riser with the BOP and fifteen bare joints experiences a top tension of 1750 kips. Table 9.4 shows the summary results for the four analysis cases presented for the deploy- ment configuration, including minimum top tensions, maximum top tensions, and minimum tensions along the length. Riser analysis for the storm configuration was done to compare the riser response in the “soft” and the “hard” hang-off configurations in a 10-yr winter storm. As noted earlier, the design limits on top tension used for the storm configuration are different from those used for the deployment Configuration. In this case, the hard and soft hang-off configurations were analysed with 10 bare joints and the LMRP on the bottom of the riser string. Due to the lighter hanging weight of the LMRP, this configuration is more prone to riser compression than the configuration with the BOP. The hard hang-off configuration is simply modelled with the top of the riser moving vertically and laterally with the vessel. In this configuration, the first axial natural period of this riser configuration is about 5 s. As noted earlier, the soft hang-off configuration with the LMRP is modelled using a spring that connects the top of the riser to the vessel. The soft hang-off configuration has a first axial natural period of about 45 s. Figure 9.17 shows the tension envelopes vs. depth along the riser string, with the LMRP only, for the hard and soft hang-off configurations. The envelopes show the minimum tension on the left side and the maximum tension on the right side. As shown, the envelope for the hard hang-off is much wider than that for the soft hang-off, indicating a large difference in tension variation between them. Additionally, the hard hang-off envelope shows a minimum tension that is below zero (in compression) at the top of the riser and over a large portion of its length. Although this is not considered a failure, it can lead to high bending stresses and lateral deflections. With the LMRP, peak top riser tensions are 750 kips for the soft hang-off and 1620 kips for the hard hang-off.
- 95. 142 zyxwvutsrqp Chapter 9 z 8000 7000 zyxwvutsrqp 6000 -5000 zyxwvutsrq 0 = zyxwv m 5 4000 a 3000 2000 Max Tension, Soft Hang-Off 1000 0 zyxwvutsrq -4 E+05 -2 E+05 0 E+OO 2 E+05 4 E+05 6 E+05 8 E+05 1 E+06 1 E+06 1 E+06 2 E+06 Effective Tension Envelopes (Ibs) zyxw Figure 9.17 Tension %ariation during 10-yr winter storm Figure 9.18 Riser vertical motion with hard hang-off, 10-yr winter storm Figures 9.18 and 9.19 show plots of vertical LMRP motion versus time for a portion of the simulation in which the peak heave motion occurred. For the hard hang-off (fig. 9.18), the peak LMRP motion is 1.23 times the vessel heave motion, which roughly indicates the level of dynamic amplification. For the soft hang-off (fig. 9.19), the LMRP motion is 0.04 times the vessel heave motion. Table 9.5 gives a typical results summary for the hard and soft configurations in the storm hang-off mode (IO-yr winter storm conditions) with the 10 bare joints and the LMRP. Next Page
- 96. Drilling and Production Risers 143 z Tensions zyxwvu 1kips Max. top tension 1620 Min. top tension 143 zyx E zyxwvuts 0 zyxwvuts .P 2 E O zyxwv c kips 750 732 -6 1 Max. heave amplitude ~ 9.2 Max. LMRP vertical ~ 11.3 amplitude i 700 1720 1740 1760 1780 1800 1820 1840 1860 1880 1900 Time (Sec) zyxw Figure 9.19 Riser vertical motion with soft hang-off, 10-yr winter storm 9.2 0.4 Table 9.5 Hang-off results for storm configuration, 10 bare joints and LMRP Max. Von Mises stress 124 1 Storm Configuration. 10 Bare Joints with LMRP [ 13.2 IMin. tension along length ~ -10 1(n/a) [ 1 Motions (double amplitude) Ift Ift 1 1Stress 1ksi lksi I This table shows the maximum top tensions, minimum top tensions, minimum tensions along the length, riser motions, and riser stresses. The peak tensions are consistent with the figures discussed earlier. This shows that for storm hang-off conditions, the soft hang-off configuration has much less riser motion and tension variation than the hard hang-off configuration. Related work has also been carried out by Miller and Young (1985) studying the effects of a column of mud contained in the riser during hang-off. Previous Page
- 97. 744 zyxwvutsrqpo Chapter z 9 9.2.6.2.6 Operational Proceduresfor Hang-Off zyxw If heavy seas are encountered during riser running or retrieval operations, typical procedures for going into the soft hang-off configuration (load shared between the tensioners and the CMC) are listed below. 1. 2. Engage the tensioning ring. 3. 4. Make up the telescopic joint in the riser string. Make up a landing joint. Lower the riser string until the tensioning lines support about half of the riser string weight and the tensioners are at mid-stroke. Activate the CMC and set it to support the other half of the string weight. Monitorladjust tensioner stroke and set point. zyxw 5. 6. After an emergency disconnect, assuming the vessel is moved off location per existing procedures, typical procedures for going into the soft hang-off configuration (on tensioners only) are as follows: 1. 2. 3. An alternate procedure that uses hard hang-off would call for installing the diverter, lifting the riser string with drill pipe, and locking the slipjoint so that the riser is supported under the drillfloor. A second alternate procedure for hard hang-off calls for landing the riser string in the spider; however, this configuration does not provide resistance to uplift (compression at the top of the riser). De-activate the riser recoil valve and open all Air Pressure Vessels (APVs). Reduce pressure on the riser tensioners until they are at approximately mid-stroke. Monitor/adjust the tensioner stroke and set point. 9.2.7 Connected Riser This section discusses the drilling riser in the connected configuration. In this configu- ration. the riser provides a conduit for drilling operations that guides the drill pipe and casing strings into the well and contains a column of drilling fluid (mud) for well pressure control and circulation of drill cuttings up from the bottom of the well. The assurance of riser structural integrity is provided by an understanding of the riser response in this configuration. Structural integrity is maintained under metocean conditions that include wind, waves, and currents that apply forces to the riser. The associated lateral motions from the vessel are also imposed at the top of the riser. In addition to the external forces and motions, drill string rotation and other operations impose wear and other degradation within the riser. Analysis of the connected riser configuration is routinely carried out to demonstrate that a rig's top tensioning capacity is sufficient to support the riser at a specific well site or in its design water depth, mud weight, and metocean conditions. In addition, if the metocean conditions include high currents, vortex-induced vibration (VIV) analysis (discussed in Section 9.5) can be carried out to further verify the riser's structural integrity.
- 98. Drilling and Production zyxwvutsrq Riser3 zyxwvuts 145 z 9.2.7.1 Performance Drivers zyxwvu The integrity of the connected drilling riser is largely driven by its deflected shape during the various operations that are carried out with it. During drilling operations, greater restrictions are placed on the riser’s deflected shape due to the need to rotate drill pipe or strip (run or pull) drill pipe through the drilling riser. When drilling operations are suspended, restrictions on the deflected shape of the riser are reduced significantly. Due to its length, the stiffness of the drilling riser is derived largely from its tension (similar to a cable), rather than its cross-sectional properties. In the absence of current, the mean deflected shape of the drilling riser is driven by the applied top tension, the mean offset at the top of the riser, the in-water weight of the drilling riser (“effective” tension gradient). A current profile applies force to the riser that further influences the mean shape. The dynamic motion of the riser is driven by the top motion of the vessel coupled with the fluctuating force resulting from the waves and current. Other factors such as end con- straints at the top and bottom of the riser also influence the riser’s mean shape and the dynamic motion. 9.2.7.1.1 Tensioned Beam Model Due to its length, the drilling riser is the most accurately modelled as a tensioned beam. The tensioned beam model combines the behaviour of a cable with the local stiffness of a beam. The equation for the tensioned beam is given later. 9.2.7.1.2 Concept of “Effective Tension” Due to the column of mud inside the drilling riser, differential pressure effects are accounted for in the tensioned beam model for a drilling riser. As discussed by McIver and Olson (1981), differential pressure caused by the mud has a profound effect on the shape of the riser. Instead of using the tension in the wall of the pipe, the “effective tension” includes the internal and external pressures as noted in the equation below. A simple calculation of the effective tension at any elevation along the riser can be carried out. The effective tension is the top tension minus the “weight” of the riser that is installed above the specified elevation. The “weight” of the riser is the in-air weight of the portion of riser (and contained mud) that is above the water and the in-water weight of the portion of riser (and contained mud) that is below the water. 9.2.7.1.3 Top Motion Drilling riser analysis includes vessel motions, since the top of the riser is connected to the vessel. The vertical motions, primarily due to heave, roll, and pitch, are not included in drilling riser analysis because of the motion compensation provided by the actions of the slip joint and the marine riser tensioners. However, lateral motions caused primarily by surge, sway, roll, and pitch are accounted for. The lateral motions imposed on the top of the drilling riser influence the direct wave and current forces applied to the riser by virtue of their phase with the waves. For example, the direct wave and current forces are relatively low if the motion of the vessel is “in-phase’’ with the water particle motions in the wave. This “in-phase’’ vessel response generally
- 99. 146 zyxwvutsrqpon Chapter 9 z occurs with surge in large waves. “Out-of-phase’’ response can occur with smaller, short- period waves and can lead to relatively high direct wave and current forces. zy 9.2.7.1.4 Hydrodynamic Loading The direct wave and current forces on the riser are calculated using formulas in Krolikowski and Gay (1980). zyxwvu A drag coefficient and a drag diameter are characteristics of the riser. Similarly an inertial coefficient and inertial diameter are also characteristics of the riser and are used in the formulas that determine the dynamics of the riser under the action of the current, waves, and top motion. 9.2.7.1.5 Rotational Stiffness zyxwvu -Top and Bottom Flex joints at the top and bottom of the drilling riser reduce the angle of the riser at its top connection to the vessel and at its bottom connection to the BOP. This local angle reduction provides a moderate reduction in angle that extends the conditions in which drilling operations can be conducted. The flex joint is a passive, elastomeric component, which has become popular for deep water. Riser flex joints are also used at an intermediate location at the elevation of the keel on dynamically positioned vessels. The purpose of a flex joint at this elevation is to prevent damage in case the riser is disconnected in high currents or while the vessel is drifting after an emergency disconnect. The purpose of the intermediate flex joint is to provide an articulation rather than restrict the angle with its stiffness. 9.2.7.2 Analysis of a Tensioned Beam Model Mean shape and dynamic motion of a drilling riser are calculated through finite element analysis of a tensioned-beam model. This analysis can be done using static analysis, frequency-domain analysis, or time-domain analysis. Static analysis can be accurate in cases in which no dynamics are expected. For steady-state dynamics, frequency-domain and time-domain solutions are alternatives that depend on solution time requirements, as described below. Time domain analysis is also used to simulate transient processes. 9.2.7.2.1 Time zyxwvu vs. Frequency-Domain Analysis Time-domain analysis generally provides the more accurate solution than frequency- domain analysis at the expense of more computational time. In time-domain analysis, the equations of motion are solved at each of many small time steps that are used to describe a process such as an extreme storm. Typically, an analysis models an extreme storm with 1000 wave cycles, which roughly corresponds with a 3-h duration. In the frequency-domain analysis, an extreme storm is described as a spectrum and the equations of motion of the riser are solved at each of many frequencies used to describe the process. The key approximation used in a frequency-domain approach is the technique for linearising any non-linear features in the process. For drilling risers, the most impor- tant non-linear feature is the drag force from the waves and current. A commonly used approximation for the drag force is described in Krolikowski and Gay (1980).
- 100. Drilling and Production zyxwvuts Risers 141 9.2.7.2.2 Coupled vs. Uncoupled Analysis zyxwv Traditional riser analysis has been performed in an “uncoupled” fashion in which the riser is considered to have no effect on the vessel at its top connection and no effect on the top of the BOP stack at its bottom connection. Usually, these effects are negligible and an uncoupled riser analysis is adequate. However in certain situations, the riser has an effect on the vessel or on the BOP stack that is considered in a “coupled” analysis. Coupling effect is generally the most important to consider in EDWdrift-off conditions. In these conditions, the riser can take on a large top angle and apply a significant lateral force to the vessel. It can also take on a large bottom angle and thus a significant lateral force to the top of the BOP stack that causes the BOP, wellhead, and conductor pipe to take on an angle. To accurately analyse the riser under the above conditions, a coupled analysis is required. At the top of the riser. the coupled analysis is carried out in combination with a vessel analysis program. As the vessel moves laterally away from the wellhead, the lateral force from the riser is applied as a restoring force, which reduces the speed of the vessel. This provides a more accurate estimate of the time available to disconnect the riser. At the top of the BOP stack, the lateral force from the riser causes the BOP, wellhead, and conductor pipe to take on an angle. This angle depends on the soil foundation properties, the conductor dimensions and the elevation of the top of the BOP stack. As the BOP angle increases, coupled analysis considers that the bottom flex joint angle allowable also increases since the “stop” of the flex joint has rotated. In addition, coupled analysis provides an accurate assessment of the loading on the conductor and wellhead. Although, uncoupled analysis generally provides a conservative assessment, the coupled analysis provides an assessment that has many of the unnecessary conservatism removed, particularly in soft soil conditions. zyxwv 9.2.7.3. Operational Limits 9.2.7.3.1 Minimum and Maximum zyxwvu API Tensions This section discusses the API guidelines that have been established for minimum and maximum tension. Minimum tension is established to prevent buckling of the riser. Maximum tension is established to prevent top tensions in excess of the installed capacity of the riser. To prevent buckling of the riser, criteria have been established within the industry to prevent the effective tension in the riser from going below zero. API RP zy 164 (1993) provides guidance on this, which provides a margin to account for uncertainties in the weight of the riser steel and the lift of the riser buoyancy. This margin also provides adequate tension in case a tensioner fails. API RP 16Q (1993) also distinguishes the rated capacity of a tensioner and the vertical tension applied at the top of the riser. (The ratio is often in the range of 90-99%.) All of these factors are considered in the calculation of the API minimum tension that is used to prevent buckling. In practice, the API minimum tension is rarely used as the riser’s operating tension. An added margin on tension is warranted to improve the riser performance in high seas or high currents, as will be discussed later in this chapter.
- 101. 148 zyxwvutsrqpon Chapter 9 z As discussed earlier in this chapter, the lowest effectivetension (usually at the bottom of the riser) is calculated as the top vertical tension minus the in-water weight of the riser plus the contained mud. The weight of the riser string and the mud column in the riser must both be supported by the tensioners to avoid riser buckling. To calculate the weight of the mud column, an estimate is made of the capacity (gallons/ft) of the riser pipe and the other lines (choke and kill lines and boost line) that contain mud. Table 9.6 shows how the in-air weight of the mud column above the water line and the submerged weight of the mud column below the water line are added to the string weight to determine the riser string weight with mud. The API minimum riser tensions are calculated using the installed weight of the riser with mud. The values calculated are vertical tensions at the top of the riser. For this calculation, the following information was used: Tolerances zyxwvu - 1% on the weight of steel in the riser and 1% on the net lift from the buoyancy material. Tensioners Down - Positive tension is maintained in the riser if one out of twelve tensioners goes down. Maximum Tension Limit - API RP 164 guidance is that top tension should be no more than 90% of the dynamic tensioning limit (same as rated tensioner capacity). This tension multiplied by a reduction factor for fleet angle only (in this example, the tensioner system compensates for mechanical losses, so that the estimate is 0.99) gives the maximum API tension in terms of vertical tension at the top of the riser. A maximum tension limit 90% of the installed capacity prevents the relief valves from popping under most conditions. In practice, a lower maximum tension limit is generally applied. Table 9.7 shows the calculation of minimum and maximum API tensions for a range of mud weights. Figure 9.20 shows a plot of the results. Table 9.6 and fig. 9.20 show a slightly higher tension than the API minimum tension at very low mud weights. In this range, a nominal tension (higher than the zy API minimum tension) is applied to the riser to assure that the riser can have a “planned” disconnect carried out successfullywithout increasing the tension. This tension is sufficient to support the in-water weight of the riser plus the LMRP (excluding the weight of the mud in the riser). A significant factor in proper tensioning of an ultra-deep water riser is compaction of the buoyancy material leading to a reduction in the net lift of the buoyancy. API RP 16Q uses the weight of the riser string, the weight of the mud column in the riser and the auxiliary lines, and tolerance values to determine the riser weight installed in seawater. API’s specified tolerance values of 5% on steel weight and zyx 4% on buoyancy net lift can be overridden if an accurate weight of the riser is taken during deployment. In one recent example, when comparisons were made to manufacturers’ values, weights recorded during deployment of a riser showed that the actual installed weight of the riser string can be matched by using 1% additional steel weight and slightly more than a 3% decrease in net lift due to buoyancy. Although this is within the API tolerance levels, when compared to
- 102. Drilling and Production Riser5 zyxwvutsrq Riser capacity (gallft) Seawater density (ppg) zyxw Table 9.6 Installed riser weight with mud 18.23 zyxw I , 8.55 Izyxwvutsrqpo 149 1Mudweight I I 1Riser string weight with mud in seawater Weight of mud in Weight of mud IWeight of seawater from in air from 'riser string flexjoint to waterline' waterline to lwith mud in diverter2 Iseawater3 'PPg 8.55 I 1Length (ft) (FJ to WL) 18940 I kip kip kip 0.00 7.79 482.77 I /Length (ft) (WL to DH) 150 I I9 173.34 I 1Wt. of riser string (kip) 1474.98 I 8.20 556.52 10.5 I1 317.80 9.57 ~ 802.36 399.29 10.03 '884.30 I9.5 1154.83 18.66 1638.47 113 13.5 14 14.5 15 15.5 16 110 1236.32 19.12 1720.41 725.24 11.85 11212.08 806.73 12.31 11294.02 888.22 12.76 '1375.96 969.71 13.22 1457.91 1051.20 13.67 1539.85 1132.68 14.13 , 1621.79 1214.17 14.58 i 1703.74 I 111.5 1480.78 110.48 1966.24 ~ 12 1562.27 110.94 ~ 1048.19 i12.5 1643.76 111.39 11130.13
- 103. Table 9.7 API riser tensions zyxwvuts - vertical load at slip zyxwvu ring zyxwvu In-water weight of bare joints Nct lift of 3 k buoyant joint Net lift of 5 k buoyant joint Net lift of 7.5 k buoyant joint Net lift of IO k buoyant joint Remainder of String wt. (excl. LMRP) __- 30.97 kips 30.33 kips 30.57 kips 30.34 kips 27.94 kips 162.66 kips Mud wt. Weight of Steel riser string weight with mud in tolerance' seawater 721 721 774 863 953 1042 PPg kip kip 8.55 482.77 37.87 __ 2673 2673 2673 2673 2673 2673 - 9 556.52 37.87 638.47 37.87 10 720.41 37.87 1.091 1.091 802.36 37.87 884.30 37.87 952.81 721.03 1042.21 721.03 loss/ slip ring 33.12 553.76 33.12 627.51 33.12 1709.45 33.12 1791.40 33.12 955.28 # of joints # of 3 k buoyant joints # zyxwvu c- of zyxwvuts 5 k buoyant joints _________ ____ # of 7.5 k buoyant joints I # of IO k buoyant points 1-250 k API min. rec. tensioner tension w/ 1- loss 250 k down factor4 (T,$ Tension required for disconnect' 1.091 [ 604.15 (721.03 1.091 1684.61 1721.03 1.091 774.01 zyxwv t 1.091 863.41 721.03 I117 zyx (Continued) 3 E 2
- 104. Table 9.7 Continued zyxwvut ~~ zyxwvutsrqponmlkjihgfedcbaZYXW Min. rec. tension7 1132 1221 1310 1400 - Maximum slip ring tension8 kip 2673 2673 2673 2673 Mud wt. 1966.24 11.5 137.87 133.12 ~ 1037.23 11.091 12 1048.19 37.87 33.12 1119.17 1.091 Weight of Steel Buoyancy Minimum 1-250 k riser string weight loss/ slip ring tensioner with mud in tolerance' tolerance2 tension3 loss seawater factor4 1131.62 API min. rec. tension w/l- 250 k down (Tm,")S 1221.02 Tension required for LMRP disconnect6 721.03 1489 I579 1668 1757 1130.13 37.87 33.12 1212.08 37.87 33.12 1294.02 37.87 33.12 1375.96 37.87 33.12 14.5 1457.91 37.87 33.12 2673 2673 2673 2673 - 1310.42 721.03 1578.62 721.03 15 15.5 16 1668.02 (721.03 1539.85 37.87 33.12 1610.83 1.091 1757.42 721.03 1621.79 37.87 33.12 1692.78 1.091 1846.82 721.03 1703.74 37.87 33.12 1774.72 1.091 1936.22 721.03 zyxw 1 ~ 1.0% In-water wt. of steel: O.Ol*(wt.of bare joints plus remainder of bare string) 2 ~ 1.0% Net lift of buoyancy: 0.01*(net lift from all buoyancy) 3 ~ In-watcr weight plus stccl weight tolerance plus buoyancy loss/tolerance 4 ~ Factor of 1.091 covers loss of one out of twelve 250-k tensioners zyxwvutsrqp 5 ~ Minimum recommended tensions that satisfy zyxwvutsrq API 16Q guidelines for buckling stability: 6 ~ In-water string weight with seawater plus in-water LMRP weight plus 50 kips 7 ~ Maximum value of 5 and 6 8 ~ 90% of dynamic tensioning limit (rated tensioner capacity) times reduction factor (0.99) Min. slip ring tension times tensioner loss factor 1847 12673 1936 12673
- 105. 152 zyxwvutsrqpon Chapter 9 z 3000 I I I I I I 2800 2600 2400 2200 zyxwvuts h zyxw g 2000 1800 zyxwvutsr E .g 1600 v) 5 1400 c L 1200 zyxwvutsrq 3 1000 800 600 400 200 0 9 10 11 12 13 14 15 16 Figure 9.20 API Riser tensions - vertical load at slip ring (Glomar C. R. Luigs - GoM 9000 ft) z the manufacturers’ values in 9000 ft of water, this can amount to 150 kips of additional weight for the entire riser string. 9.2.7.3.2 Riser Angle Limits API RP 164 has established riser angle limits for drilling and non-drilling operations with the riser connected. The basis for these is generally to minimise wear during rotation of the drill pipe and during tripping of the drill pipe. Angular limits are also necessary in order to conduct certain operations such as landing casing hangers and production equipment. When no drilling operations are being conducted, the limits can be relaxed to simply avoiding bottom-out of the flex joints. Figure 9.21 shows extremely large riser angles on a connected riser during high currents. Riser wear incidents have continued to occur in drilling operations, with several “keyseating” failures occurring near the bottom flex joint. The key measures for avoiding wear are adequate riser top tension and vessel positioning. The areas susceptible to wear are the inner surfaces of the riser and BOP stack, particularly near the bottom flex joint. API RP 16Q specifieslimits on the bottom flex joint angle and top flexjoint angle. During drilling operations, mean top and bottom flex joint angles of 2.0” are specified in API RP 16Q. In ultra-deep water, operationspersonnel generally use more restrictive targets for top
- 106. Drilling and Production Risers 153 z HIP STATIONARY -CURRENT- zyxwv Figure 9.21 Excessive Top and bottom angles on connected riser zyx and bottom angles, such as 0.5-1.0" due to the cost consequences of tripping the riser. For non-drilling operations, maximum riser angle limitations are generally zy 9", based on avoidance of flex joint bottom-out. 9.2.7.3.3 Stress Limits Riser stresses are checked during the riser analyses. Maximum stresses are generally limited to 67% of yield strength. This limit ensures that the maximum tension applied to the riser is within the capacity of the riser connector. In this check, axial, bending, and hoop stresses are considered. In addition to maximum stresses, alternating stresses are limited by a recipe given in the API RP 164. This recipe is intended to limit the fatigue damage in the connector and the riser pipe. Explicit fatigue analyses are often carried out to provide additional checks of the fatigue damage in a riser under wave loading conditions. As noted earlier, VIV fatigue analysis is carried out on risers to check the fatigue damage done under high current conditions. The fatigue damage done by VIV is generally consi- dered to be more severe than that done under wave loading. 9.2.7.3.4 Riser Recoil Limits ion DP vessels) The minimum top tension in a connected riser is often governed by riser recoil considera- tions whose limits are calculated through analysis. The top tension must be high enough to ensure that the LMRP will unlatch cleanly from the BOP during an emergency disconnect. The limiting value in such a process is the clearance between the LMRP and the BOP after disconnect, if and when the LMRP cycles back downward toward the BOP due to vessel heave motion. A reasonable clearance is chosen to avoid damage based on the physical dimensions of the LMRP and BOP.
- 107. 154 zyxwvutsrqpon Chapter z 9 The maximum top tension based on riser recoil is limited to no more than the value that could cause excessive slack in the tensioning lines as the riser disconnects and moves upward. The slack could occur soon after disconnect as the riser accelerates upward and the tensioning system cannot keep up. Slack also could occur as the riser is stopped. Finally, the maximum top tension is limited to no more than the value that the riser recoil system can stop during an emergency disconnect. The riser can be stopped by a combina- tion of the riser recoil system and an arrangement in which the tensioners bottom out before the telescopicjoint collapses. This arrangement, sometimes called a deadband, pro- vides for the riser having no force applied to it after the tensioners have bottomed out. This provides some assurance that the riser does not apply force to the rig floor even at relatively high tensions. These topics will be discussed further under the riser recoil discussion in Section 9.2.9. z 9.2.7.3.5 Tensioner StrokelTelescopic Joint [TJ) Stroke Limits During EDS/drift-off conditions, the limits on tensioner stroke and telescopic joint stroke become important. The amount of allowable stroke-out depends on how far the telescopic joint is stroked out when it is in its nominal (Le. calm seas) position at the site. Several factors can cause this nominal position to be “off centre” including the placement of pup joints in the string leading to the outer barrel to be slightly high or low on the inner barrel. As the telescopic joint is stroked out, a margin before complete stroke out of either the tensioners or the telescopic joint must be maintained to allow for wave-frequency variations and other uncertainties. This will be discussed further under the EDS/drift-off discussion in Section 9.2.8. 9.2.7.3.6 BOP, Wellhead, and Conductor Limits The BOP, wellhead, and conductor pipe are often designed by the loading experienced during EDS/drift-off conditions. The BOP manufacturers provide curves that indicate the rated capacity of the flanges when loaded in tension, bending, and pressure. The wellhead manufacturer provides a similar rated capacity for the wellhead. Finally, the conductor has its connectors and pipe rated for tension and bending. The riser analysis results (including BOP, wellhead, and conductor loading) are compared against these ratings to determine whether the rating of the system is exceeded. Analysis can be conducted to determine whether riser loading at the bottom flex joint is within the capacity of each of the BOP connectors, the wellhead, and the conductor casing. This analysis is conducted for combinations of vertical load, lateral load, and pressure load conditions specified by the operator. Depending on the component designs, the highest loading occurs during drift-off and the weakest link for bending loads is often either the wellhead connector or the casing connector closest to the wellhead connector. As noted under operating limits, the assumption of a rigid, vertical BOP is generally a conservative approach for BOP component loads, but a more rigorous approach involves coupled analysis. After the rig is on site, the misalignment angle of the conductor casing from vertical could be large enough to warrant reanalysis to determine its influence on the component loads.
- 108. Drilling and Production zyxwvutsrqp Risers zyxwvuts 755 z This could assist in establishing a vessel position that would lead to improved bottom flex joint angles for drilling. Another topic on operating limits involves torsional loading in special situations in which the vessel rotates and applies torsion to the riser and the wellhead. Depending on the component designs, the weakest link with torsional loading in the system could be the wellhead connector or the casing connector. Operational procedures and limits are set to avoid rotation or damage to these components. zyxw 9.2.7.4 Typical Operating Recommendations 9.2.7.4.1 Recommended Top Tension vs. Mean Vessel Offset Recommended riser top tensions are determined based on the limits defined in Section 9.2.7.3, except riser recoil limits which will be introduced in a later section. These recommended top tensions are discussed in the example below. Riser analysis for a connected riser configuration was conducted to determine whether the rig’s top tensioning capacity is sufficient to support a riser in 9000 ft of water under some representative design metocean conditions. This assessment was done for drilling operations with up to 16-ppg mud in the Gulf of Mexico. The riser stack-up described in Section 9.2.4.3.2 was modelled in a typical riser analysis program. As noted earlier, vessel RAOs from Section 9.2.4.1 are used in the analysis for the riser configuration in 9000 ft of water. Analysis was carried out for the following conditions, one with extreme waves and the other with extreme current. As noted earlier, the operational limits that apply for the non-drilling conditions are substantially less restrictive than those that apply for drilling conditions. Also, the high current conditions have a much different influence on the riser than the storm conditions. Besides the high drag loads, vortex-induced vibration of the riser pipe cause increase riser drag coefficients, causing larger riser angles. For the conditions discussed above, state-of-the-art riser programs are available to calcu- late the riser’s deflected shape, angles, and stresses. As noted earlier, these programs often carry out a solution in the frequency-domain or in the time-domain. Both types of solutions can be used in conducting large parameter studies for determining recommended top tensions with various offsets and mud weights. Frequency-domain programs tend to use less computer time, so they have become more popular. Results from riser analysis pro- grams can be used to assemble parametric results that show plots of top angle vs. top tension, bottom angle vs. top tension, stress vs. top tension, and other relationships for various mean vessel offsets and mud weights. Top tensions that satisfy the operational limits can be derived from these results. 1-yr Winter Storm zyxwvu - Connected, Drilling IO-yr Winter Storm - Connected, Non-Drilling High Current - Connected, Drilling Extreme Current - Connected, Non-Drilling
- 109. 156 zyxwvutsrqpon Chapcer z 9 z -o- Rig Capacity -zyxwvu m - Max Setting +Min. Rec Setting zyxwvut -3 zyxwvutsrqp -2 -1 0 1 2 3 Figure 9.22 Drilling operations window, 1-yr storm zyx Figure 9.22 shows a curve of tensions that satisfy operational limits for various offsets. As shown, these tensions are within the API maximum tension for a range of offsets. If the vessel can keep station within this range of offsets, the operating tension can be established for that mud weight. A vessel’s mooring system can typically keep the vessel stationed within & 2 % of water depth. If the riser angles are too large at these offsets, the vessel can be positioned at a more favorable offset by using “line management”. The mooring lines of the vessel can be “managed” by being pulled in or payed out to position the vessel over the well. This requires additional action on the part of the crew and can be restricted under severe metocean conditions. If a vessel is dynamically positioned, it can typically hold station within an offset circle of 1% of water depth from its set point (not always directly over the well). Given an offset circle of this size, the top tensions needed to satisfy the riser’s operating limits vary with metocean conditions. The top tension needed with 16-ppg mud in a one-year winter storm is about 1700kips, as shown in fig. 9.22. In high currents, (fig. 9.23), the top tension needed to satisfy the same conditions is about 2400 kips. 9.2.7.4.2 Top Tensions for Various Mud Weights Curves such as those above are generated for various mud weights and are compiled to form a curve of top tension vs. mud weight. This curve is useful because the crew can adjust the top tension as the mud weight is changed, whereas top tension cannot be practically changed as offset varies. A graph of top tension vs. mud weight, specific to each well, is considered a key document on the rig. Figure 9.24 shows a typical curve of top tension
- 110. Drilling and Production Risers zyxwvutsr 3200 2800 2400 2000 1600 1200 800 400 - -zyxwvu Max. zyxwvutsr Setting +Min Rec Setting zyxwvut -3 -2 -1 0 1 2 3 Figure 9.23 Drilling operations window, high current zyx Recommended Vertical Tensions Ultra Deepwater Drillship -GoM -8143 feet zyx 3000 -2500 .- % x - 5 2000 .- v) C E 1500 m 0 i ? 3 1000 zyxwvutsrq 8 v) , ' * Recoil Min Il-Yr Storm 1 l -Min API Tension jzyxwv 7 z a 500 ~ 0 1 8 9 10 11 12 13 14 15 16 Mud Weight (ppg) Figure 9.24 Curve of operating tensions vs. mud weight 1 Recoil Max 151 I versus mud weight for a dynamically positioned vessel, including riser recoil limitations that will be discussed in Section 9.2.9. 9.2.8 Emergency Disconnect Sequence (EDS)/Drift-Off Analysis When a dynamically positioned drilling vessel loses power in ultra deepwater, the resulting motion of the vessel and the response of the riser depends on the intensity of the wind, waves, and current. A "drift-off' begins and the vessel tends to rotate from a heading with
- 111. 758 zyxwvutsrqpo Chapter 9 the bow into the weather to a heading turned 90” with the weather on the beam. Under the effects of the increasing vessel offset from the wellhead, the riser’s deflected shape changes with time and is significantly affected by the vessel’s drift-off speed. Through analysis, the riser’s deflected shape can be shown to govern the time at which emergency disconnect limits are exceeded. By allowing a specific time for the emergency disconnect sequence to be carried out, yellow and red alerts are established to protect the system. This section discusses the practical application of EDS/drift-off analysis and the techniques that are used. The response estimates of the vessel and the riser during drift-off conditions are used for setting emergency disconnect limits. The yellow and red alerts are set at vessel offsets and riser limits that will allow an emergency disconnect sequence (EDS) to be carried out while assuring the integrity of the drilling riser and its associated equipment. Potential drift-off scenarios are analysed to establish the yellow and red alert settings. Results from an EDS/drift-off analysis are generally used to guide the captain in deter- mining DP settings for each well. An analytical simulation of the response of the riser and the vessel is used to determine the time available to disconnect the riser. The resulting prediction of available time are used by the captain to set alert circles for planning the “emergency disconnect sequence” or EDS. The EDS defines a series of alert circles, each of which has required procedures for the crew to prepare for riser disconnect. For example, a yellow alert circle includes a procedure for discontinuing drilling and hanging the drill pipe off in the BOP stack. zy A red alert circle signals the captain or the driller to “activate a red button” to start an automatic sequence that causes the drill pipe to be sheared by shear rams in the BOP stack and the riser to be disconnected. The EDS ensures the integrity of the riser and the related equipment, particularly the BOP stack, connectors, and conductor pipe that provide well pressure containment. The disconnect times are governed by exceedance of limits on top riser angle, bottom riser angle, slip joint stroke, wellhead moment, and conductor moment. The vessel is considered to be in either one of the following two modes when a drift-off occurs: the first mode can be termed “drilling operations” and is associated with metocean conditions that are suitable for drilling; and the second mode can be termed a “state of readiness” and is associated with metocean conditions or other conditions that prohibit normal drilling activities. When comparing drift-offs in the two modes, starting with drilling operations, more time is required to carry out the procedures required to dis- connect the riser (say 150 s). When starting from a state of readiness, the captain or driller is ready to activate the red button to start the emergency disconnect sequence so that less time is required (say 60 s). Because of this time difference, drilling operations are discon- tinued in certain metocean conditions and a state of readiness can be continued into larger metocean conditions. (Please consider the times quoted above as examples only; actual times vary with drilling vessel.) An example set of metocean conditions used for the state of readiness mode is a 10-yr winter storm with a 1-minwind speed of 50.5 knots, a significant wave height of 19ft, and a surface current of 0.6 knots. An example set of conditions for drilling operations is a one-minute wind speed of 25 knots, a significant wave height of 7.6 ft, and a surface current of 0.3 knots.
- 112. Drilling and Producrion Risevs 159 z 9.2.8.1 Drift-Off During Drilling Operations zyxwv For a drift-off that occurs during “drilling operations”, the yellow and red alert circles are set using the time history of vessel motion and riser response resulting from the EDS/drift-off analysis. The point in time at which the first riser allowable limit is exceeded is termed the “point of disconnect”, or POD. Disconnect at any later time would exceed a system allowable. With the POD as the basis, the vessel motions data time history is used to move backward according to the time required from “activating the red button” to the POD. As noted earlier, an example time allowed for this portion of the sequence is 60 s. This determines the time and offset position associated with the red alert circle. From the red alert circle, the vessel motions data time history is used to move backward again according to the time required to move from ongoing “drilling operations” to a “state of readiness”. An example time for this portion of the sequence is 90 zyx s. This determines the time and offset position associated with the yellow alert circle. As noted above, the point of disconnect (POD), which drives the yellow and red alert circles, is governed by first exceedance of an allowable limit within the system. In this process, allowable limits are set for any component whose integrity could be compromised as the vessel drifts off. The limits are generally set for the top riser angle, the bottom riser angle, stroke-out of the slipjoint, stroke-out of the tensioners, loading on the BOP. loading on the wellhead connector, loading on the wellhead, and loading on the conductor pipe. Typical limits for top and bottom angles are 9” (90% of the flex joint stop, per API RP 16Q) and stroke-out values of say zyxwvu 25 ft based on some margin within a 65-ft stroke capacity, for example. In 4500 ft of water and 9000 ft of water in the Gulf of Mexico, summary results for an EDS/ drift-off analysis in a reasonable set of metocean conditions (Le. the 95% non-exceedance environment) used for drilling operations are as follows: 4500 ft - Red Alert Circle = 225 ft (5% WD); Yellow Alert Circle = 72 ft (1.6% WD) 9000 ft - Red Alert Circle = 360 ft (4% WD); Yellow Alert Circle = 180 ft (2% WD) In these examples, the results in 4500 ft of water are governed by yield of the conductor pipe; whereas the results in 9000 ft of water are governed by stroke-out of the slip joint. As shown above, drift-offs tend to be more difficult to manage in the shallower water depths. In 4500 ft of water, the size of yellow alert circle has reduced to a relatively low, but manageable level when compared to the larger yellow circle in 9000 ft. 9.2.8.2 Drift-Off During a State of Readiness For a drift-off that occurs during a state of readiness, the metocean conditions used are design values in which the riser will remain connected. An example of this is the 10-yr winter storm in the Gulf of Mexico. In the state of readiness mode, only the red alert circles are set and this is done using the time history of vessel motion and riser response resulting from the EDSidrift-off analysis. The point in time at which the first riser allowable limit is exceeded is termed the “point of disconnect”, or POD. From the POD, the vessel motions
- 113. 760 zyxwvutsrqpo Chapter zy 9 data time history is used to move backward according to the time required from “activating the red button” to the POD. As an example, the time allowed for this portion of the sequence is 60 s. This determines the time and offset position associated with the red alert circle. The allowable limits for the system are the same as they are in the drilling operations mode. In 4500 ft of water and 9000 ft of water in the Gulf of Mexico, summary results for an EDS/drift-off analysis in a reasonable set of metocean conditions used for a state of readiness are as follows: 4500 ft zyxwvut - Red Alert Circle = 90 ft (2% of WD) 9000 ft - Red Alert Circle = 225 ft (2.5% of WD) As with drift-offs from a drilling operations mode, drift-offs from a state of readiness tend to be more difficult to manage in a shallower water depths. In 4500 ft of water, the size of red alert circle is reduced to a relatively low, but again manageable level when compared to the larger red circle in 9000 ft. Figure 9.25 shows how much more rapid the drift-off in a 10-yr storm is when compared to the drift-off in a one-year storm. The comparison of results in the 4500-ft and 9000-ft water depth cases is also influenced by the larger riser restoring force in shallow water. 700 zyxwvutsrqpon 00 zyxwvutsrqp 60000 500 00 a zyxwvut c 40000 a 4 ! 30000 200 00 100 zyxwvutsrqp 00 0 00 + -9000-ft WD,IO-yr Storm -A- 45004 WD,1-yr Storm 0 20 zyxwvutsrq 40 60 80 100 120 140 160 180 200 Time (s) zyxw Figure 9.25 Drift-off analysis time histories
- 114. Drilling and zyxwvutsrqpo Production Risers zyxwvutsrq 761 z 9.2.8.3EDS/Drift-Off Analysis Technique This section describes a transient coupled analysis technique for calculating drift-off of a dynamically-positioned vessel and the associated effect on the emergency disconnect sequence for a drilling riser. The drift path of the vessel is calculated in the time-domain, taking into account the transient response of the riser and the vessel's change of heading under the influence of current, wind, and waves. The effect of vessel rotation on horizontal motion is important in calculating the yellow and red alert offsets for the EDS. Also, the effect of riser restoring force on the vessel will be shown to be significant. 9.2.8.3.1 Riser Response Analysis Transient dynamic analysis in the time domain provides a reasonable estimate of riser and vessel response during drift-off. An alternative approach is the quasi-static technique in which inertial forces are approximated and applied as loads distributed along the riser. A third alternative, the static analysis technique, is accurate only for certain combinations of very slow drift speeds or shallow water. A transient riser analysis can be used to model the inertial effects of the riser and the relative velocity effects between the current and the speed of the riser. Wave-frequency forces are often not a significant factor in these results. The vessel's linear (offset) mation time history is specified at the top of the riser and the analysis is run to generate the riser analysis results including top riser angle, bottom riser angle, slip joint stroke, riser stresses, and wellhead loads. Figure 9.26 shows the time history of slip joint stroke for conditions associated with a 10-yr storm (non-drilling, state of readiness) and with a reduced storm (drilling operations). Note that the slip joint stroke does not show any appreciable movement until about 50 and 100 s into the drift off, for the 10-yr storm and the reduced storm, respectively. As shown, the rate of increase in the slip joint stroke is much higher for the IO-yr storm. A typical allowable limit for slip joint stroke is between 20 and 30 ft depending on its stroke limits, the water depth, the top tension, and the space-out of the pup joints. Figure 9.27 shows the time history of bottom flex joint angle for both the 10-yr storm and the reduced storm. Note that the bottom flex joint angle does not show any motion until about 70-80 s into the drift off, regardless of the storm size. After the initial response, the rate of increase in flex joint angle is higher in the 10-yr storm, as expected. A typical allowable limit used against this curve of bottom flex joint angle is 9". 9.2.8.3.2 Importance of Coupled Riser Analysis The riser and vessel motions analysis programs are coupled to include the effects of riser restoring force on vessel motion. Depending on the water depth and specific conditions, this can provide a 15-20% reduction of offsets in the time history of vessel motion. A simplistic coupled analysis is illustrated below. First, the vessel analysis is done with no riser loads. Second, the resulting vessel motions are used in the riser analysis.
- 115. 162 zyxwvutsrqpo Izyxwvutsrqponmlkjih Chapter 9 0 zyxwvuts -10 -20 -30 -40 -50 -60 -70 ’ I i l i i l t l i Y I I I I 1 I I I I I 0 50 100 150 200 250 300 350 400 450 500 Time zyxw (s) zyxw Figure 9.26 Time history of slip joint stroke during drift-off E. B ’ 0 100 200 300 400 500 Time (s) Figure 9.27 Time history of bottom flex joint angle during drift-off
- 116. Drilling and Production Risers 163 Third, the vessel analysis is redone with lateral riser loads from the previous riser analysis. Fourth, the riser analysis is redone with the updated vessel motions. zy 0 zyxwvutsrqp A more sophisticated analysis approach would solve for the complete system (vessel and riser) at each time step. This would result in a fully coupled analysis. 9.2.8.3.3 Importance of Vessel Rotation The results of vessel motions analysis depend heavily on the heading of the vessel with respect to the incident weather (wind, waves, and current). The force on a drillship is much lower when it is headed into the weather than when the weather is on its beam (turned by 90’). To minimise force and vessel motions, the captain generally heads the vessel into the weather. When a vessel loses power, it will tend to rotate such that the weather is on the beam zyxwvuts - a stable orientation. The speed at which this rotation takes place can be calculated through the vessel motions analysis. Due to the differing force coefficients in the different headings, the rotational speed has an influence on how quickly the vessel translates away from its set point over the well. Vessel motions analysis can be carried out simply using the equations of motion for a rigid body based on Newton‘s 2nd law. The translational and the rotational motions are described by: zyxwvu mx = F(t) I@= M(t) where m represents the mass of the vessel, zyxw x represents the translational acceleration of the vessel at the centre of gravity (CG) in the surge and sway modes, I is the vessel mass moment of inertia, ii, denotes rotational acceleration in the yaw direction, the “dot” represents differentiation with respect to time (t),and F and M represent the exciting force vector and moment vector acting in the horizontal plane. The applied forces and moments are due to: Environmental forces and moments due to wind, current, and mean wave drift; Hydrodynamic forces and moments proportional to the vessel acceleration represented by the added mass term and added inertia terms at zero frequency; Hydrodynamic drag forces and moments proportional to the vessel velocity; and Riser reaction forces in the horizontal plane. The wind, current, and waves are applied collinearly and concurrently. The initial conditions of the vessel heading and velocity are defined. In a fully coupled analysis, as discussed in 9.8.3.2,the forces (including the riser restoring force) and moments are updated at each time step and the corresponding vessel motion and rotation in the horizontal plane are calculated. Example vessel characteristics are shown in table 9.8 The added mass and added mass moment of inertia at zero frequency are calculated using a diffraction program. The current and wind force and moment coefficients can be determined from a wind tunnel model test.
- 117. 164 zyxwvutsrqpo Table 9.8 Vessel principal particulars zyxw Chapter 9 z Length (perpendiculars) Breadth zyxwv 1Depth / m 117.8 I IDraft IDisplacement Iton 154,7091 9.2.8.4 Trends in Analysis Results with Water Depth Trends show that EDS/drift-offs are more difficult to manage in shallow water than in deep water because, in deeper water, a specific amount of distance traveled by the vessel results in a lesser percentage offset and a lesser angle. Not all of this advantage can be retained, however, because of the shape of the riser and the different allowable limits involved. In waters shallower than 5000 ft, the wellhead or conductor moment may be the governing limit that establishes the point of disconnect (POD) discussed earlier. The moment values are determined by the soil properties and the dimensions and yield strengths used in these components. Figure 9.28 shows a typical conductor pipe bending moment profile based on the drift-off trajectories for beam sea and the rotating ship conditions. Figure 9.29 shows a summary of drift-off analysis results for site in 4227 ft of water in the Gulf of Mexico, with a riser top tension of 1371 kips and a mud weight of 10 ppg. The curve represents the horizontal vessel excursion (offset) vs. time. A vertical line is drawn at the time of POD, which is the minimum of the times at which the allowable limits for stroke, angles, wellhead bending moment, and conductor bending stress were reached. In this example, the POD occurs at 254 s and the associated offset is 467 ft. If the time is reduced by 60 s (to 194 s), the red circle radius is established as 290 ft. If the time is reduced by a further 90 s (to 104 s), the yellow circle radius is established as 89 ft. In dynamic- positioning operations, the yellow circle defines the offset at which drilling operations are suspended and the red circle defines the offset at which the EDS sequence is initiated. 9.2.8.5 Operational and Analytical Options If the yellow or red circles are not large enough to be practical, options may be available by looking at the system as a whole. A first option is usually to find an analytical fix and the second to propose an operational fix. Analytical fixes can include exploring options for reduced top tensions, which if set too high initially, could cause difficulties in either riser recoil or connected riser recommendations. Reduced riser tensions and other such compro- mises may be needed to reduce the loads on conductors for EDS/drift-off, for example. In many regions of the world, metocean conditions are so severe that they cause difficulties in managing the possibility of EDS/drift-off.In areas such as the Gulf of Mexico, Trinidad, Brazil, and the Atlantic margin, high currents can cause a vessel to drift off rapidly. If currents exceed conditions associated with a state of readiness mode, steps to provide operational management might be necessary such as positioning up current or simply
- 118. Drilling and Production Risers 165 z 20 0 zyxwvuts - -20 2 -40 -60 zyxwvuts 2 -80 & E. P zyxwvu s zyxwvu 13 -100 -120 zyxwvuts 0 -140 -160 -1000 0 1000 2000 3000 4000 5000 6000 7000 8000 zy 900Cz Bending Moment (kips-ft) Figure 9.28 Conductor pipe bending moment profile during drift-off 600 500 5 'E 400 # 2 .t 200 YI - 300 I I" 100 0 0 50 100 150 200 250 300 Time (recs) Figure 9.29 Summary of drift-off analysis results
- 119. 766 zyxwvutsrqpo Chapter 9 z disconnecting the riser in some conditions. However, a disconnected riser in high currents is also difficult to manage due to the large angle that it will take on. In areas of the world that have high wave conditions that build rapidly, the possibility of an EDS/drift-off event poses another type of riser management issue. If the riser can survive EDS and hang-off in design level wave conditions, the management issue is simply a matter of when to disconnect and ride out the storm. Disconnection of the riser protects the pressure-containment components, i.e. the BOP, wellhead, and conductor. However, when a site has design wave conditions in which EDS and hang-off can jeopardise the free- hanging riser, the riser is pulled before the storm is encountered. Depending on the water depth and the forecasted seastates, the riser pulling operations are begun well in advance. z 9.2.9 Riser Recoil after EDS This section covers the response analysis of the riser as the LMRP is released from the BOP during an emergency disconnect sequence (EDS). An understanding of this process is important in order to maintain safety and avoid damage to the riser and its related components. Additionally, riser recoil considerations often dictate the top tensions that are pulled on the drilling riser. Riser recoil analysis is conducted to determine the axial response of the riser after an emergency disconnect of the LMRP from the BOP at the seabed. In practice, this analysis is used to optimise riser tensioner system settings and define riser top tensioning bands to prevent excessive response of the riser. Typical allowable limits are aimed at ensuring the system behaves as follows after disconnect: the LMRP connector lifts off the BOP mandrel without reversal that could cause re-contact; the riser stops before impacting the drill floor, and slack in the tensioner lines is limited. To check these limits, some form of riser recoil analysis is generally done for each well site. This section provides a discussion of the riser recoil process, riser response analysis, allow- able limits, results and interpretation of some example cases, and sample operational recommendations. Although some sample guidelines are discussed here, general guidance would be highly dependent on the riser tensioning system and site-specificguidance would depend on the site and the selected operating parameters. The process and its analysis are discussed in more detail in Stahl (2000). 9.2.9.1 Definition of Process As the riser goes through an emergency disconnect sequence (EDS), it automatically dis- connects near the seabed. This disconnect is carried out at the interface between the lower marine riser package (LMRP) and the lower portion of the blowout preventer (BOP). As the riser releases, it responds with upward axial movement that is managed through the tensioners and the associated riser recoil system. Management of the riser’s upward movement is carried out by adjusting the stiffness and/ or damping of the tensioner system. This can be done in a variety of ways and the examples below do not cover all of them. In one example system, the EDS includes an automatic command to close air pressure vessels (APVs) normally kept open to maintain small tension variations during operations. This causes a sudden increase in the system’s vertical stiffness. Also, a so-called “riser recoil” valve is shut to increase the damping by
- 120. Drilling and Production zyxwvutsrqp Risers 161 z constricting the orifice for fluid flow. In another example system, the riser’s upward movement is managed by changing the orifice size based on tensioner stroke or velocity, with no closure of APVs. Several properties of the riser also influence the riser’s vertical response. First, the in-water weight of the riser string and the LMRP affect the dynamics of the riser. In certain cases, bare joints of riser are included in the riser stackup to help control the upward movement. Secondly, the weight of mud contained in the riser alters the response after disconnect; the frictional effects of the mud stretch the riser downward for some duration after disconnect. Thirdly, in deep water, stretch in the riser can be significant (several feet) and this leads to a rapid upward response (slingshot effect) after disconnect. zyxw 9.2.9.2 Riser Response Analysis Some of the key modelling parameters and analysis cases are considered in a riser recoil analysis. Riser recoil analysis is generally carried out assuming only axial response, with fluid flow through the tensioning system, vessel heave, effects of offset on vertical tension, and mud flow all playing a big part in the response. For this discussion, due to its rig- specific nature, the tensioning system is simply considered a spring-damper device. Heave is an important input parameter, with its selection generally based on a relationship to a vessel in a design storm. Top tension used in the analysis is altered depending on the offset that is of interest. This is due to the build up of tension that can be caused in some systems when the APVs close some time prior to disconnect. Mud flow is typically modelled in the analysis, with higher mud weights give higher frictional loads on the sides of the riser as they fall out, thereby pulling the riser downwards for some duration after disconnect. 9.2.9.3 Allowable Limits The allowable limits on riser recoil set the following riser top tensions: minimum top tensions to keep the LMRP from damaging the BOP during disconnect; maximum top tensions to avoid slack in the tensioner lines just after disconnect; and maximum top tensions to avoid the riser impacting the drill floor. Minimum tensions are limited by avoidance of contact between the LMRP and the BOP, as the LMRP cycles back downward toward the BOP after disconnect. Such movement could occur if the disconnect were to occur at the “worst phase” of a vessel’s heave cycle. Such phase considerations cannot be controlled because of the duration (about 60 s) of the EDS sequence. Allowable limits on such motion are dependent on the BOP equipment and the tolerance for damage, but leaving a few feet of clearance is generally considered reasonable. Maximum riser top tensions are limited by avoidance of slack in the tensioner lines during riser recoil. The upward motion associated with this limit could be exacerbated if the disconnect were to occur at the “worst phase” of a vessel’s heave cycle. As noted above, such phase considerations cannot be controlled because of the duration (about 60 s) of the EDS sequence. Reasonably small amounts of slack are allowed with certain systems, but no specific limits have been established. To avoid the riser impacting the rig floor, a “deadband” might be available to provide further protection. This deadband can be defined as an arrangement whereby the ten- sioners and slip joint stroke ranges are offset. In this arrangement, when the tensioners
- 121. 768 zyxwvutsrqpon Chapter 9 z have pulled their line to their full upward extent, the telescopic joint should still have some travel available (say 5 ft) before it bottoms out. Thus, the tensioners would apply no force to the riser while in this deadband. This arrangement provides a cushion that would help to slow down the riser if it strokes upward further than expected. Slack in the tensioner lines would have to be managed, however. Through means of this deadband arrangement, further limits on maximum top tension can be avoided. zyxw 9.2.9.4 OperationalIssues As noted above, some form of riser recoil analysis is generally used for every deepwater well site. Due to the impact of the results on riser top tensions, sensitivity cases are sometimes run to investigate ways to allow a larger band of allowable tensions thus making better use of the rig’s installed tensioner capacity. The nature of these sensitivity cases would depend on the rig’s tensioner and recoil system. Examples of such cases could include closing varying numbers of APVs, thus altering the stiffness at the time of disconnect; or a larger orifice or a different program for changing the orifice size. 9.3 Production Risers Four types of production risers were mentioned in the introduction: 1. Top-tensioned (TTR) 2. Free Standing 3. Flexible 4. Steel Catenary (SCR) Figure 9.1 illustrates the various kinds of risers. All are designed to convey well fluids to the surface. Each type has unique design requirements. Flexible risers are the most common type of production riser. They may be deployed in a variety of configurations, depending on the water depth and environment. Flexible pipes, long the standard riser for floating production, have traditionally been limited by diameter and water depth. Deepwater projects in the Gulf of Mexico and Brazil are now employing SCRs for both export and import risers. Figure 9.30 shows the capability of flexible pipes as of this writing. This will undoubtably grow in the future. The choice between a flexible riser and an SCR is not clear cut. The purchase cost of flexible risers for a given diameter is higher per unit length, but they are often less expensive to install and are more tolerant to dynamic loads. Also, where flow assurance is an issue, the flexible risers can be designed with better insulation properties than a single steel riser. Flexible risers and import SCRs are associated with wet trees. Top tensioned risers are almost exclusively associated with dry trees and hence are not usually competing with flexibles and SCRs except at a very high level: the choice between wet and dry trees.
- 122. Drilling and Production zyxwvutsrqpo Rrsers zyxwvutsrq 769 z Figure 9.30 Capability of flexible pipe (Technip Offshore) 9.3.1 Design Philosophy and Background 9.3.1.1 Metocean Data Each location may have critical design conditions; e.g. loop currents in the Gulf of Mexico and highly directional environments in the West of Africa. Vessel motions and offsets have a major influence on riser design and should be paid due attention (see Section 9.3.1.4). Metocean data used in riser analysis are water depth, waves, currents, tide and surge variations and marine growth. For the extreme waves and currents, the 1, 10, 100-yr and higher return periods may be considered. The 95% non-exceedance values may be used as temporary installation design condition. Long-term waves are defined by an zyx HsTp scatter diagram, with directionality if required. Interfacing between the riser analysts and the metocean specialists at an early stage in the design process is recommended, so that riser-critical environmental conditions do not get overlooked. The importance of both directionality and of joint wavelcurrent behavior varies from one location to another and should always be carefully considered. Riser response is period sensitive, and analyzing the maximum wave-height case with a single wave period may not result in the worst response of the riser by reference to vessel RAOs, ensuring that important peaks in vessel response are not missed. It should be recognised that the confidence with which metocean design data is derived varies considerably from one geographical location to another. Currents in the deepwater Gulf of Mexico, for example, are considerably higher than on the shelf. This has a large impact not only on the design of risers and mooring systems but also on the methods used for installation, and this emphasises the need for reliable site specific data. It is recommended that currents specified for the riser design include an allowance for uncertainties in the derivation of data. No general rule for this is laid down here; such decisions should be taken in consultation with metocean specialists. 9.3.1.2 Materials Selection (This section contributed by David Rypien, Technip Offshore, Inc., Houston, TX) Materials for riser pipe and components are selected based on design criteria, environmental conditions, and economics. In most cases, the governing criterion is
- 123. 770 zyxwvutsrqpo Chapter 9 the economics determined by trade-offs for the type of material, e.g. using carbon steel vs. titanium. Titanium was selected for stress joints (Oryx Neptune Spar, Placid Green Canyon 29), and in one case for an entire drilling riser (Heidrun). However, it is generally uneconomic for normal applications. Composite material has also been proposed for risers, but until now has been considered too expensive or immature. A composite string is currently being tested on the Magnolia TLP in the Gulf of Mexico. Once the material type is selected, a material specification is developed that considers the operating environment; lowest anticipated service temperature, sour service, and/or cathodic protection. The key material properties include: 1. hardness, 2. strength, 3. toughness Weldability considerations generally limit use of steel to yield strengths of 80 ksi or less. Higher strength steels may be used with threaded and coupled joints; however, these joints have higher stress concentrations, lower fatigue resistance than is typically required for floating production systems. Finally, inspection, testing (including fatigue testing) and packaging requirements need to be specified. Common standards and specifications used for carbon steel riser pipe and components are listed below: API RP 2 RD API 5L API RP 2 2 ASTM A370 BS 7448 DNV-OS-F101 DNV-OS-F201 NACE MR-01-75 Design of Risers for Floating Production Systems and Tension Leg Platforms Specification for Line Pipe Recommend Practice for Preproduction Qualification for Steel Plates and Offshore Structures Methods and Definitions for Mechanical Testing of Steel Products Fracture Mechanics Toughness Tests. Methods for determination of fracture resistance curves and initiation values for stable crack extension in metallic materials Offshore Standard zyxw - Submarine Pipeline Systems Standard for Dynamic Risers Sulphide Stress Cracking Resistant Metallic Materials for Oilfield Equipment Line pipe material specifications are often combined with casing pipe sizes to be compatible with well systems. API Spec 5L specifies two classification levels: PSL 1 and PSL 2 to define, generally, lower and higher strength steels. Most riser applications call for PSL 2 classification, typically X52, X60 or X80.
- 124. Drilling and Production zyxwvutsrqp Risers zyxwvutsrq X80 I7 z 1 80-100 90-120 zyxw Table 9.9 Strength range for API zyxw 5L pipe 1Yield strength, ksi IGrade IUltimate strength, ksi 1 152-77 166-1 10 I 160-82 175-1 10 1X65 165-87 177-1 10 I 1x70 170-90 182-1 10 I Specifications of chemistry and heat treatment that will achieve the required material strength, hardness, and toughness need to be developed with the assistance of the pipe manufacturer. 5.3.1.2.1 Strength Tensile strength is defined in terms of yield, zyxw oy, and ultimate, zyx q,. Yield strength is defined as the tensile stress required to produce a given percentage of strain, e.g. API 5L determines oy,corresponding to the value is 0.5% E (strain). If a tensile test continues past the point of yield, the material elongates and, in a ductile material, the area is reduced. The stress, based on the original area, is the ultimate tensile strength. API 5L specifies a minimum range of strength levels for the various steel grades as shown in table 9.9. X65 or X80 are the most common steel grades for top tensioned production risers. The amount of elongation before failure is a measure of ductility. API 5L specifies a minimum elongation, e, in 2 in. length as A0.2 e = 625,000- zyxwvu u o - 9 (9.4) where e =Minimum elongation in 2 in. to the nearest percent. A = Specimen area, in’, U= Minimum ultimate tensile strength, psi. For example, the elongation of a round bar specimen with A =0.2 in’, and U= 100 ksi would be 14%. API 5L also requires that the ratio of oJou shall be less than 0.93 to insure a level of ductility. 5.3.1.2.2 Hardness The following discussion is taken from www.tpub.com,’doematerialsci I . “Hardness is the property of a material that enables it to resist plastic deformation, penetration, indentation, and scratching. Therefore. hardness is important from an engineering standpoint because resistance to wear by either friction or erosion by steam, oil, and water generally increases with hardness. Hardness tests serve an important need in industry even though they do not measure a unique quality that can be termed hardness. The tests are empirical, based on experiments and observation, rather than fundamental theory. Its chief value is as an inspection device
- 125. 112 zyxwvutsrqp ~ zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA Chapter 9 - ~ - z 600 RockwellB / z y / 140 120 100 E 8o - - 60 8 zyxwvut U 40 20 0 zyxwvutsr able to detect certain differences in material, when they arise, even though these differences may be undefinable. For example, two lots of material that have the same hardness may or may not be alike, but if their hardness is different, the materials certainly are not alike. Several methods have been developed for hardness testing. Those most often used are Brinell, Rockwell, Vickers, Tukon, Sclerscope, and the files test. The first four are based on indentation tests and the fifth on the rebound height of a diamond-tipped metallic hammer. The file test establishes the characteristics of how well a file takes a bite on the material”. The indentation tests are most commonly used in material qualification. Each method uses a different indentation ball size and results in a different value. Figure 9.31 shows the relationship of Rockwell and Vickers hardness numbers to the Brinnel Hardness. Hardness is directly correlated with strength, and inversely correlated with ductility. This is shown in fig. 9.32. Although hardness is normally used for testing purposes and not as an independent design criteria, a maximum hardness of 22 Rockwell C (275 HV 10 maximum at cap pass) is specified for risers and pipelines in sour service. High strength is desirable for weight reduction in deepwater. High hardness, however, increases the risk of brittle fracture. This is a critical concern for tensile members like risers where it is generally desirable to be ductile against failures, allowing time for detection and corrective action (e.g. a through wall crack should not cause fracture of the pipe). Also, high strength and hardness typically require an increase in the carbon content. Figure 9.33 shows the maximum attainable hardness for quenched steel as a function of carbon
- 126. Drilling and Production Risers zyxwvutsr 773 z 0 zyxwvutsrq 50 IW 15.0 zoo Z M zyxwvu Jw 350 zyxw ~~~ 30 20 10 -~ 0 zyxwvutsrq 0 50 1W 150 ZW 250 300 350 zyxw Figure 9.32 Tensile strength (Rothbart, 1964) content. API 5L specifies a maximum carbon equivalent for use in lin 0.43%, where carbon equivalent, CE is defined as M n (Cr +Mo + V)zyxwv (Ni+Cu) 5 + 15 CE=C+-+ 6 pipe to be less than (9.5) The carbon equivalent provides a guideline for determining welding preheat to minimise hardenability issues and reduce the cooling rate. 9.3.1.2.3 Toughness “The quality known as toughness describes the way a material reacts under sudden impacts. It is defined as the work required to deform one cubic inch of metal until it fractures. Toughness is measured by the Charpy test or the Izod test.
- 127. 114 zyxwvutsrqp 1zyxwvutsrqponmlkjihgfedcbaZY 10 - zyxwvuts 0 1 zyxwvutsrq Chapter 9 z -1 z Izyx 2 I zyxwvuts Figure 9.34 Charpy impact test (www.tpub.com/doematerialsci/) Both of these tests use a notched sample. The location and shape of the notch are standard. The points of support of the sample, as well as the impact of the hammer, must bear a constant relationship to the location of the notch. The tests are conducted by mounting the samples as shown in fig. 9.34 and allowing a pendulum of a known weight to fall from a set height. The maximum energy developed by the hammer is 120 ft-lb in the Izod test and 240 ft-lb in the Charpy test. By properly
- 128. Drilling and Production Risers I75 z calibrating the machine, the energy absorbed by the specimen may be measured from the upward swing of the pendulum after it has fractured the material specimen. The greater the amount of energy absorbed by the specimen, the smaller the upward swing of the pendulum will be and the tougher the material is”. A history of impact tests is given by Siewert, et a1 (1999). Charpy test results (CVN) are reported as absorbed energy for a standard test specimen. Results are presented in units of ft-lbs, or Joules for SI units (1 ft-lb zyxwv = 1.35582 J). Charpy Impact tests are not required for PSL 1 pipe. PSL 2 pipe must meet minimum requirements for the absorbed energy as spelled out in API 5L. Charpy tests are a fast and low cost method for measuring the toughness of steel plate. More elaborate CTOD testing (BS 7448) is sometimes used for measuring toughness of weld heat affected zones. API 5L specifies a Weld Ductility Test, which requires that a pipe be flattened with the weld at 90” to the point of application of the forces. In this test, no cracks or breaks of greater than 1/8” are allowed until the pipe is flattened to a prescribed distance. Increased demands on strength while maintaining an acceptable hardness for sour service (e.g. Vickers Hardness <275 HV lo), and toughness performance in deep water operations are currently on the edge of formulating a material chemistry that will meet these requirements. This dilemma has promoted the use of corrosion resistant alloys (CRA’s) and the use of cladding to try and meet service requirement while trying to keep the material costs down. zyxwvut 9.3.1.2.4Manufacturing Capability Another issue is the actual ability to manufacture riser pipe with the specified wall thickness and diameter for riser applications in deep water. Most pipe manufacturers cannot produce or handle these sizes. There are only a few, to date, that have material handling capacity for thick-walled, large diameter seamless pipe. Two manufacturers, which are currently capable of supplying pipe in these sizes, include: SUMITOMO PIPE & TUBE CO., LTD. 23-1 Sugano 3-Chome Ichikawa 272-8528, Chiba 272-8528 JAPAN zyxwvut http://www,sumitomokokan.co.jp/ +81 47 322 3322 +81 47 322 2448 and Tenaris Pipeline Services Carretera Mexico-Veracruz Via Xalapa, zyxwvuts km 433.7 (91697) Veracruz, Ver. Mexico www.tenaris.com (51) 2 989 1255 (52) 2 989 1600
- 129. Chapter z 9 z 776 zyxwvutsrqpo Many heat treat, tempering, and quench facilities are not capable to produce pipe with uniform material properties along the length and through thickness of the pipe. For heavy wall pipe, a pre-manufacturing test using the material specification should be conducted to verify the capability of potential riser pipe manufacturers. The pipe produced during these tests can be used as test pieces for weldability and fatigue testing, keeping some of the material testing costs down. 9.3.1.2.5 Field Welding It used to be assumed that once riser pipe is produced with acceptable mechanical properties that we are ready to conduct welding tests using the welding procedure specification called out for installation. This philosophy is changing. Prior to conducting installation weld procedure qualification testing, a weldability test of the material is conducted as outlined in API RP 22. The test enables the project to understand the material response to welding conducted at a low, medium, and high range window of heat inputs, e.g. ranging from 15 to 75 kJjin, and representing heat inputs for manual to automatic weld processes to be used. The weldability of the riser material is verified prior to installation welding as a result of these tests. After the riser material is supplied to the installation welders, formal weld testing is conducted using the actual installation welding procedures and conditions. The next step is to verify the welding procedure specification with appropriate welding procedure qualification records and testing. A review of these results will determine if the weld procedure is adequate to meet the material specification and produce high quality welds. One area of concern is to minimise the use of pre-heat during welding in order to facilitate installation of the riser. This is often difficult to do because of the carbon equivalent or chemistry of the material, Le. keeping in mind the purpose of pre-heat to reduce the cooling rate and hardenability of the material. Inspection of riser welds in the US by automated ultrasonic testing (AUT) has replaced manual and radiographic testing. Prior to conducting AUT, it is recommended that the project review the AUT procedure and the acceptance criteria. zyx A demonstration test by the AUT contractor should be conducted, and a follow-up verification of indications by manual ultrasonic testing should be performed to insure a reliable test. Follow-up audits should be conducted to verify continued weld quality during fabrication of riser sections. 9.3.1.3 Analysis Tools Riser analysis tools may be classed as frequency or time domain. Most tools for riser response to waves and vessel motions require vessel motions input in the form of Response Amplitude Operators (including phase angles), which permits appropriate marriage of vessel motions with forces from the wave kinematics. Analysis of riser VIV is widely carried out using the program SHEAR7 developed at MIT under a joint industry research study, and with the more recently developed program VIVA (2001). The programs enable prediction of riser VIV response under uniform and sheared current flows. Whilst time-domain analysis remains the preferred option in some cases (e.g. confirmatory extreme storm response analysis) the most commonly used VIV software (including
- 130. Drilling and Production Risers zyxwvutsrq 111z SHEAR7, VIVA and VIVANA) are frequency-domain programs. Reasonable accuracy may well be provided by such programs under many conditions, since VIV motions are typically small, as are the associated structural non-linearities. Furthermore, the reasonable allowance can often be made for some non-linearities by suitable post-processing of results where fatigue prediction is the main concern. Programs such as Flexcom-3D and Orcaflex are used for analysis to determine bending and deflection of the productioin riser systems. z 9.3.1.4 Vessel Motion Characteristics Characteristic vessel motions and their applicability to different design checks are discussed in table 9.10. Vessel RAOs are used throughout the whole design process and it is impor- tant for them to be well-defined. Spacing of periods in the RAO curve must be sufficiently close zyxwvuts - especially near peaks - to maintain good accuracy. A useful reference on this subject is Garrett, et a1 (1995). Noting that the riser attachment location can have a significant influence on both the riser extreme and fatigue response, as may vessel orientation relative to waves and current, it is important to be able to correctly and efficiently manipulate and transform the RAO data. 9.3.1.5 Coupled Analysis Vessel, risers and mooring lines make up a global system, which has a complex response to environmental loading. The interaction of these components creates a coupled response, which may be significantly different to that predicted by treatment of each component on its own. Fully coupled analysis may be conducted as part of the final riser verification. However, it may be worth considering a coupled analysis at an earlier stage in the design process so that problems with the riser, vessel or mooring line design are highlighted and possible cost savings identified. The design of offshore structures operating in hostile environment and in water depth more than 5000 ft requires the development of integrated tool which are accurate, robust and efficient. A hull/mooring/riser fully coupled time domain analysis may meet such require- ments. For some systems, the coupling effects may magnify the extreme hull responses. Whereas, for most platforms in deep waters, the coupling effects more likely lead to smaller extreme responses due to additional damping from slender members, which results in less expensive mooring/riser system. Not accounting for the riser stiffness, drag or damping when calculating vessel offsets may result in a conservative estimate of extreme vessel offset which may or may not be acceptable for storm analysis. Conversely, increased vessel offset may indicate that riser fatigue damage from first order effects is spread over a greater length of riser than is truly the case, resulting in an underestimate of riser fatigue damage. The effects of current and damping are interlinked. Current loads on risers can significantly affect vessel offset (e.g. current loading on risers accounts for 40% of total loading on one FPSO known to the authors) and may increase due to drag amplification if the risers are subject to VIV. On the other hand, the riser hydrodynamic damping is related to the riser drag and will tend to reduce the amplitude of riser first-order response to wave loading. In general, simplifications cannot be assumed to be either conservative or unconservative in
- 131. 778 zyxwvutsrqpo Table 9.10 Characteristic vessel motion summary table zyxw Chapter z 9 z Characteristic First order/ wave frequency (RAOs) Extreme offsets Low frequency/ second order motions Vessel springing, ringing (for vertically tethere vessels; e.g. TLP zyxwvu IVessel VIV 1Coupled motion Relevant design case Extreme, clashing, fatigue Extreme, clashing, fatigue Fatigue Fatigue Fatigue All Discussion RAOs describe vessel response to wave-frequency excitation. They are typically determined by diffrac- tion analysis and are used in all stages of riser design. Extreme offsets represent expected extreme positions at the riser’s point of attachment. To avoid undue conservatism, any first order contribution should be removed prior to riser dynamic analysis. Horizontal offsets are usually given, but TLP set-down and spar pitch can also be important. Flooded compartment conditions can give rise to appreciable set-down. Drift data should be used in detailed riser fatigue analysis. The typical format is mean offset + one standard deviation with period for a range of sea-stateslbins in the scatter diagram. The offset data is typically given for surge/sway but may include other degrees of freedom, such as, pitch for a spar. Current, wind and wave forces should be considered as contributors to these motions. The amplitude of vessel springing may be relatively small but could cause high levels of fatigue, especially at the TDP of an SCR, if it occurs a large proportion of the time. Vessel VIV is theoretically possible with any floating vessel subjected to current loading that has cylindrical sections with aspect ratios (L/D) greater than three. The frequency of excitation will be equal to the vessel’s natural frequency, which is typically 200-400 s depending on the mooring system. The implication for riser design is high levels of fatigue damage. More important in deepwater. their overall effect. However, larger errors can be expected as water depth and the number of risers increase. A global coupled analysis may be conducted using riser analysis software, though there may be limitations in representing the vessel. Alternatively some seakeeping codes could include risers and moorings, though it may be necessary to simplify these in order to limit computer time.
- 132. Drilling and Production zyxwvutsrqp Risers zyxwvutsr 779 z The issues of coupled analysis have been addressed in the Integrated Mooring and Riser Design JIP using a range of example vessellmooringlriser systems. The results of this work are available in Technical Bulletin (1999) describing the analysis methodology for preliminary and detailed analysis of integrated mooring/riser systems and outlining the relative importance of the various parameters and integration issues involved. 9.3.2 Top Tension Risers Top Tensioned Risers (TTRs) are long flexible circular cylinders used to link the seabed to a floating platform. These risers are subject to steady current with varying intensity and oscillatory wave flows. The risers are provided with tension at the top to maintain the angles at the top and bottom under the environmental loading. The tensions needed for the production risers are generally lower than those for the drilling risers. The risers often appear in a group arranged in a rectangular (or circular) array. 9.3.2.1 Top Tension Riser Types Top tensioned risers are used for drilling and production. Figure 9.35 shows the various types. Conventional exploration drilling risers use a low pressure riser with a subsea BOP. The subsea BOP was also used on the Auger TLP [Dupal, 19911, however most floating production platforms with drilling may now use a surface BOP. The Hutton TLP was the first floating production, dry tree unit, to use a surface BOP [Goldsmith, 19801. The Split BOP has recently been used for exploration drilling in relatively benign environments [Shanks, et a1 2002; Brander, et a1 20031. 9.3.2.2 Dry Tree Production Risers The earliest use of top tensioned risers was for offshore drilling in the 1950s. The first tensioners consisted of heavy weights attached to cables. These cables ran over pulleys to support the riser. These “deadweight” tensioners were replaced by pneumatic tensioners as shown in fig. 9.36. These tensioners used hydraulic cylinders to control the stroke of a block and tackle system. The riser was suspended by cables as in the deadweight system. This method has been replaced by direct acting hydraulic cylinders (fig. 9.37), which has been used on TLPs, and on the Genesis spar drilling riser. The hydraulic rods are in tension Figure 9.35 Types of drilling and production top tensioned riser systems
- 133. 780 zyxwvutsrqpo Chapter z 9 z Figure 9.36 Pneumatic tensioner for the direct acting tensioners. The Diana drilling riser was supported by the first ram style tensioner, fig. 9.38. This more compact arrangement of the tensioner was possible on the spar, because the riser does not take on an angle at the deck. The angle is taken at the keel of the spar and bending is accommodated by intermediate guides in the centrewell. The first top tensioned production riser was used on the Argyll field in the North Sea [NationalSupply Company, 19751.This was actually a tubing riser connected to a wet tree.
- 134. Drilling zyxwvutsrqpo and Production Risers zyxwvutsr Figure 9.37 Direct acting tensioner 78 z 1 9.3.2.3 Single vs. Dual Casing zyxwvu Figure 9.39 shows typical cross-sections for top tensioned risers, production and drilling. Top tension production risers utilised to date do not have separate insulation. However, the annulus between the tubing and the inner casing is, sometimes, filled with nitrogen to provide thermal insulation. The choice of single vs. dual casings is a trade-off between the capital cost and the potential risk of loss of well control. There is almost no risk of loss of well control during normal operation because of the sub-surface safety valves [SCSSV, see, e.g. Deaton, 20001 and the dual barrier effect of the tubing and riser (remember that a single casing represents dual barriers during normal operations). The risk of blow out occurs during workover. In this phase the tubing and SCSSVs are pulled and mud is introduced into the riser to provide overpressure in the well relative to the formation pressure. The amount of overpressure offered by the mud is termed the “riser margin” (or “riser loss”). Goldsmith, et a1 (1999) describe a methodology for analysing the risk cost (RISKEX) and capital cost (CAPEX) trade-offs for a single and dual casing risers. Figure 9.40 shows riser loss as a function of mud weight and water depth. Typical riser margins are 300400 psi as indicated. Existing dry tree units, spars and TLPs, use both single and dual casing risers in about equal numbers [Ronalds, 20011. Next Page
- 135. 782 Figure 9.38 Ram style tensioners [Bates, et zyxw a1 20011zyxw Chapter zy 9 Previous Page
- 136. Drilling and Production Risers zyxwvutsr 183 zy Figure 9.39 Cross sections of TTR 9.3.2.4 Codes and Standards zyxwvu A valuable source for tracking industry codes and standards for risers and all sorts of oilfield systems may be found at zyxwv http://www.rigcheck.com/codespecs.html. The primary industry recommended practices for production of riser design are API RP2RD and DNV OS-F201. These apply to all tensioned risers from floating production systems. Flexible risers are covered in API RP17B and Bulletin 17J. DNV has separate rules for flexible pipe, and recommended practices for titanium (RP F201) and composite (RP F202) risers. MODU completion/workover risers are covered in API RP 17G. Subsea tiebacks are covered in API RP 1111. The API Recommended Practice is based on a Working Stress Design (WSD) method. This is a prescriptive approach using a single utilisation parameter to account for all the failure
- 137. 784 zyxwvutsrqpon Low Normal High zy Chapter 9 SLS 1Annual per riser lo-' zyxw Figure 9.40 Riser loss vs. mud weight and depth [Goldsmith,et zyx a1 19991 10-'-10-2 1 zy o-~-1o - ~ mechanisms. The DNV standard allows either the WSD or a Load and Resistance Factor Design method (LRFD). LRFD is aimed at achieving a particular target safety level by utilising partial safety factors for each failure mode, or limit state. The intent of the DNV code is to achieve a certain reliability level by applying probabilistic analysis to the various failure mechanisms. Table 9.11 shows the target failure probabilities for different limit states and safety classes. The limit states are defined as: Serviceability Limit State (SLS) - Acceptable limitations to normal operations Ultimate Limit State (ULS) - Structural failure Fatigue Limit State (FLS) - Cyclic loading Accidental Limit State (ALS) - Infrequent loading Annual per riser 'Annual per riser Annual per riser Table 9.11 Target safety levels (OS-F201) 1Limit state 1Probability bases j Safety classes 1o - ~
- 138. Drilling and Production Risers zyxwvutsrq Riser status (phase) Testing Temporary with no pipelinelwell access zyxwvu Table 9.12 Classification of safety cases (OS-F201) Riser content Fluid category 1,3 Fluid category 2 Fluid category 4,5 z 1 Location class Location class Location class 1 2 1 zyxw 2 1 2 Low Low Low Low NA NA Low Low Low Low Low Normal I I - zyxw 785 In-service with pipeline/well access The classification of safety cases is defined in table 9.12. Fluid categories are defined by the International Standards Organisation (ISO) as: 1. Water based fluids 2. Oil 3. Nitrogen, argon and air 4. Methane 5. Gas LRFD, while more complex to apply, allows for optimisation that is not achievable using WSD methods. Figure 9.41 shows an example plot of utilisation vs. water depth for a top tensioned TLP oil production riser under combined extreme North Sea conditions and external overpressure. In this example, API RP2RD is seen to be the most conservative approach. LRFD would allow optimisation of the riser. 9.3.2.5 Riser Components The conventional TLP production riser is made up of the following components (fig. 9.42): Tieback connector at the bottom The bottom tapered joints or flex joints The riser joints and connectors The tensioner spool pieces The tensioner load rings The guide rollers at platform deck The surface tree The tubing strings inside Flowline connectors at deck level to trees or valves The Spar production riser is made up of the following components (fig. 9.43): Tieback connector at the bottom
- 139. 786 zyxwvutsrqpo Chapter z 9 z Figure 9.41 Example of the application of LRFD to riser design (Courtesy of DNV) zy Surface Tree MWL Splashzone Joints Standard Joints Stress Join? Mudline zyxw 7Tdmdzm Figure 9.42 TLP top tensioned riser
- 140. Drilling arid Production Risers zyxwvutsrq 787 z Surface Tree zyxwv Standard zyxwv Joints Stress Joint Mudline Figure 9.43 Spar top tensioned riser The bottom tapered joints or flex joints The riser joints and connectors The keel joint or lower stem The air cans The upper stem The surface tree The tubing strings inside Flowline connectors at deck level to trees or valves TLP and Spar riser systems are virtually identical below the floater. The differences are in the manner of supporting and tensioning the riser at the top. TLP risers are supported by the hull buoyancy. The tension is provided at the load ring, which is supported by tensioners (see fig. 9.42). Rollers at the TLP deck centralise the riser and accommodate the angle between the riser and the hull. Spar risers are supported by air cans (sometimes called buoyancy cans), not by the spar hull itself. Buoyancy cans may be either “Integral” or “Non-Integral” type, fig. 9.44. The
- 141. 788 zyxwvutsrqp Figure 9.44 Integral and non-integral spar buoyancy cans zyx Chapter zy 9 Integral air can consists of an air can attached to a riser joint. The air can is installed along with all the rest of the riser joints. This type of air can has only been used once on the Genesis spar. “Non-integral cans”, fig. 9.4413, consists of an inner pipe called a “stem” which supports the air cans. The non-integral cans are deployed separately. The riser is run through the stem and landed on a shoulder at the top of an extension of this stem called the “upper stem”. The upper stem carried the entire tension of the riser and the weight of the surface tree. The angle between the spar and the riser is accommodated by a keel joint. Early keel joints on classic spars consisted of a singlejoint made up of a large diameter pipe on the outside connected to the riser pipe with flexible connections at the ends [Berner, et a1 1997; Bates, et a1 20021. The outer riser contacts the spar hull at a keel guide; both items include a sufficient wear allowance to accommodate the loss of material caused by the relative motion over the lifetime of the project. This type of keel joint is not used on truss spars. Instead, the lower stem of the buoyancy cans is extended through the keel. The riser is centralised near the bottom of this stem with a ball joint, fig. 9.45. Mini-TLPs using top tensioned risers have also used keel guides. These have a similar function to those on spars, but typically require less stroke. One design is described by Jordan, et a1 (2004), fig. 9.46. The riser pipe typically follows standard casing dimensions and materials (API Bulletin 5C3), however they may be designed to line pipe specifications as well (API RP 5L). The choice of riser size and materials is discussed below. Riser pipe may be joined by threaded or bolted connections. Bolted connections are heavy and expensive. They are used primarily where superior fatigue performance is
- 142. Drilling and Production Risers zyxwvutsr 789 z Figure 9.45 Keel centraliser for the Matterhorn TLP [Jordan,et a1 20041 zy [++ 3oU zyxwvuts I Figure 9.46 Lower stem and ball joint for truss spar [Wald, et zyx a1 20021.
- 143. 790 zyxwvutsrqpo Chapter 9 External sealing with pressure energized metal- to-metal seal. External torque shoulder could be considered a redundant metal-to-metal seal Easy trouble free make up provided by creep stabbing coursethread design and 45 deg stab flank. Connector makes up in 3.7 turns and is free running Low stress concentration factors provided by elliptical transition areas. low stress areas Torque up on connector, not pipe. Functional for wide make-up torque range large load flanks and even load distribution Fatigue resistance enhanced by internal elliptical load redirection grooves, generous thread root radii and very tight tolerance bard. High flank-to- flank radial thread interference to prevent back-out Internal pressure sealing with metal-to-metal pressure energized radial seal High load bearing external shoulder capable of supportinga minimum of 6000' in air zyxwvutsrq Figure 9.47 Typical weld-on upset connector required, e.g. for stress joints and keel joints. Threaded connections may be either weld-on (fig. 9.47) or couplings with threads machined into the pipe (fig. 9.48). Casing couplings are cheaper but until recently were normally not suitable for fatigue sensitive applications [Cargagno, et a1 20041. For example, stress concentration factors (SCFs) for the thread root of the casing connectors are typically five or greater. SCFs are usually stated relative to the nominal pipe wall stress. Upset weld-on connecters, on the other hand, can achieve SCFs as low as about 1.2. Both type of connectors can achieve 100% of the strength of the pipe. Weld-on connectors, however, are not suitable for high strength steel, greater than 95 ksi yield strength, because of insufficient data for weld performance in riser applications. Thus the choice of connector is tied to the overall performance requirements of the riser, which requires a significant amount of analysis. 9.3.2.6 Riser Sizing Riser sizing requires consideration of a number of load cases. The size may be dictated by pressure, collapse, tension, bending or a combination of these factors depending on the
- 144. Drilling and Production Risers zyxwvutsrq VAM TOP zyx FE-NA 791 z Box zyxw CriticalAreazyx 1 G m n / , 0.463 Wall zyxwvutsrqpo 5 - 1 7, / T t Pin/ Critical Area I (Connection Connection O.D. I.D. 10.486 8.925 Figure 9.48 Typical casing connector OD at Fa& 1 1 10.333 Pipe Pipe O.D. ID. 9.625 8 . m system. It is important to develop as early as possible a design and functional specification which spells out the various load cases, and that this document become a primary reference document throughout the course of the project. Changes in functional requirements can have “ripple” effects and should be communicated to designers and analysts as soon as they occur. It is also important to keep a close communication between the riser designers and others involved in the riser interfaces throughout the process: Oceanographers Vessel and mooring designers Global response analysts Process engineers Riser design, especially in deep water, is an iterative procedure. Initial assumptions about topsides weight, vessel size and response and even well characteristics might change several times in the course of a project. A rapid and accurate model for riser sizing is important for keeping up with these inevitable changes. Riser definition starts with specification of 1. 2. Concentric or non-concentric tubulars 3. Well layout and layout on the seafloor The inner riser size will be dictated by the size of tubulars, umbilicals, subsurface valves and connectors that have to fit within the internal diameter. In some deepwater applications, the production riser has been used for drilling [Craik, et al 20031. If this is the case, the outer riser size will be dictated by the drilling program. Figure 9.49 shows the layout for concentric and non-concentric tubulars. The diameter selected should allow sufficient clearance for the connectors of the inner tubing, Le. the space should allow for the drift diameter of the inner tubulars. Number of tubulars: single casing or dual casing (see discussion above)
- 145. 192 zyxwvutsrqpon Figure 9.49 Concentric and non-concentric tubulars zyxw Chapter 9 z For the minimum performance properties for common casing sizes used for risers, consult API RP 2RD. The collapse resistance (in psi) and pipe body yield values (in lbs.) are listed for different riser outside diameters in six tables. There is no requirement that riser sizes fit the standard casing dimensions. However, special sizes normally increase the cost and schedule. Drift diameters in these tables do not reflect weld-on, fatigue resistant connectors. The wellbay layout and seafloor spacing have primary impact on the size of the vessel and the method of running risers. While initial sizing of the risers may be performed independent of the wellbay layout, e.g. by reference to pressures and operating conditions alone, important parameters like riser stroke, local bending at the seafloor and the keel (in the case of a spar), process deck height, etc. will depend on these parameters. Also, the vessel cannot be sized until the wellbay is determined (see Chapter 7). This means that the vessel motions can not be finalised, and hence the final dynamic stresses cannot be determined. The importance of early consideration of the wellbay layout on the whole design of a floating production system cannot be overemphasised. Once the basic configuration of the number of tubulars, their makeup and a minimum ID for the inner riser are determined, analysis of a set of load cases is required to determine the controlling environment. At this point a selection of the governing design guideline is required, e.g. API RP2RD, OS-17201 or other. As was shown above, an LRFD code allows room for optimisation; however, the selection will usually be a function of the certifying agency and country and their familiarity with various codes. The following discussion assumes that API RP 2RD is the governing design code. Table 9.13 shows the recommended minimal design matrix in API RP2RD. Table 9.14 shows an example Load Case Matrix for the Matterhorn TLP. The factor C, or the allowable load stress increase, indicates the increase in the allowable load from the nominal value given in Section 5 of API RD 2RD. The basic allowable stress is
- 146. D r i l l i n gzyxwvutsrqp and Production Risers zyxwvutsr 2 3 4 zyxwvuts 5 zyxwvutsrqpon Table 9.13 Design matrix for rigid risers (API RPZRD) Extreme Extreme Design No 1.2 Extreme Maximum operating Extreme No 1.2 Extreme Maximum operating Design Yes 1.2 Temporary Temporary Associated No 1.2 193 Design Load Environmental case categoryzyxwvut ~ condition ~ Pressure ~ Reduced tensioner ~ Cph ~ capacity or one mooring line broken 11 1Operating 1Maximum operating 1Design 1No 11.0 I 6 lTestd 1Maximum operating 1Testd 1No 11.35 1 17 ISurvival 1Survival IAssociated 1No 11.5 I 18 1Survival IExtreme 1Associated IYes I 1.5 I i9 1Fatigue 1Fatigue 1Operating 1NO lNoteC1 Notes: Anisotropic materials may require special consideration W s e of Cris described in Section 5: strength issues are discussed is zyxwv 5.2. deflections in 5.3. collapse issues in 5.4 and 5.5, fatigue in 5.6 bPipeline codes may require lower C , for risers that are part of a pipeline ‘Not applicable dPlant testing for rigid risers should be agreed between user and manufacturer where C,=2/3, and oJis the material yield stress, defined in API RP2RD, for steel and titanium, as the stress “required to produce an elongation of 0.5% of the test specimen gage length”. API RP 2RD defines three stresses: primary membrane, primary bending and secondary. A “Primary” stress is “any normal or shear stress that is necessary to have static equilibrium of the imposed forces and moments. A primary stress is not self-limiting. Thus, if a primary stress substantially exceeds the yield strength, either failure or gross structural yielding will occur”. A “Secondary” stress is “... any normal or shear stress that develops as a result of material restraint. This type of stress is self-limiting, which means that local yielding can relieve the conditions that cause the stress, and a single application of load will not cause failure”. A primary membrane stress is the average value of the stress across a solid cross section, excluding effects of discontinuities and stress concentrations. For a pipe in pure tension this would include the total tension divided by the cross section of the pipe. For a pipe in tension and global bending, the membrane stress would include the global bending effect as well. The primary bending stress is the portion of primary stress proportional to the distance from the centroid of the solid section, excluding stresses due to discontinuities and stress concentrations.
- 147. 194 zyxwvutsrqp 100-yr.winter storm 1.2 and shut-in with PNO surface tree and completion tubulars PNO 100-yr. loop current 1.2 zy Chapter z 9 z ;:Pi l[ormal production P-N5 supported from the top of riser P-N4 zyxwvuts Table 9.14 Design load cases, Matterhorn TLP [Jordan, et a1 (2004)] PNO, PSI zyxw 1 1-yr. winter storm 11 PNO Cold core eddy 1.2 PSI 100-yr.hurricanc 1.2 PSI 1000-yr.hurricane 1.5 description P-K3 p - ~ 4 P-L1 p-L2(2) p-L3‘3’ p-~4(4) Design environment contents(’) increase factor completion tubulars supported from the top of the riser. Shut-in with the surface tree and completion tubulars supported at the top of the riser PK PK 100-yr. hurricane 1.2 1000-yr. hurricane 1.5 IPK Il-yr. winter storm 11 I IPK 1100-yr.loop current 1 1.2 I ~ PSLNO 195% non-exceedance 1 1.2 I 1PSL 1 l-yr. winter storm 11.2 I 1PSL 1 100-yr. loop current I1.5 I IPSL 1100-vr. hurricane 11.5 I ,Veilkilled with 1 ; ~ 10-yr. winter storm ~ lli 1 p-C3 supported from the 10-yr. hurricane p-c4 sap of the riser PK 100-yr. hurricane 1.5 BOP stack and completion tubulars 10-yr. loop current 1P-TD 1Tensioner damage(4) 1PSI 1 100-yr. hurricane I1.5 I Notes: The follouing arc definitions of the “Pressure & Contents“ column abbreviations: PNO PSI PK PSL PSLNO Normal operating surface shut-in tubing pressure with a tubing leak. To overcome this situation and replace the leaking tubing. a”Bu1lhead Pressure” must be imposed at the surface wellhead that is greater than the shut-in tubing. “Bullhead Pressures” will not be present during a hurricane when the platform is abandoned Add a 15 kip snubbing unit BOP, a 41.5 kips snubbing unit, and 25 kips of work string to the top of tree Operator can remove everything but the 15 kip snubbing unit BOP off the tree before a hurricane or a severe loop current situation Tensioner damage is defined as the loss of one tensioner element without any adjustment to the remaining elements Normal surface operating pressure under normal flowing conditions Shut-in tubing pressure with no tubing leak Well killed with heavy liquid in the tubing and the annulus Surface pressure shut-in tubing pressure with a tubing leak
- 148. Drilling and Production Risers 195 z Cf factor Primary membrane plus bending Primary membrane Stresses due to discontinuities and stress concentrations fall into the category of secondary stresses. Primary stress components are combined using an equivalent von Mises stress. 1.0 1.2 1.6 Cf (Sm) 53.3 64.0 80.0 1.5 Cf (Sm) 80.0 96.0 120.0 1 zyxwvuts Oe zyxwvutsr = - ( 0 1 - 0*12+( 0 2 - 0 3 1 2 +(03 - 01)* zyxwv A S J Primary membrane plus bending plus secondary Range of primary membrane plus bending plus secondary plus peak Average bearing stress (9.7) 3.0 Sm 1160.0 1160.0 160.0 Based on Based on Based on Based on fatigue curve ,fatiguecurve fatigue curve fatigue curve 0.9 Sy 172.0 72.0 72.0 where oe=von Mises equivalent stress, 01, 0 2 , 0 3 =principal stress. A common combination of stresses includes the hoop and axial tensile stresses, both of which are primary stresses. Average primary 10.6 Sm 32.0 32.0 shear stress (9.8) (9.9) (9.10) 32.0 where op=Primary membrane stress, o b =Primary bending stress, oq=Secondary stress. Table 9.15 summarises the allowable stress criteria for API RP2RD. Primary bending stress and secondary stresses are typically associated with changes in riser section near connectors and transitions. Their evaluation requires an assessment of the through thickness stress profile, and separation graphically or mathematically of the average, linear (bending) and non-linear components of the stress distribution. The example in Annex C of RP 2RD should be consulted for application of these criteria zy Table 9.15 Summary of allowable stress criteria Stress category Stress Allowable stresses (ksi) allowable Normal Extreme Survival operating event event
- 149. 196 zyxwvutsrqponm Chapter z 9 z Initial riser sizing, excluding stress joints, typically considers only the primary membrane stresses. The steps include the following: Select nominal tubular sizes (OD) as described above. Make a trial selection of IDS For each load case, Compute the top tension required to achieve a zero effective tension at the mudline (this is the weight in water of the submerged portion of the riser and its contents, plus the dry weight of the riser and contents above the waterline to the tensioner ring). Apply a nominal “tension factor” (TF) to insure positive bottom effective tension under dynamic loadings. The tension factor for floating production systems is typically in the range of 1.3to 1.6.However, the tension factor could change upon further analysis of riser interference or Vortex Induced Motions. This is an iterative process. The top tension is the tension required to yield a zero effective tension at the mudline times the tension factor. Considering the top tension and pressure in the riser, compute the combined primary stresses and a utilization factor Vary wall thickness and repeat the above procedure until zyxw U zyxw < 1.0. Selection of the material yield strength is required at this point. zyx As was mentioned above, weld-on connectors are presently limited to strengths below 95 ksi. For very high- pressure wells where dynamic stresses are likely to be low, higher yield strength may be selected provisionally to reduce riser tension. However, the fatigue of the riser couplings will need to be checked before this decision is validated. Ductility and toughness are also critical concerns for dynamic risers to avoid the possibility of brittle fracture. The majority of deepwater risers are designed for 80 ksi yield strength. Another important consideration in deepwater is the minimum effective tension at the seafloor. In deepwater the riser may have a negative effective tension without failing. This is because the bending that occurs is limited by the displacement of the top of the riser, i.e. it is a secondary stress rather than a primary stress. Detailed analysis may indeed indicate that suitable criteria may be met with reduced tension factors. This is part of an optimisation process. For low motion platforms such as spars, initial sizing for the main body of the riser might ignore dynamic effects. Experience may indicate that higher safety factors applied to this “static” riser sizing approach will lead to good results with a minimum of iteration. In any event, the next step is to perform dynamic analysis of the risers. This analysis requires 1. 2. Vessel motions Definition of the seastates corresponding to the load cases defined above,
- 150. Drilling and Production zyxwvutsrqp Risers I91 z The analysis may be frequency domain or time domain, coupled or uncoupled. Often, global vessel motions will be computed by one group and riser dynamics by another. This paradigm is very risky. For example, vessel motion programs often use different coordinate systems than riser programs. Translating motions from the origin of a vessel to the riser hangoff point using a different coordinate system can and often does lead to errors. It is best to perform a quality check of the procedure by analysing a simple case, e.g. a monochromatic sine wave, and performing some hand calculations prior to doing the bulk of the analysis with random wave input. Another issue with frequency domain analysis is that it can neglect important non-linear effects such as slowly varying motions and damping. Coupled analysis is the simultaneous solution of the vessel and riser motions. In deep water the riser loads may actually reduce the vessel responses [see, e.g. Prislin, et a1 (1999) Halkyard, et a1 (2004) for comparison of full-scale data with calculations]. The above discussion focuses on strength design. The riser may also fail from fatigue, hydrostatic collapse, buckling and thermal effects. Fatigue analysis needs to consider fatigue for vessel motions, wave loadings on the riser and vortex induced motions. The analysis is similar to that for the drilling riser discussed above; however, the difference is that production risers are designed to remain in place for the life of the field, whereas drilling risers are routinely retrieved and inspected. API RP2RD requires the fatigue life of the riser to be: Three times the service life (usually the life of the field) for areas accessible for inspection (or, where safety and pollution risk are low), or, Ten times the service life for areas not accessible for inspection (or, where safety or pollution risk are high). In practice, deep water risers are invariably designed for ten times the service life. Figure 9.50 shows the DNV fatigue curves used for the analysis of riser components. These are identical to the DOE curves discussed in Chapter 7 [DNV C N 30.2, HSE, 19951. Riser connectors without welds use the “B” curve with an appropriate stress concentration factor. Preliminary fatigue analysis is often performed using the “B” curve to determine an allowable SCF to achieve the required design life. If this allowable SCF is five or greater, it is likely that threaded couplings may be used for the connectors. New couplings with lower SCFs are becoming available as was mentioned above. The difference between the C, D and E curves, which apply to welded connections, depends on the quality of the weld. Special considerations need to be given to the presence of H2S and corrosion. These fatigue curves assume cathodic protection. The reader is encouraged to review the literature on riser sizing and analysis for deepwater floaters [e.g. Wald, et a1 2002; 0’Sullivan, et a1 2002; Jordan, et a1 2004; and Bates, et a1 20021,
- 151. 798 1 m . ozyxwvu Chaprer 9 z Figure 9.50 DNV fatigue curves 9.3.2.7 TTR Analysis Procedures The outer geometry of the riser is not uniform because of various elements attached to it. The equation of motion for a riser and its different components is given on the assumption that the riser represents a bent tubular member in one plane and only one plane of motion is considered. Similar equations may be applied to the orthogonal plane, and the two motions may then be combined with the coupling between them, coming from the external forces. The equation of motion is explicitly written as zyxw $ zyx 2 x ) (flexural rigidity) - (axial tension force) dxl d Y d - - [{A,( y)p,( y ) - Ai(y)pi(y ) }- (external & internal fluid pressure) dY +m( zyxwvu y)x (riser inertial resistance) =fxs(x,y, t) (external horizontal force) (9.12) Additional constraints are needed to solve this equation, which are specified at the top and bottom joints as end restraints. The restraints could be fixed, pinned, free or a specified top offset from the vessel displacement. This horizontal equation may be solved for both the static and dynamic analysis of the riser. For the static analysis, the fluid inertia is absent and the external loading is due to
- 152. Drilling and Production Risers zyxwvuts 199 z the current load. In this case, the equation becomes, (9.13) where the right hand side is the current force. zyxwv Feb) is called the effective tension due to axial and pressure force, zyxwvu Ub)is the current velocity as a function of the vertical coordinate y, and C D b )is the corresponding drag coefficient for the riser. In order to solve the dynamic problem due to oscillatory excitation, the right hand side of equation (9.12) should represent the dynamic load, e.g. from wave and vessel motion. The two solutions are combined into one, when the static and dynamic external loads, on the right hand side are combined. Either finite difference or finite element methods are used to solve for the deflected riser mode shapes and structural properties under static or dynamic loads. Because of its versatility, the finite element method (FEM) becomes an obvious candidate for the numerical tool. Indeed, most of the general-purpose riser analysis packages are based on the FEM, and the reader is referred to the vast literature that exists on the FEM for details of these analyses [see, for example, Meirovich (1997), and Moe, et a1 (2004)l. Example Problem: Transverse envelope The maximum and minimum transverse displacements of a top tensioned riser are computed for several current speeds. The following input are considered as example: z e e e e e e e zyxwvutsrq 0 e 0 e 0 0 e water depth zyxwvut = 100 m, riser length = 120 m, outside diameter =0.25 m, inside diameter = 0.2116 m, top tension =200 kN (-1.5 times the riser weight), modulus of elasticity of riser pipe=2.10 x 108 kN.mP2, specific weight of the fluid outside = 1025 kg.m-3, specific weight of the fluid inside =800 kg. mP3, specific weight of the riser wall material =7700 kg. m-3, mass, m* =2.7 damping parameters, m(* =0.054, riser model elements = 80 below, 20 above still water, riser ends =fixed but free to rotate, uniform flow velocities =0.16 to 0.93 mis, Reynolds numbers=4.0 x lo4 5 Re 5 2.3 x zyxw lo5. The results for the first four mode shapes are shown in fig. 9.51. The current speeds for these modes and the corresponding reduced velocities are included in the figure. The typical drag coefficients used for a production riser for a SPAR of draft of 2000 ft are summarised in table 9.16.
- 153. 800 zyxwvutsrqp Figure 9.51 First four mode shapes of the example top tensioned riser zyx Chapter 9 Empirical formula [API, 19921for the blockage based on the relative spacing of the risers with respect to the diameter is given by: 0.25SID for 0 zyxwv < S/D < 4.0 1.0 for S / D = 4.0 CBF= (9.14) where zyxwvuts S = centre to centre distance of risers of diameter D. The value of the current blockage factor for a row of cylinders is given in Table 9.17. 9.3.2.7.1 EfSective Tension The effective tension, physically a very meaningful quantity, represents a composite tension, which incorporates the effects of internal and external fluid pressures. It is defined for a single-walled riser as follows: Teff = Twall zyxwvutsrqp - p i 4 zyxwvu + P A (9.15) where Teff is effective tension, Twall is tension in the riser wall, p and A denote pressure and enclosed section area respectively and subscripts i and e mean internal and external respectively. In general, all of these quantities vary along the riser length. Broadly, Teffis used in force calculations and Twallis used in stress calculations. However, engineers are advised to familiarise themselves fully with these terms and associated interpretations, particularly with regard to multi-pipe risers, and are referred to Sparks (1983) for a detailed explanation.
- 154. Drilling and Production zyxwvutsrqp Risers zyxwvutsrq Straked section Bare section zyxwvu 801 1.4 450 Below SPAR keel 1.2 1450 Down to sea bed zy Table 9.16 Typical drag coefficients for a production riser zyx 4 6 1Riser section 1CD 1Length (ft) 1Remarks I End-on 0.80 Diagonal 0.85 Broadside 0.80 End-on 0.75 Diagonal 0.85 Broadside 0.80 IUmer section 10.9 I100 /Inside SPAR 1 8 End-on 0.70 Diagonal 0.85 Broadside 0.80 Table 9.17 Current blockage factor for cylinder group 1No. of cylinders 1Current heading 1Blockage factor 1 13 1All 10.90 i Care is required in communicating tensions to others, since many mistakes have been made. Where there is room for doubt (except discursively) it should always be made clear which tension quantity is meant. Special care is required for riser terminations, where load- paths are diverted and for riser connectors, where component manufacturers may assume a different terminology from analysts. 9.3.2.7.2 Soil Riser Modelling The ability to predict the behaviour of laterally loaded conductor casing embedded in the seabed is an important consideration in the design of conductor casing systems and in the prediction of lower flex-joint angle and wellhead bending moments. If the soil immediately below the mudline has low strength, as is frequently the case, little resistance is provided against lateral deflection in this region and the area of highest bending of the structural casing can occur some distance below the mudline. For this reason, the characterisation of lateral resistance of the soil near the mudline is an important input to a reliable structural model of a coupled casing-riser system. Under lateral loading, soils typically behave as a non-linear material, which makes it necessary to relate soil resistance to conductor casing lateral deformation. This is achieved by constructing lateral soil resistance-deflection zy (p-y)
- 155. 802 zyxwvutsrqpo Chapter z 9 curves, with the ordinate of these curves being soil resistance per unit length, p and the abscissa being lateral deflection, zyxwvu y. The analysis of such a problem may be accomplished by structural analysis of the casing structure using non-linear springs to model the p-y behaviour of the soil and by the solution of the following equilibrium equation: (9.16) where y =casing deflection, zyxwv x zyxwv =length along casing, EZ= equivalent bending stiffness of casing system, p =soil resistance per unit length. This equation is solved applying the casing geometry and soil stiffness boundary conditions, typically in terms of a family ofp-y curves developed for the soil. These p y curves, which represent the increasing non-linear soil stiffness with depth, are typically based on empirical formulations proposed for soft clay, stiff clay and sand respectively. The draft zyxwv API Technical Report (API 16TR1) provides guidance on the derivation of these curves [Kavanagh, et a1 20041. 9.3.3 Steel Catenary Risers (Portions contributed by Thanos Moros & Howard Cook, BP America, Houston, TX) The steel catenary riser is a cost-effective alternative for oil and gas export and for water injection lines on deepwater fields, where the large diameter flexible risers present technical and economic limitations. Catenary riser is a free-hanging riser with no intermediate buoys or floating devices. Flexible riser is a free-hanging riser with intermediate buoys or floating devices. See fig. 9.52 below. A typical profile of a SCR is shown in fig. 9.53. In 1998, a 10-in. steel catenary riser (SCR) was installed in P-18 platform, a semi- submersible production vessel moored in Marlim Field, at 910 m water depth. This was the first SCR installed at a semi-submersible platform. Figure 9.52 Free hanging SCRs with and without intermediate buoys
- 156. Drilling and Pvoducrion Risers zyxwvutsrq 803 z ";Lz 22 deg En End zyxwvutsr M zyxwvutsrq Figure 9.53 Typical profile of steel catenary riser Figure 9.54 shows the worldwide population of floating production systems (FPS) with steel catenary risers of 12 in. or greater in diameter. The FPSs are ranked by criticality of the SCRs in terms of water depth divided by the diameter of the largest SCR on the platform. The smaller this ratio, the more critical the touchdown point fatigue is likely to be. As can he seen from the chart, Typhoon's 18 in. gas export riser in approximately 2100 ft of water has the smallest ratio of any SCR installed to date. This section provides guidelines for the design of simple and lazy wave steel catenary risers (SCRs). Such risers are being considered or built for use in many deepwater fields. There are currently two dedicated riser design codes relevant to SCR design, API RP 2RD (1998) and DNV-OS-F201 (2001), and their scope is similar. They provide recommendations on structural analysis procedures, design guidelines, materials, fabrication, testing and operation of riser systems. The steel catenary risers (SCRs) are designed by analysis in accordance with the API codes (API RP 1111 and API RP 2RD) or the DNV codes. The analysis generally follows the following steps: Size the SCRs for pressure and environmental loads; Select the minimum top angle required for resisting environmental loads and providing adequate fatigue performance; Generate design parameters (angles and loads) for the flex joints and their attachments to the floater;
- 157. 804 zyxwvutsrqp Figure 9.54 Large diameter SCRs for FPS zyxw Chapter 9 Compute fatigue life of the SCRs based on “suitable” design fatigue curves for the proposed welds and assess the criticality of the welds; Compute cyclic load histograms for use in fracture mechanics analyses for defining weld acceptance criteria; Assess procedures for abandoning (laying on bottom), retrieval (lifting off bottom) and installation of SCRs; and Perform Vortex Induced Vibration (VIV) analyses to determine if vortex suppression devices were required and if so in what quantity. The principal difference between the codes is in the approach to structural design. API is based on a working stress design approach. DNV provides a limit state approach, which is less conservative, although a simplification to a working stress approach is allowed for in the document. There are several factors that influence the sizing of the riser diameter and its wall thickness. Some of them are the following: Metocean conditions, Host vessel offsets and motions, and
- 158. Drilling and Production Risers 805 Structural limitations zyxwvu - burst, collapse, buckling, post-buckling, Construction issues - manufacturability, tolerances, weld procedures, inspection. Installation method - tensioning capacity of available vessels, Operating philosophy - transportation strategy, pigging, corrosion, inspection, Well characteristics ~ pressure, temperature, flowrate, heat loss, slugging. The producing well characteristics determine variations in line contents and properties over time, which should be defined for operation in normal and abnormaljshutdown conditions. The designer should take into account the full range of contents for all stages of installa- tion, commissioning and operation. zyxwv 9.3.3.1 Influence of Construction/Installation Method The designer should take account of the effects of construction and installation opera- tions, which may impose permanent deformation and residual loadsjtorques on the riser system whilst consuming a proportion of the fatigue life. In-service requirements deter- mine weld quality, acceptable levels of mismatch between pipe ends and out-of-roundness, whilst NDT requirements are determined from fatigue life and fracture analysis assessments. The following, in particular, should be quantified: In collapse design, the effects of the sag bend strain levels during installation as well as extreme loading, shut down/depressurised and minimum wall thickness cases. Residual torque resulting from curves in the pipeline, installation vessel tensioner crabbing or plastic deformation during laying operations, as regards components such as flex-joints. SCFs from geometric discontinuities, regarding pre-weld fit up (hi-lo) limits resulting from out of roundness (UOE pipe), non-uniform wall thickness (seamless pipe) and tolerances of weld preps. Stress concentrations induced by plastic deformation during installation (reeling, S-Lay). Residual ovality induced by plastic deformation during installation (reeling, S-Lay). Installation load cases. Weld procedure and tolerances, NDT methods and thresholds, which should be related to the required fatigue resistance. Connelly and Zettlemoyer (1993) and Buitrago and Zettlemoyer (1998) may be found useful in the determination of SCFs for girth welds. Annealing after seam welding may reduce residual stresses with consequent improvement in hydrostatic collapse resistance. Mechanical properties of protective coating or thermal insulation systems should be able to accommodate all construction activities. For example, where thermally insulated risers are to be installed from a reel barge, environmental conditions at the spool base may differ considerably from those in the field, particularly if reeling is done in winter in northern Europe or the northern USA. External pipe insulation systems are often made up of several
- 159. 806 zyxwvutsrqpo Chapter 9 layers of material zyxwvu - with field joints having a different make up. How the system will behave, when reeled and unreeled, can only be reliably assessed by carrying out bending trials under the worst conditions (usually the coldest). zyxw 9.3.3.2 Geotechnical Data As an SCR comes in contact with the seafloor at the touch down point (TDP), an interaction (force-reaction) takes place between the riser and the seafloor. This interaction is usually characterised through the use of three sets of perpendicular “springs”, which represent the axial (or longitudinal), horizontal (or transverse lateral), and vertical (or transverse vertical) soil restraints against the riser motions. The soils at or close to the seabed in deep water are generally very soft, to soft clays, although the presence of sand layers cannot be discounted. The interaction of the SCR with the seabed will depend on the riser motions and soil conditions. The riser may cut a trench several riser diameters wide and may load or severely disturb soil up to 5 or more riser diameters below the mudline. It is therefore important that any geotechnical data that are obtained from the site are representative of the conditions within the riser zone of influence. Arguably the most significant soil parameters for modelling the interaction of the riser with a clay seabed are the undisturbed and remolded undrained shear strengths. However, other soil properties such as plasticity, particle size and permeability are important for characterising soil suction and dynamic response, including viscous damping effects. Soil chemistry may be important in some cases in designing for external corrosion. For sands the most important mechanical properties for assessing riser interaction are the relative density and permeability, as characterised by the angle of internal friction and the particle size distribution or grading. The definition and units of spring stiffness used in structural codes are not necessarily consistent, which may lead to misinterpretation and misuse. In order to reduce the risk of analytical ambiguities and errors, the units commonly used to describe soil springs are discussed below. Guidelines for selecting soil spring stiffness are also given. 9.3.3.2.1 Soil Springsfor Modelling Riser-Soil Interaction One of the simplest and most popular ways of modelling the support or restraint provided by soil surrounding a pipeline or riser pipe is by using discrete uncoupled soil springs. Many structural codes can handle non-linear soil springs such as those frequently used to model interactions with offshore piles and conductors, usually called zyx p-y curves (lateral springs ) and t-z curves (axial springs). Others may be limited to equivalent linear elastic soil springs, often referred to as Winkler Springs. Experience indicates that the definition and units of spring stiffness used in structural codes are not necessarily consistent. This may result in some misinterpretation and misuse, particularly if the spring stiffness is obtained from independent sources.
- 160. Drilling and Production Risers 807 z The aim of this section is to summarise the units commonly used for soil stiffness to reduce the risk of analytical ambiguities and errors. For this purpose it is helpful to assume the soil as elastic and to consider the classical problem of a flexible strip or beam on an elastic foundation. 9.3.3.2.2 Soil Stiffness The modulus of elasticity of an elastic material E is defined as: E zyxwvuts = zyxwv 0,’s (9.17) The units of E are Stress or Forcellength squared, e.g. kN,’m2. The deflection of a vertically loaded area supported on a semi-infinite elastic half space is related to E by the following expression: 6 = ZpqB(1 - v2)/E (9.18) where zyxwvut Zp =an “elastic” influence factor, q =the average stress applied over the loaded area, B =the width of the loaded area, v =Poisson’s Ratio. The deflection at the centre of a uniformly loaded flexible strip or beam, such as a riser pipe, on a quasi-elastic seabed is given by the following specific solution: 6 = 2qB(1 - zyxw v2)/E (9.19) The traditional way of defining soil stiffness for a beam on an elastic foundation is by the Modulus of Subgrade Reaction, Ksu, which can be obtained by re-arranging equation (9.19): E 2B(1 - v2) Ksu = q/6 = (9.20) The units of Ksti are Forcellength cubed, e.g. kN,’m3. zyxw An alternative way of describing the same soil stiffness is by Ku, defined as: KU= Q/6 (9.21) The quantity Q is the total load on the strip or beam so Q = B L q (9.22) where: B and L are the width and the length of the loaded area, respectively. Substituting for Q in equation (9.21) KU= BLq/F or Ku = B L KSU From equation (9.20) (9.23) (9.24) EL Ku=- 2(1 - zyx 9) (9.25)
- 161. 808 zyxwvutsrqpo Chapter 9 The units of Ku are force/unit length, e.g.kNjm and it is independent of the strip or beam width, B. The main potential source of confusion with units arises from a variant of equation (9.25) obtained by considering a unit length of the strip or beam, Le. by assuming L is 1.0. In this case equation (9.25) reduces to: E Ku” zyxwvut = ~ 2(1 - zyx v2) (9.26) The dimensions of Ku* are force/length squared (stress), but the actual units force/length/length. In the case of a riser pipe the units are force per unit deflection are Per unit length of pipe, e.g. kN/m/m length ofpipe. Note Ku* is also independent of width. The use of stress units for Ku* can be and has led to misinterpretation. Therefore, when expressing soil stiffness in this form it is important to use units of force/unit deflectionjunit length of pipe. Preliminary indications from recent research work are: (i) Soil stiffness under vertical compressive loading is important for wave-related riser fatigue. An increase in soil stiffness reduces fatigue life. (ii) Suction effects due to riser embedment appear to be less important for riser design, but may in some circumstances need to be accounted for. Interaction between the seabed and the riser is dependent on many geotechnical factors, including non-linear stress-strain behaviour of soil, remolding, consolidation, backfilling, gapping and trenching, hysteresis, strain rate and suction effects. Riser analysis codes presently in use have limited seabedjriser interaction modelling capabilities, but typically allow the use of soil springs to model load-deflection response and the product of submerged weight and friction coefficients or soil-bearing capacity theory to calculate maximum resistance force, as follows: Guidance on seabed friction coefficients can be obtained from BS 8010 (1973), which gives ranges for lateral and axial coefficients based on experience in shallower waters. However, as stated in the standard, these coefficients are an empirical simplification of actual pipe/ soil interaction, particularly for clays. Soil models that capture some of the key features of riser-clay interaction much better are currently being developed in recent industry research programs, such as “STRIDE” and “CARISIMA”. These models may include refinements such as soil nonlinearity, hysteresis, plastic failure, suction and viscous damping. Meanwhile, a simplified modelling approach combined with sensitivity checks that can bound the problem and identify key parameters can be used. An analysis method with a two-step approach is: 1. Conduct global riser analysis using simplified soiljriser interaction model - for example, linear elastic soil springs with maximum soil resistance based on sliding Friction coefficient for lateral movement across the seabed Friction coefficient for axial movement along the seabed Seabed resistance or stiffness to bearing loads.
- 162. Drilling and Production Risers zyxwvutsr 809 z friction or bearing capacity. In lieu of other data, a rigid or very stiff seabed is recommended for fatigue analysis, as this provides a conservative estimate of damage. Conduct analyses for critical storm load cases and fatigue sea-states, using a detailed soiljriser model, such as that being developed by STRIDE or CARISIMA. If a detailed model is not available, conduct sufficient analysis to bound the seabed interaction problems. 2. zyxwvutsr 9.3.3.3 Buoyancy Attachments and Other Appurtenances Lazy-wave risers are similar to simple catenary risers except that they have an additional suspended length that is supported by a buoyant section. This provides a compliant arch near the TDZ on the seabed. Analysis is required to optimise the arrangement and to define the required arch size and buoyancy distribution. The arch height and riser response can be sensitive to variations in the density of fluid contents. In addition, there may be a loss of buoyancy with time through water intake and degradation of the buoyant material. Analysis should be conducted to confirm that the riser has an acceptable response for the complete range of fluid contents and buoyancy. Buoyancy modules affect normal and tangential drag, mass and buoyancy upthrust. When modelling auxiliary buoyancy, consideration should be given to hydrodynamic loading at the bare pipe,/buoyancy interfaces. Buoyancy is typically supplied in modules that provide a discontinuous distribution of buoyancy and hydrodynamic properties. Analysis may be conducted assuming a continuous distribution. But it is recommended that sensitivity analyses be conducted to confirm that this is an acceptable assumption both in terms of storm and fatigue response. Care should be exercised in modelling to ensure accurate representation of (distributed) buoyancy, mass, added mass/inertia and drag. 9.3.3.4 Line-end Attachments SCR attachment to the floating vessel may be achieved using a flex-joint or a stress joint: Flex-joints. Correct understanding of the flex-joint stiffness is important in determining maximum stresses and fatigue in the flex-joint region. Flex-joint stiffness for the large rotations which typically occur in severe storms is much less than for the small ampli- tudes occurring in fatigue analysis. Temperature variation can also result in significant changes in flex-joint stiffness. In addition, it should be verified that the flex-joint can withstand any residual torque that may be in the riser following installation or released gradually from the seabed section of the line. Steps may be taken to relieve torque prior to attachment. Another design consideration for flex-joints, especially in high-pressure gas applications, is explosive decompression. Under pressure, gas may permeate into any exposed rubber faces of the flex-element. When de-pressurised, the gas expands and can migrate outwards with time. However, if the reduction in pressure is rapid the expanding gas can cause breakaway of the rubber covering the steel/rubber laminates. With repeated, rapid de-pressurisation, the steel laminates become exposed, the edges of the rubber laminates become damaged and functionality of the flex-joint is impaired. Explosive decompression risk is increased in
- 163. 810 zyxwvutsrqpon Chapter 5 z gas applications and at high pressures (say 3000 psi) may cause structural damage to the flex-joint. Suppliers may apply proprietary methods to avoid these problems. Stress joints - may be used in place of flex-joints, but they usually impart larger bending loads to the vessel. They are simple, inspectable, solid metal structures, and particularly able to cope with high pressure and temperature. They may be either steel or titanium, the latter material having the advantage of good resistance to attack from sour and acidic well- flows and, of course, gas permeation. Titanium gives lower vessel loads than steel and typically has better fatigue performance than steel. When conducting analysis with either flex-joints or stress joints as the attachment method, sufficient load cases should be considered to define the extremes of response. Angle change across the component is a key input for both types of termination, as are tension, pressure and temperature. An assessment of long-term degradation is also important from both a technical and an economic standpoint. zyxwv 9.3.3.5 Pipe-in-Pipe (PIP) SCRs Thermal insulation is required for some production risers to avoid problems with hydrate, wax or paraffin accumulation. The use of external insulation may in some cases impair a riser’s dynamic performance by increasing drag and reducing weight-in-water. However, PIP thermal insulation technology can often be used to satisfy stringent insulation require- ments (lower U-values) whilst maintaining an acceptable global dynamic response with the penalty of a heavier and perhaps more costly structure. Inner and outer pipes of a PIP system may be connected via bulkheads at regular intervals. Bulkheads limit relative expansion and can separate the annulus into individual compartments, if required. Use of bulkheads can be a good solution for pipelines, but for dynamic SCRs one must consider the effects of high stress concentrations, local fatigue damage and local increase in heat loss. Alternatively, regular spacers (centralisers) may be used, which allow the inner and outer pipes to slide relative to each other whilst maintaining concentricity. A detailed discussion of all analysis issues is beyond the present scope, but a checklist follows. The items listed are in the most cases additional to those for single-barrier SCRs and it is not claimed that the list is exhaustive. Also, according to engineering judgement, some of these effects may be omitted in the early phases of design, though justification for doing that should be given wherever possible. Residual curvature (which may change along the SCR) following installation Residual stresses due to large curvature history Residual axial forces between the two pipes Connection between the inner and outer pipes, including length and play of centralisers Boundary conditions and initial conditions at riser terminations Fatigue life consumed during installation Pre-loading of inner and outer pipes Axial forces and relative motions during operation, due to thermal expansion and internal pressure
- 164. Drilling and Production Risers zyxwvutsr 811 z 6x1 (XI (xi) (xii) (xiii) (xiv) (xv) (xvi) (xvii) (xviii) (xix) (xx) (xxi) Poisson’s ratio effect on axial strains Local stresses in inner and outer pipes due to centraliser contact, including chat- tering effects Frictional effects between inner and outer pipes Thermal stresses and thermal cycling effects Buckling checks (including helical buckling) due to thermal and general dynamic loading Soil forces on outer pipe Internal and external pressures having different effects on stress in inner and outer pipes Effect of packing material in reversal of lay direction on a reel should be assessed and cross-section distortion minimised; the pipe yields as it is reeled and it is very soft at the reel contact point Effects of PIP centralisers on pipe geometry during reeling Wear of centralisers Validity of VIV calculations (e.g. as regards damping) Possible effect of any electrical heating on corrosion rates Effect of damage (e.g. due to dropped objects striking outer pipe) on thermal and structural performance The capabilities of software intended for performing PIP analysis should be carefully considered, since commercially available programs vary widely in this respect. A PIP SCR can be modelled as a single equivalent pipe (EP), although it should be recognised that the technology is new, and careful attention must be paid to several aspects of the analysis. Development of stress amplification factors, to estimate loads and stresses in individual pipes following global EP analysis, is acceptable in the early stages of design. However, it is important to appreciate the conditions under which such factors become inaccurate. which will vary from case to case. Ultimately. full PIP analysis is required for verification. Two useful references on PIP SCR analysis are Gopalkrishnan, et a1 (1998), and Bell and Daly (2001). The first of these illustrates the large disparities, which can arise between the simplified EP approach and full PIP analysis, especially regarding the static stress. A JIP on deepwater PIP (including tests) named RIPIPE has been conducted by Boreas and TWI in the UK, and results will become public domain in due course. zy 9.3.3.6 Analysis Procedures The SCR is a 3-D structure, which in terms of design planning implies that directionality of loading (wave and current) must be included in the engineering analysis. Analysis methods for flexible risers include a complex finite-element structural method coupled with more simplistic hydrodynamic models (e.g. Morrison equation or potential theory). Empirically derived hydrodynamic coefficient databases are combined with the structural dynamic models. CFD method for computing excitation is combined with finite- element dynamic response analysis.
- 165. 812 zyxwvutsrqponm Chuptev z 9 z The analysis process typically falls into two phases. A preliminary analysis is performed in which the global behaviour of the riser is examined to confirm that the proposed configuration is practical and to provide preliminary data relating to key components in the system. A detailed analysis refines the definition of components and further examines all aspects of riser operations. In the preliminary analysis, the riser behaviour is generally examined in the normal opera- ting mode using extreme loading conditions, and design changes are made accordingly. Several combinations of riser configuration and loading conditions may be required to complete this initial assessment and to determine preliminary design load data for specific components. Initial VIV and fatigue life assessments should also be included. zy A flow chart showing the interaction between all aspects of the riser design and analysis is given in fig. 9.55. 9.3.3.7 FEA Codes and Modelling Methods A number of commercial finite element or finite difference codes are available that may be used for SCR analysis, mostly time-domain. Frequency-domain analysis uses minimal computational effort, but does not account for non-linearities in riser response. Time- domain analysis accounts for non-linearities in riser response and the computational time and effort, whilst much greater than the frequency-domain analysis, can be acceptably low. When modelling SCRs the element mesh should be refined at locations of high curvature and dynamic response; e.g. directly below the interface with the vessel and in the TDZ. Guidance is given in API RP 2RD on calculating the required element mesh. Appropriate convergence checks should be conducted in any case at a suitable stage in the analysis. Riser boundary conditions are the connection to the vessel and the termination and contact at the seabed, and care should be taken to model these accurately. Flex-joints can be modelled as articulation elements, and the designer should be aware of the sensitivities of flex-joint stiffness to both temperature and dynamic loading. Stress joints with a conti- nuous taper may be modelled as a series of stepped sections, again paying due regard to convergence as well as accuracy. The orientation of the vessel attachment can have a large effect on end loading and termination sizing and should be optimised. 9.3.3.8 Analysis Tools The software that are used in the design of risers are listed later. In the following. discussions are included in brief in order to illustrate general procedures for the analysis. For details of their capabilities the readers should consult the manuals of the specific software. Static configuration and mode-shapes should be calculated using an FE model. Alternatively, for quick evaluation studies only, an analytical solution to the catenary equation may be used. In-plane and out-of-plane mode-shapes should be calculated. Such externally generated mode-shapes can account for soil-riser interaction, property changes along the riser, lateral constraints. The FE model should properly model boundary conditions at the top of the riser. If a flex- joint is used, a suitable rotational stiffness should be implemented (stiffness depends on
- 166. Drilling and Producrion zyxwvutsrqp Risers zyxwvutsrq 813 Client Spocification zyxwvutsr & Docunimtatioti Conceptud Dcsign ' Global Configuration zyxw Preliminary Design ' Preliminat). IntwfaceData zyxw r> Analysis ' Prelitninar).Component &sign Data Component &sign Confirm Global -, Reports Detalled zyxwvuts D e s i g n Configuration Analysis 3 Confirm InterfaceData - - Component Btailzd Design Envelopes & Limitations Fabrication & Inspxtion Rapiremails Xianufacturing Procedures Installation Procedures, Limits and Contingencies QAW. Inspection V Standards & Critnia I I Risw &sign Report Figure 9.55 Flow chart on riser design response for flex-joints). The most appropriate bottom boundary condition may vary from case to case. If the modelled riser is terminated at TDP, the use of a fixed (built-in) end is likely to produce a lower fatigue life than use of a pinned end. However, it should not be assumed that a fixed end will produce the lowest of all possible fatigue lives; sometimes an intermediate stiffness case may be worse. Even for calculations with 2D currents, both in-plane and out-of-plane loading should be considered, which should yield reasonable accuracy, although it is not necessarily conservative compared to a true 3D behaviour. Out-of-plane loading (in-plane response) is often the most critical. For out-of-plane loading the current needs no modification. For in-plane loading the component of current normal to the riser axis should be used: V,- = Vsina (9.27)
- 167. 814 zyxwvutsrqp Chapter 9 Structural damping where V zyxwvut = horizontal current velocity, V,- = velocity normal to riser, CI = angle between riser and horizontal. Suppression devices are discussed later. The way in which they can be included in VIV modelling programs, such as SHEAR7, is evolving. Calculated damage for each profile should be factored by its probability of occurrence, then added to obtain the overall damage (taking account of the location on the riser periphery where damage accumulates). Sensitivity to profile shape and current intensity should also be evaluated. The following values of structural damping and stress concentration factor may be often suitable: 0.003 (Le. 0.3% of critical)* 1Parameters 1Value I Subcritical, Re < lo5 11.2 Critical, lo5< Re < lo6 Post-critical, Re > lo6 0.7 0.6-1.2 1.o 1.o 1.o zy 9.3.3.9 Hydrodynamic Parameter Selection Typical hydrodynamic coefficients for flow normal to the riser axis are given in table 9.18. Two exceptions to these general guidelines are: (i) (ii) For first-order fatigue analysis of non-VIV-suppressed riser sections, a CD= 0.7 may be appropriate. For straked risers or parts of risers where strakes have been applied, especially where Keulegan-Carpenter Number, KC, is low, an increased CDmay be appropriate and application-specific data should be sought. Further data on hydrodynamic coefficients for single risers and riser in arrays, showing dependence on KC, roughness, turbulence, spacing and strakes are also available (See Chapter 4). Effects, which can further influence the drag coefficient, are pipe roughness (due to marine growth, for example), VIV due to current or vessel heave, and interference from adjacent risers and structures. Reynolds number, in this regard, is defined in terms of the relative velocity between riser and water particle. The tangential drag of a riser is typically small as the structure is slender and the outer profile is even. Buoyancy elements, other appurtenances or marine growth can result in Table 9.18 Hydrodynamic coefficients for flow normal to riser axis 1Flow regime 1zyxwvutsrqponmlk CD' IC2 I Note 1: Reference area is area projected normal to riser axis Note 2: Reference volume is displaced volume of riser per unit length
- 168. Drilling and Production Risers zyxwvutsr Component Riser Riser/buoyancy interface zyxwv 815 C D zyxwv c a 0.01~ 0.0 zyx 0J4 Note zy 5 Table 9.19 Hydrodynamic coefficients for flow tangential to riser axis Note 3: Reference area is surface area of riser Note 4: Reference area is exposed annular area Note 5: Reference volume and C , to be agreed for each case local increase of tangential drag coefficient. Some typical values for modelling the buoyancy modules are given in table 9.19. Care should be taken to ensure that the reference areal volume associated with the hydrodynamic coefficients is correct for the software being used. Further guidance is provided in DNV Classification Notes 30.5, whilst the FPS 2000 JIP Handbook (2000) produced a wide-ranging survey of hydrodynamic data applicable to riser design and analysis. As a general rule, if doubt remains about the selection of CD,the value used should tend towards the conservative side. This means use of a higher value, when and where drag acts as an excitation and use of a lower value, if it acts to produce damping. An increased added mass coefficient, typically C, = 2.0 (Le. twice the value given in table 9.18) should be used for straked risers. 9.3.3.10 Sensitivities The sensitivity of riser response to changes in design and analysis assumptions should be evaluated. Parameters that should be considered include the following: Riser length - including installation tolerances, thermal expansion effects, tide and surge Weight - including corrosion, fluid density variations and slugging Flex-joint stiffness, including sensitivity to deflection, rate of deflection and temperature Seabed stiffness and soil/riser interaction effects Current directionality Drag coefficients Marine growth Vessel motion (draught and mass distribution dependence). Expected extremes of the parameters identified above should be incorporated into the riser model. This will allow the effects of parameter changes to be quantified. 9.3.3.11 Installation Analysis Limiting installation seastate or current, hand-over limitations and expected loads and stresses at each phase of the installation process should be established and the effect of installation methods and operations on fatigue life should be determined.
- 169. 816 zyxwvutsrqponm Chapter 9 z The installation analysis should establish functional requirements for installation equipmefit, identify operational sensitivities and establish limiting conditions and key hold points in the procedure. In addition, the analysis should identify contingency procedures/escape routes to be implemented in the event that safe operational limits are exceeded. Venkataraman (2001) discusses a number of issues relevant to installation of steel risers by reeling, iticluding reeling strain, buckling due to bending on the reel, strain amplification, elastic-plastic fracture mechanics, fatigue and hydrogen cracking. zyx 9.3.3.12 Extreme Storm Analysis Riser response is period and direction sensitive and highly dependent on vessel motions. Analysis using the maximum wave height with associated wave period may not result in the greatest response. Extreme storm analysis can be conducted using either regular or random waves. Regular wave analysis is a good preliminary design tool, as required design changes can be quickly evaluated. Regular wave analysis may be validated using random wave analysis, as the latter is able to provide a more realistic representation of the environment. However, if the wave period range is adequately covered, regular wave analysis is sufficient for early feasibility checks. Due to the period sensitivity of dynamic catenary riser systems it is recommended that a range of periods be analysed to confirm riser extreme storm response. A typical random wave analysis is: 1. Establish riser structural model. 2. Select spectrum type and parameters based on available data. Associated current and directional data should also be established. Representation of wave spreading is not usually required. Apply extreme vessel offset corresponding to assumed environmental conditions. Simulate random sea and calculate resulting vessel motion and riser structural response. In selecting the random sea, consideration should be given to its duration and its statistics. Where practical, these should be reported. Postprocess sample response statistics to estimate extremes (see below). Repeat for other cases, ensuring that period content is suitably represented. 3. 4. 5. 6. Compatible low-frequency motions may be included, depending on the software used. 9.3.3.12.1 Short-Term Extreme Responses Short-term extreme responses are those occurring in storms of relatively short duration, typically three hours. There is sometimes a need to post-process sample random dynamic analysis, results (or measurements) in order to establish an extreme response prediction. Alternatively, the sample extreme may be considered a good enough estimate. Unless it can be demonstrated that a simpler method, e.g. the Rayleigh method gives sufficient accuracy for the particular response under investigation (taking due account of
- 170. Drilling and Production Risers zyxwvutsr 817 z the stage to which a project or study has progressed) a (three-parameter) Weibull method is more appropriate. Analyses should be long enough to get satisfactory convergence of response statistics, noting that variability between different realisations of the same sea-state is reflected in the response, and extreme order statistics converge more slowly than lower order statistics. Convergence may best be achieved (and observed) by performing several different three- hour simulations for the same sea-state; i.e. using different seed numbers to produce different realisations of the same wave spectrum. 9.3.3.12.2 Long-Term Extreme Responses In principle, full prediction of long-term (e.g. lifetime) extreme responses requires probability-weighting and addition of all short-term extreme distributions, including those corresponding to the low and moderate wave height. However, this is usually not possible because not all of the short-term extreme response distributions will have been developed. In fact, very few may be available - perhaps some corresponding to an extreme wave envelope, perhaps just one or two. Some judgement is needed regarding the extent to which limited short-term data can be adapted and extrapolated to provide a suitable long-term extreme prediction. For a detailed coverage of this see Chapter 5. zy 9.3.4 Diameter and Wall Thickness The first parameter that should be determined for the design of an SCR is the wall thickness. The minimum wall thickness is calculated on the basis of external and internal pressure and buckling requirements. However, for SCRs the dynamic and fatigue life are, in most cases, the determining factor for the wall thickness [Chaudhury, 19993. This wall thickness should include for corrosion allowance. Initial wall thickness estimates are made using assumed riser loads. Further increases in riser wall thickness or changes of material grade may be required for a satisfactory response, based on later and more detailed analysis. Refer to Section 9.3.4.1 for further details. Wall thickness must account for all potential modes of failure as follows: Burst under hydrotest, In-service collapse, Collapse during installation, Burst at maximum internal pressure, Propagation buckling in-service and during installation', 'Propagation buckling checks may be performed, but need not be acted on for the dynamic part of the SCR unless required by regulatory bodies. The emphasis should be on preventing a buckle from initiating, although buckle arrestors in the static flowline, beyond the TDZ, may be advisable to prevent a buckle propagating between pipeline and riser. The primary function of buckle arrestors in pipelines is to restrict damage to limited lengths,which can then be replaced,whereas a buckle in a riser would require replacement of the whole riser, even if only a short length were damaged.
- 171. 818 zyxwvutsrqponm Chapter zy 9 Combined modes; e.g. external pressure with bending and tension. Calculations should allow for reduced wall thickness due to manufacturing tolerances, corrosion and wear, although corrosion may be neglected for installation and hydrotest conditions. Increased wall thickness may be required, perhaps only locally, to comply with dynamic response criteria. More generally, optimisation of wall thickness over the full riser length may help reduce cost and vessel interface loads. However, such an exercise is likely to be more beneficial for designs governed by collapse, and may yield no benefit at all for high-pressure cases governed by burst. zyxwvut 9.3.4.1 Nominal Wall Thickness The nominal wall thickness of pipeline is the specified wall thickness taking into account manufacturing tolerance. Corrosion Allowmce The external surface of submarine pipelines is generally protected from corrosion with a combination of external coating and a cathodic protection system. The internal surface, depending upon the service, may be subject to corrosion. This is accounted for by the addition of corrosion inhibitors or applying a corrosion allowance to the pipeline wall thickness. The corrosion allowance is calculated from the anticipated corrosion rate and the design life of the pipeline system. Manufacturing Tolerance Manufacturing or mill tolerances are the specified acceptance limits for the linepipe wall thickness during manufacture. The tolerance will depend upon the size of pipe and manufacturing process involved. A negative wall thickness tolerance should be taken into account when calculating wall thickness required for hoop stress criteria. The specified nominal wall thickness is calculated from the minimum required wall thickness as follows: (9.20) where ttol = negative manufacturing tolerance as specified by codes DNV, IP6 etc. Typical values for the wall thickness tolerance for seamless and welded pipe are &12.5% and & 5%, respectively. Consideration should be given to the nature and consequences of post-buckling behaviour. Under combined loading a pipe may buckle only locally in shallow water, but fail completely under the action of continuing hydrostatic pressure in deeper water. Guidance on wall thickness sizing against collapse and burst criteria is given in the following. This is drawn from several sources on standard pipeline practice and is
- 172. Drilling und Production Risers zyxwvutsrq DNV (1976, 1981) ASME 31.8 Det Norske Veritas, H+vik, Norway, 1981 and 1976. American Society of Mechanical Engineers, Liquid Transportation Systems for Hydrocarbons, Liquid Petroleum zy Table 9.20 Design codes zyxw I 819 Gas, Anhydrous Ammonia, and Alcohols. 1IP6 1Institute of Petroleum Pipeline Safety Code (UK). I BS 8010 (1973) DNV (2000) British Standard 8010, Code of practice for pipelines. Part 3. Pipelines subsea: design, construction and installation. Det Norske Veritas, Offshore Standard OS-F101, “Submarine Pipeline Systems”, 2000. 1 American Society of Mechanical Engineers, Gas Transmission and Distribution Piping Systems. suitable for initial sizing. However, project-specific requirements or guidance developed, more specifically for risers, may take precedence where there is justification. Propagation buckling, maximum D / t ratio, corrosion allowances, manufacturing tolerance, hydrotest pressure and API standard wall thickness are also discussed. The design criteria for the wall thickness calculations are as follows: zyx 6 Minimum wall thickness shall be the larger wall thickness determined from the above design criteria. The design codes in table 9.20 are used for wall thickness design: These codes are briefly covered below, except DNV (2000), which is relatively new and applies a Limit State approach. Its relevant section is referred to for each of the above design criteria. Section zyxwvu 5, fig. 5.3 of DNV (2000) gives an overview of the required design checks. Limiting hoop stress due to internal pressure. Hydrostatic collapse due to external pressure. Buckle propagation due to external pressure. 9.3.4.2 Maximum Diameter to Thickness Ratio The pipeline and riser wall thickness may be specified independently of the static design criteria due to installation stress limits. If the pipeline or riser is to be installed by the reel method, a maximum diameter to thickness ratio is recommended to avoid excessive out of roundness of the line during reeling. The ratio will depend upon the line size, reel diameter and total length of line to be reeled. As a general guideline, a diameter to thickness ratio of less than twenty three (23) is used for reel barge installation. The American Petroleum Institute (API) specification for line pipe is based upon a range of standard diameters and wall thickness. These values are different for imperial and metric sizes. Pipe mill tooling and production is set up around this specification.
- 173. 820 zyxwvutsrqpo Chapter zy 5 Non-standard line sizes are, sometimes, used for risers, where a constant internal diameter is specified, or for alloy steels, which are manufactured to special order and sized to meet production requirements. zyxwvu API RP 2RD (1998) also gives recommendations on collapse pressure and collapse propagation. 9.3.4.3 Resistance to Internal Pressure (Hoop Stress Criteria) Two load cases, namely, maximum design pressure and hydrotest pressure, need to be considered with respect to resistance to internal pressure. Design codes and standards stipulate that the maximum hoop stress in a pipeline shall be limited to a specified fraction of yield stress. The design pressure used in the analysis is based upon the maximum pressure occurring at any point in the pipeline and riser system. The maximum operating pressure will be limited by pump capacity or reservoir pressure and determined during a hydraulic analysis of the system. Design pressure may also take into account the transient surge pressure effects due to valve closure or shutting down of the transfer pump. The minimum or nominal wall thickness required to resist internal pressure may be cal- culated from any of the formulas given in table 9.21 below. Alternatively DNV-OS-FlO1 Table 9.21 Formulas for nominal wall thickness Code zyxwvuts 1Formula IP6 DNV zyxwvut (Pi- Po) 2nhoy BS 8010 (1973) ?,In = ~ Dnom +tcor I I - t ' - t +tcor Comments I Ratio D,,, /tmin greater than 20. Ratio D,,, Itmi,, less than 20. Positive root of quadratic equation is used. Next Page
- 174. Drilling and Production Risers zyxwvutsrq Riser Linepipe Det Norske Veritas, DNV 1981 0.5* 0.72* 821 z Table 9.22 Usage factors for internal pressure zyxw 1Design code IUsage factor zyx (nh) I lANSI/ASME 31.4 & 49 CFR195 10.6 10.72 I /ANSI/ASME 31.8 & 49 CFR192 Ios# 10.72# 1 1British Standard 8010, Part 3 10.6 10.72 I * DNV (1981) specifies the riser (zone zyxwvut 2) as the part of the pipeline less than 500 m from any platform or building and the pipeline (zone 1) as the part of the pipeline greater than 500 m from any platform building =ANSI ASME 31.8 specifies the riser zone as the part of the pipeline. which is less than 5 pipe outside diameters from the platform and the pipeline zone as the part of the pipeline. which is more than 5 pipe diameters from the platform can be used. It must be observed that DNV and ASME codes specifically refer to a nominal wall thickness, while IP6 and BS 8010 (1973) refer to a minimum wall thickness. If a minimum wall thickness is specified, the nominal wall thickness may then be calculated using a corrosion allowance and a manufacturer’s tolerance (see Section 9.3.4.1). In table 9.21 t,,, = corrosion allowance, t,,, = minimum wall thickness, t,,, = nominal wall thickness, zyxwvu D , , , = nominal outside diameter, zyxw oL= specified minimum yield stress (SMYS), nh = usage factor or fraction of yield stress, P,= internal design pressure, Po = external design pressure. The usage factor nh, which is to be applied in the hoop stress formulae, is specified by the applicable design code and the zone or classification of the pipeline. For submarine pipelines and risers the code requirements governing design usage factors are summarised in table 9.22. Temperature de-rating shall be taken into consideration for the risers and pipelines operating at high temperatures (typical > 120’C). The pressure containment check (bursting) should be performed according to DNV 2000 DNV-OS-F101 Section zy 5 D400 [equation (5.14)]. 9.3.4.4 Resistance To External Pressure (Collapse) Two load cases need to be considered with respect to resistance to external pressure: In-service collapse Collapse during installation Failure due to external pressure or hydrostatic collapse is caused by elastic instability of the pipe wall. For wall thickness determination the external pressure is calculated from the Previous Page
- 175. 822 zyxwvutsrqpo Chaptev zy 9 hydrostatic head at extreme survival conditions. The maximum water depth taking into account the maximum design wave height and storm surge should be used. The minimum wall thickness required to prevent hydrostatic collapse is determined from Timoshenko and Gere (1961) for DNV, API 5L and IP6 and from BS 8010 (1973). In the first case, the Timoshenko and Gere (1961) formula is used to calculate the minimum wall thickness as follows: zyxwvu P; zyxwvut - [20) (2yrn) - + (1 +0.03e- (9.29) where Po=external hydrostatic pressure, P, =critical collapse pressure for perfectly circular pipe given by: (9.30) zy v = Poisson's ratio, E =Young's modulus of elasticity, E =eccentricity of the pipe (Yo) (see below), 0,= specified minimum yield stress (SMYS). BS 8010 advocates use of the formula described in Murphy and Langner (1985) and this is described as follows: where the notation is as above and in addition, (9.31) (9.32) zy f, = the initial ovalisation (see below). zyxwv 9.3.4.5 Pipe Eccentricity, Out of Roundness and Initial Ovalisation Pipe eccentricity is a measure of pipe out of roundness. This is generally a specified manufacturing tolerance, which is measured in a different way depending on which code or standard is used. The various ways of measuring it and permitted values are given by various codes. The out of roundness definitions and tolerances with reference to API, DNV and BS 8010 are given in table 9.23. Summarising: IP6% eccentricity = API% out of roundness 2 x API% out of roundness = DNV% out of roundness BS 8010% initial ovalisation = API% out of roundness
- 176. Drilling and Production Risers zyxwvutsr 823 z Table 9.23 Out of roundness formulas zyxw 1Out of roundness or initial ovalisation 1Code or standard ITolerance 1 x IUU and x IUU D n o m 1 D n o m I IAPI 5~ D n o m - Dmin ~ ''Oy0 ~ x 100 and 'Dmax - D n o m 1 D n o m D n o m 1 DNV &ax - Dmin Dmax +Dmin BS 8010 (1973) 2.5% 1 l l The combined loading criterion is to be performed according to DNV 2000 DNV-OS-F101 Section 5 D500 [equations (5.22)-(5.26)]. If the riser is in compression between the supports the global buckling shall be checked according to Section 5 D600. 9.3.4.6 Resistance to Propagation Buckling Two load cases need to be considered with respect to resistance to internal pressure: in-service and during installation. The required buckle propagation wall thickness is the wall thickness below which a buckle, if initiated, will propagate along the pipeline or riser, until a larger wall thickness or a reduced external pressure is reached. The wall thickness required to resist buckle propagation can be calculated from the formulas in table 9.24. Generally, the DNV formula is marginally more conservative than the Shell Development Corporation formula and they are both considerably more conservative than the Battelle formula. The degree of conservatism required depends on the installation technique in terms of risk, the length and cost of the line and the water depth in terms of how easily a repair can be made. In practical terms: changing of any criteria will change the required wall thickness. However, since designers are normally limited to selecting from API pipe sizes, there is quite often no actual change in the pipe specified. If during design, a pipeline is found to be governed by the buckle propagation criteria, then there are two options open to the designer; the first option is to make the wall thickness of the pipe sufficient so that a buckle once initiated will not propagate. The second option is to make the wall thickness of the pipe sufficient to only withstand external pressure (hydrostatic collapse) and to use buckle arresters. Buckle arresters consist of thick sections of pipe or welded fittings, which a buckle cannot propagate through. If buckle arresters are fitted, the damage will be limited to length of the pipeline between arresters, should a buckle initiate.
- 177. 824 Table 9.24 Wall thickness to resist buckle zyxw Chapter z 9 z Code or standard zyxwvu 1Formula 1 Remarks DNV 1981 and DNV 1976 Conservative tnom =- kDnom l + k = F 1.15~~0, Battelle Columbus Laboratories [Johns, et a1 19761 Shell [Langner, 19751 11.0% BS 8010 Part 3 (1973) Note: The propagating buckling check should be performed according to DNV 2000 DNV-OS-F101 Section 5 D500 [equation (5.27)] This risk, however, is considerably reduced after installation. The choice between the two is determined by considering the potential cost saving in wall thickness and possibly installation benefits due to the reduced submerged weights. This is paid off against the risk of having to replace a relatively large section of pipe or riser possibly in deep water. 9.3.4.7 Hydrotest Pressure The BS 8010 (1973) design code gives criteria to calculate the minimum hydrostatic test pressure for a pipeline. The test pressure required to qualify for a MAOP (Maximum Allowable Operating Pressure) equal to the specified design pressure is either the lower of 150% of the internal design pressure, or the pressure that will result in a hoop stress (based on specified minimum wall thickness) equal to 90% of the specified minimum yield stress. The test pressure should be referenced to the Lowest Astronomical Tide (LAT) and due allowance made for the elevation of the pressure measurement point and parts of the system above LAT. For definition of design pressure see DNV 2000 DNV-OS-F101 Section 1. Hydrotest criteria are discussed in Section 5 B200. 9.3.5 SCR Maturity and Feasibility Three views of SCR maturity and feasibility are given in figs. 9.56 and 9.57. Figure 9.56 shows the existing SCRs against diameter and water depth. The choice of diameter and depth as axes is largely motivated by collapse considerations, although installation capabilities are also relevant. Figure 9.3 shown earlier (similar to fig. 9.56) puts more emphasis on water depth and diameter than on waves and vessel-types. It also shows recent
- 178. Drilling zyxwvutsrqpon and Production Risers zyxwvutsr 825 z Figure 9.56 Existing SCRs and technology stretch Figure 9.57 Estimated feasibility [from Spolton and Trim (2000)l and maturity of SCRs [Note: feasibility colour scheme developed for 10-in. HPHT oil production SCRs (left shading zyx = steel, middle light shading =steel-titanium, right = unproven)] zyxw chronological development of flexible-pipe riser regimes, indicating that feasibility for typical production sizes is now approaching 2000 m water depth. In fig. 9.57 [Spolton and Trim, 20001, titanium parts were progressively substituted for steel parts as environments and vessel motions considered became worse. Thus, in the left and
- 179. 826 zyxwvutsrqpon Chapter z 5 Extreme Survival the lower light shading parts of the chart, relatively small amounts of titanium are used; e.g. in TSJs and TDZs, whereas the top right region represents SCRs composed mostly or completely of titanium. These figures are guidelines only but give a reasonable first impression of what has been and what can be achieved, and what “technology stretch” is required for harsher conditions. Soil-structure interaction and VIV effects (separately and together) are major uncertainties remaining in the SCR design and analysis, although considerable progress has been made in both areas through the STRIDE and CARISIMA JIPs (among others). Findings continue to be extended and assimilated into mainstream practice and software codes. Trenching of SCRs in the TDZ has been observed and may represent an additional critical case for careful examination during detailed design. 100 yr. Associated Associated Product Design pressure Associated 100 yr Associated Product Design pressure Associated Associated 100 yr. Product Design pressure 100 yr. Associated Associated Product Failed mooring 100 yr. Associated Associated Variable Accidental** 1000 yr. Associated Associated Product zyx - zy 9.3.6 In-Service Load Combinations The in-service design cases of table 9.25 should be assessed for the most severe loading direction, which may vary according to the response quantity of interest. Allowable stresses are given in table 9.26.Latest versions of the API and the DNV riser codes, such as API RP 2RD, and DNV-OS-F201 are commonly used. Von Mises stress is defined in accordance with API RP 2RD (1998) as: von Mises stress = [(S- h)’+(S - ~ ) ~ + ( h - zy Y ) ’ ] /; (9.33) Table 9.25 In-servicedesign cases 1Design case 1Wave 1Current 1Wind IContents 1Other* I Normal To be agreed Associated Product Design pressure operating (typically 1 yr.) IPressure test ~ 1 yr. 1Associated Associated 1Water 1Test pressure 1 *Use associated pressure for survival. Maximum pressures are given; i.e. lower pressures should also satisfy checks **Accidental conditions are discussed further in Section 9.3.7
- 180. Drilling and Production Risers zyxwvutsr Normal operating Extreme zyxwvut Table 9.26 Allowable stresses 0.67 1.o 1.o 0.8 1.2 zyxw ~ 1.5 1 827 1Design case Von Miseslyield 1Survival 11.0 11.5 1Pressure test 10.9 1 1.352 11.35 1 ‘API RP 2RD (1998) definition. Plain pipe allowable stress is 2,3 yield x C , ’At the riser top the distinction between load- and curvature-controlled stress may not be clear. ’Where primary membrane stress exceeds yield (corresponding to C, = 1.5) a strain-based If so. stress should be considered load-controlled and C, reduced to 1.2 formulation should be used in which the strain at yield is substituted for the yield stress. Non-linear strain analysis is then required in order to demonstrate compliance. Also, for any case where yielding is predicted, further consideration and consultation should take place. and the higher value of zyxwvutsr C, = 1.8 for Survival may be acceptable if the exceedance is isolated. In general, it should not be assumed that increased sag bend factors can always be used; the effects of the various forces and motions applied to the riser should first be carefully considered where S = stress due to equivalent tension and bending stress, r = radial stress, h = hoop stress: T zyxwvut (do- t)M 21 s=- zyxw f A (9.34) Most sag bends are predominantly curvature-controlled, not load-controlled, and higher bending stresses are then allowable, since yielding does not of itself constitute failure. Increased values of API RP 2RD (1998) Design Case Factor zyx Cffor curvature-controlled sag bends are shown in table 5.2. In accordance with the API RP 2RD (1998), tangential shear and torsional stresses are not included and can be treated as secondary stresses, which are self-limiting. Torque, however, can influence the integrity of flex-joints (see Section 9.3.3.4). The term “associated” in table 9.25 is defined in API RP 2RD (1998) as “to be determined by considering joint wind, wave and current probabilities”. Often a 10-yr return period is assumed, unless there is a very strong correlation (positive or negative) between these items, or, project-specific requirements dictate otherwise. Associated Pressure is the greatest pressure reasonably expected to occur simultaneously with survival environmental conditions.
- 181. 828 zyxwvutsrqpo Chapter z 9 Other design considerations include flex-joint rotational limits, interface loads, compres- sion in the TDZ, tension on flowlines and clearance from vessels, mooring lines, umbilicals and other risers. zyxwvut 9.3.7 Accidental and Temporary Design Cases A failed mooring line with a 100-yr wave condition is an accidental design case typically used in SCR design, table 9.25. However, one failed mooring line is not the only potential failure mechanism that will have an effect on riser integrity. Other accidental design cases applicable to SCRs are listed as follows: zyxw 0 Two or more failed mooring lines Breached hull compartments 0 Failed tethers Internal pressure surge The likelihood of each accidental design case needs to be addressed on an individual basis. For example, two failed mooring lines combined with a 100-yr wave condition may be highly unlikely, especially if a failure is fatigue and not strength related. In this case an increased design allowable or less severe environmental condition may be considered. The likelihood of each accidental design case may be defined with a quantitative risk assessment. For guidance on analysis and criteria for temporary conditions; e.g. transportation and installation, see Section 4.3.3and tables 1 and 2 of API RP 2RD (1998). When calculating fatigue in towed risers, due allowance should be made for variability of environmental conditions and uncertainties in forecasting weather windows. Additional damage may be justified if there is a realistic risk that changing weather conditions will force an altered course or return to port. 9.4 Vortex Induced Vibration of Risers 9.4.1 VIV Parameters Important hydrodynamic quantities that influence VIV are: 0 0 Reynolds number, 0 Lift coefficient, and 0 Correlation of force components. For the hydrodynamic design a few important non-dimensional numbers in fluid-induced vibration are given in table 9.27. If zyxwvutsr VR<10, there is strong interaction between the structure and its near wake. If VR<1, VIV is usually not critical Shedding frequencies and their interactions, Added mass (or mass ratio) and damping,
- 182. Drilling and Productiori Risers zyxwvutsrq includes hydrodynamic Mass ratio (total mass added mass) zyxwvut Table 9.27 Basic non-dimensional VIV parameters m* zyxw = --m l total mass structure length - z pD2 829 1Structural aspect ratio zyxwvu 1 a* = - D Damping ratio 1Reynolds number i R e = T UD Strouhal number - related to the fluid 1 3=? path length per cycle structure width U ~ Reduced velocity - related to the structure - _ f D 1 Vortex shedding regions may be checked on the basis of fig. 9.58. The figure suggests that for a fixed cylinder, the vortex shedding frequency is proportional to the fluid velocity. For a cylinder at the intermediate Reynolds number of 1.18 x lo5< Re < 1.91x lo5 and St = 0.2, Le. the vortex shedding frequency is unaltered. At the transition region for Re of 105-5x106 the shedding frequency has a scatter and is broad banded. Note that drag coefficient also dips in this range (drag crisis). For large amplitude motion of cylinder, the shedding is correlated along the span and vortices become two-dimensional. 9.4.2 Simplified VIV Analysis The VIV of riser may be investigated by a simple Wake Oscillator Model (fig. 9.59). For fixed cylinder or small amplitude motion, the vortex shedding along the cylinder span is uncorrelated (no fixed-phase relationship). Equation of motion for the above model is written as (9.35)
- 183. 830 zyxwvutsrqpo Chapter 9 z Figure 9.58 Reynolds vs. Strouhal number for a fixed circular cylinder zy T Jzyx DISPLACEMENT ( Y I f -- zy 1: - UNIFORM -__ VELOCITY IU L E N G T H i t ) DIAMETER K t 5 P - d / ---. z Figure 9.59 Wake oscillator model zyx oL= cylinder natural frequency, zyxwv u3, a4 = dimensionless constant, 3 = transverse fluid flow velocity in the wake, G = transverse fluid flow acceleration in the wake. The quantities Ti, and ib are functions of the shedding frequency, which depends on UID. Note that the fluid force on the right hand side is inter-dependent on the cylinder motion. As the natural frequency of fluid oscillation approaches the natural frequency of the cylinder, resonance occurs. (See Blevins, Flow-induced Vibration, pp. 25-32 for details.) Solving equation (9.35), the amplitude ratio is given by 0.03 + 0’07’ [ (6, + 1.9)St _ - - A, zyxwvu D (6, + 1.9)St2 (9.36)
- 184. Drilling and Production Risers zyxwvuts 10 1.37 zyxwvutsrqp $1-zyxwvutsrqponml 831 - - -THEORY zyxwvut 0 RIGID CYLINDER EXPERIMENTS 0 PIVOTED ROD EXPERIMENTS zyxw 1 1 CABLE EXPERIMENTS - - 1 - - - - r 0.1 0.81, zyxwvuts 8 STRING OR CABLE CANTILEVER BEAM, 1ST MODE zyxwvut 1.305 - CANTILEVER BEAM, ZNO MODE 1.499 CANTILEVER BEAM, 3R0 MODE 1.537 zyxwvu j 7 1.155 1.291 STRUCTURAL ELEMENT - - RIGID CYLINDER - PIVOTEDROO SIMPLE SUPPORT BEAM 1.155 0 0 I I I l l I I I l l I I I Figure 9.60 Amplitude ratio vs. reduced damping (applicable 200 < Re < 2 x lo5) where y =shape factor (function of mode shape) and 6, =reduced damping, (9.37) The results are shown in fig. 9.60. The analysis shows that the amplitude ratio decreases with increasing mass ratio and increasing damping. An FE analysis problem was run with a TTR in shallow water, in which the transverse envelope, maximum and minimum transverse displacements were computed for the riser subjected to uniform current. The results of the anaylsis are shouwn in table 9.28. It shows the predominant modes of vibration for various current speeds, the corresponding reduced velocity and frequency of vibration. For the discussion purposes assume that the value of St = 0.2. Then, the Strouhal number realtionship, St =fsD/U where D is the riser diameter and U is the current velocity provides the vortex-shedding frequency f s . For example, D = 0.25m, U = 0.23m/s gives fs = 0.18 vs. f r = 0.15 D = 0.25m, U = 0.40m/s gives zyxw fs = 0.32 VS. f i = 0.20
- 185. 832 lU(m/s) Mode zyxwvut V, zyxwv fi U (mls) Mode Vr 0.16 lSt 4.51 0.22 0.54 3'd 4.00 0.23 1*I 6.76 0.15 0.62 3'd 4.57 0.31 zyxwvuts I 2nd 4.00 0.25 0.70 3'd 5.15 Chapter 9 z f r I 0.25 0.22 0.19 0.39 10.40 10.47 Znd 5.00 0.20 0.78 3'd 5.72 0.17 I 2"d 5.14 0.20 0.86 4th 4.03 0.25 1 2nd 6.00 0.17 0.93 4th 4.40 0.23 1 Therefore. in the first case, we have a VIV lock-in, while the second case shows that lock-in is avoided. zyxwvuts 9.4.3 Examples of VIV Analysis A typical example of VIV analysis is illustrated below. An analysis by VIVANA for the deflected shape of an SCR is shown in fig. 9.61. The analysis results are compared with model towing test for a towing speed of 0.13 m/s. The Strouhal number was calculated to be 0.24 for this example. 9.4.4 Available Codes There are many design codes available for the analysis of risers. A few of those are listed in table 9.29. The details of the capabilities of these codes may be obtained from their websites. 9.5 VIV Suppression Devices Several types of vibration spoilers are used in the offshore industry. To prevent the VIV and lock-in, vortex suppression devices interrupt the regularity of the shedding and stop vortex streets from forming. In a test program at the US Navy facility with cylinders in steady flow, a fiberglass cylinder model was built with a super smooth ground surface. The tests in supercritical Reynolds number demonstrated the absence of VIV. VIV of risers can cause high levels of fatigue damage but can be reduced using suppression devices such as: Strakes Fairings Shrouds The typical cross-section of a streamline fairings, such as rudders, fins, etc. (taper ratio > = 6 to 1) shown in fig. 9.62 is effective for VIV suppression of a marine riser. The slender
- 186. Drilling and Production Risers zyxwvutsr 833 z 1.n zyxwv 0.9 -0.8 zyxwvu v c : QI QI k zyxwvu 0.7 m g 0.6 2 0.5 E +” . 5 = zyxwvutsrqponm 6 Q) QI g 0 . 4 zyx y 0.3 - 0.2 0.1 0.0 n 0.1 0.2 fi,3 0.3 RMS DiapI.lDiarn. Figure 9.61 Measured and predicted transverse displacement of an SCR [Lie, et al (2001)j Glasgow (UK) Science tower was designed in the shape of an aerodynamic foil and allowed to rotate 360” with the mean wind direction with the help of a turntable at its base. One of the most common types of VIV suppression devices for the production risers is helical strakes (fig. 9.63). The width of the strake is typically about 10% of the cylinder diameter. Strakes generally increases the overall drag force as well as the hydrodynamic damping of the riser, which are counteracting for the motion of the riser. In developingidesigning a riser the questions to ask on VIV are: Is VIV a problem for the riser under the given environment at the site? 0 If VIV is a problem, will an alternative riser design avoid the problem? If suppression is necessary, what is the best practical method available? Analysis should account for effects of suppression devices on riser behaviour, via changes in weight and hydrodynamic coefficients. VIV-suppression strakes are an incorporated design element in all SCRs (fig. 9.64). Various manufacturers offer these strakes. A contracting philosophy needs to be prepared before ordering these elements. Nominated strake manufacturers should have wet tank test results of their product design in a similar diameter application, which demonstrates their efficiency.
- 187. lSoflwarezyxwvutsrqp DH zyxwvutsrqpo I BPP DNV ZENTECH MIT ~- .~ DeepVlV Numerical dynamic http://www.dhi.dk/consulting/ pipeline-seabed interaction offshore/pipelinesrisers/ Prediction of riser displacements and stress status zyxwv http://www.bpp-tcch.com/vcrtical.htm -.~ ~. http://www2.dnv.com/elni b/ http://www.zentech.co.uk/flcxrise.htm ~ ~ ~- Static and dynamic analysis of flexible risers Riser analysis einail: kimv(u)mit.edu ~- ISkaas NOBSystem DEEPLINES Mentor subsea risers ~- Riser analysis Seaflex FLEXRISER Shear zyxwvutsrqponm I Table 9.29 Some availablesoftware for riser analysis Source IDescription ~ 1Website IFP VIVs as wcll as fluid-structure interactions in riser bundles http://www.ifp.fr/IFP/en/rcsearchindustry/ Global maritime 1Riser and station keeping advisory systems Ihttp://www.globalmaritime.com/softwarc/ Analysis of floater motions and mooring-riscr system response http://www.name.ac.uk/rescarch/off-eng/ 1http://www.hks.com/ I Orcina Ltd. Analysis of floater motions and mooring-riser system response http://www.orcina.com/OrcaFlex/ MCS International 1 1http://www.mcs.com Design and code checks; J-pull ~ J-tube analysis; riser calculation Deepwater riser design product http://portal.woodgroup.com/pls/portal30/ url/page/external- jpkenny-home/techsoft http://www.stress.com/oiIgas/riser-tech. htm services, Inc. PRINCIPIA Global analysis of risers, moorings and flowlines http://www.principia.fr/principia- deeplincs-eng.html J. Ray McDermott interaction and buckling Induced vibration, soil structure http://www.jraymcdermott.com/mcntor/ mentor-riscrs.htm
- 188. Drilling and Production Risers zyxwvut 835 z I zyxwvutsrq Figure 9.62 Streamline geometry Figure 9.63 Strakes on risers Figure 9.64 The Prince SCR during hang-off installation showing the pre-installed VIV strakes [Gore and Mekha, 20021
- 189. 836 zyxwvutsrqponm Chapter z 9 z Strake suppression efficiency (percentage reduction of motion amplitude compared to bare pipe) depends on pitch (P)and height zyxwv ( H ) . Common values are PID= 17.5 and HID zyxwvuts =0.25 (where D is hydrodynamic diameter, including insulation and strake shell, and His height above this level). For these values a suppression efficiencyof 80% may typically be assumed, in addition to an increased drag coefficient compared to the underlying bare pipe. Strakes near the water surface may need to be treated against marine growth and strakes near the seabed may need to consider abrasion performance. The designer should consider various factors when planning to use strakes, including the following: Required coverage The strake suppliers and some consultants and operators now have performance data from model testing to address the above. The performance of fairings is in some respects better (e.&.lower drag) but can present increased challenge in other areas, e.&.installability. However the field use of fairings as an alternative to strakes does appear to be increasing. Both strakes and fairings can reduce the VIV induced motion, can reduce fatigue damage due to VIV by over 8O%, will, however, introduce handling difficulties. Strakes increase riser drag, whereas fairings reduce drag loading. Fairings need to rotate with current direction and add to design complexity. Strake drag and lift coefficients Alternative PID and HID values Strake and fairing suppression efficiency (including any Reynolds number effects) Performance of strakes (or fairings) in tandem zyxw 9.6 Riser Clashing Riser deflections may need to be controlled to avoid collision with adjacent risers, umbilicals, moorings or the host vessel. Often a target minimum clearance is specified e.g. five times the outside diameter of the riser. If this criterion cannot be met the designer may elect to demonstrate that the probability of collision during field life is of an acceptably low probability (e.g. less than lop4) or demonstrate that collision can be resisted without compromising riser integrity. This logic may also apply to installation operations. 9.6.1 Clearance, Interference and Layout Considerations Analysis should be conducted to confirm that interference with other parts of the production system does not occur. Interaction may occur between the following: Riser and vessel; Riser and riser; Riser and mooring lines; Riser and umbilicals.
- 190. Drilling and Production Risers zyxwvutsrq 837 z The results of a clearance analysis can have an effect on the layout of the risers, umbilicals, mooring and orientation of the flowlines. The layout of the risers should also take into account the overall field layout, the requirement for discrete flowline corridors, anchor system prohibited areas, crane locations, supply boat loading positions and the trajectory of dropped objects. The designer should usually avoid collision among risers. But, if the layout is such that this ideal cannot be achieved, the cumulative probability of risers contacting other risers, umbilicals. mooring legs, the hull or any other obstruction during field life including installation may be assessed and compared to some target value (e.g. A model test of risers in a deep-water fjord was performed to investigate riser collision [Huse, 19961. The test site was chosen at Skarnsund, 100 km north of Trondheim. The sound has a water depth of 190 m, and tidal currents well above two knots. An existing bridge spanning the sound was used as the work platform. A set of riser models were suspended from a surface catamaran with a weight attached to their bottom end, and supported by a pulley system to introduce the desired tension in the risers. The riser group consisted of an array of risers in a 3 zyxw x 4 rectangular arrangement (fig. 9.65) with equal spacing. One riser in the middle of the array represented a drilling riser, while the others were smaller diameter production risers. The array represented a riser system for a Tension Leg Platform. The spacing at the top and bottom end among the risers were maintained at equal distances in the inline and transverse directions. The drilling riser had a pretension of about 1205 kN, while that of the production risers was varied from 412 to 862 zyxwvu kN for two test conditions. Several tests were performed at different current velocities and shear type profiles encountered at the site. At low current velocities, no collision of risers was observed. As the current velocity increased, the collision between neighbouring risers was initiated and the frequency of collision increased with the increase in the magnitude of current velocity. Vortex induced vibration increases the mean inline drag force, causing large static deflection in the middle of the risers. This, in turn, induced collisions between the as well as resistance to consequence damage. R3 R6 R12 zy 0 0 0 0 R5 R8 R11 R1 R4 R7 R10 0 0 0 0 Figure 9.65 Setup of riser array in the Fjord
- 191. 838 zyxwvutsrqp 1.5 1.ozyxwvuts 0.5 0.0 -0.5 -1.0 -1.5 zyxwvutsrqp Chapter z 9 IO 20 30 40 lime zyxw (sec.) zyxw Figure 9.66 Displacement time history of drilling riser [Huse, 19961 neighbouring risers. The collision generated a loud audible noise indicating a collision between the risers. The displacement time history shown in fig. 9.66 shows that the drilling riser experienced a clear evidence of lock-in vibration at its natural frequency. The VIV amplitude was about half the riser diameter. Additionally, the risers experienced an irregular low frequency inline oscillation of large magnitude, almost of a chaotic nature. The peak-to-peak amplitudes of these motions were as much as 3040 diameters. Typically, the far upstream risers remained stationary. The next riser collided with the upstream riser and then moved far downstream in a slow motion before returning upstream and colliding again with the upstream riser. This situation arose at or above the collision velocity of current. In a practical design, of course, it is undesirable to have collisions and they should be avoided in a design. Thus, the low frequency oscillation of the intermediate risers, while of interest, should not arise in a properly designed spacing of a riser system. 9.7 Fatigue Analysis Fatigue damage verification is an important issue in riser design, demanding a high number of loading cases to be analysed. The random time domain nonlinear analysis is considered an attractive and reliable tool for fatigue analysis, as non-linearities are properly modelled and the random behaviour of environmental loading is considered. However, time domain analysis consumes large computer time. The frequency domain analysis is considered an efficient alternative tool for the initial phases of riser design used mainly for a fatigue damage verification. Riser fatigue analysis is conducted using a stress-cycle (S-N) approach. The equation used to determine fatigue life of steel components is: where S= stress range (MPa), including the effects of stress concentration due to misalignment, but excluding that due to the weld itself, zyxw N =the allowable number of cycles for the stress range and K and m=parameters depending on the class of weld, constructional detail.
- 192. Drilling and Pvoduction zyxwvutsr Risers zyxwvutsrqpo Class K ! m Reference x 1.15 zyxwvu x lOI5 4.38 API RP 2A-LRFD (1993) X' zyxwvuts 2.50 1013 3.74 839 z Figure 9.67 Titanium S-N curves Table 9.30 Basic parameters defining fatigue curves for steel in air 15.73 x lo'* 13.00 ~ HSE: 1995 Offshore Guidance Notes 1 1.04x 10l2 13.00 IF2 For titanium alloys such as Grades 23 and 29, the following S-N curve [Baxter, et a1 19971 is widely applicable for good quality girth welds, zyxwv N = 6.8 x 1019.F6 (9.39) The S-N curves for titanium are shown in fig. 9.67. The choice of fatigue design curve will depend on many factors specific to a particular design, construction detail, materials, and the level of conservatism desired. UK HSE Guidance (1995) is given in table 9.30 below for steel in air.
- 193. 840 zyxwvutsrqpo Clzapter z 9 Adjustments may be required to fatigue curves such as those above to account for the endurance-limit effect at low stressjhigh cycles in air, cathodically protected joints in sea-water: and freely corroding joints in sea-water. Other parameters that may affect riser fatigue are thickness, mean stress correction for unwelded or stress-relieved components, stress concentration factors (SCFs), and temperature. Based on published codes and standards it is recommended that for thickness T greater than 25 mm the DNV (2000) correction of (25/7)02 should be applied to the design (or allowable) stress-range obtained from S-N curves. A value of 1.3may be assumed in the absence of more detailed information, although SCFs as low as 1.1 have been achieved for some risers. Fatigue damage in risers comes from three main sources: First-order wave loading and associated vessel motions Second-order/low frequency platform motions Riser VIV due to current or vessel heave (see Section 9.3.1.5 for comments on the latter) Stresses due to 1 and 2 may in some cases be combined prior to calculating fatigue. At the present time it is not considered necessary to combine a riser VIV stresses with these, although that is possible in principle and would be the most accurate approach. Second-order effects are sometimes larger than first-order effects. Also, it is pointed out in Campbell (1999) that introducing second-order effects does not necessarily increase or necessarily decrease fatigue life. An example shows a reduced life (compared to the case where only first-order fatigue is calculated) for a spar-mounted SCR but an increased life for a semisubmersible-mounted SCR. Additional fatigue may accumulate from vessel VIV, slugging, pressure pulses and installation. The fatigue calculation methods use the above stress-cycle (S-N) approach. Fracture mechanics analysis may also be applied. The following methods are possible (among others) for obtaining the combined fatigue effects of 1 and 2: Rainflow Counting (RFC) of stress from a combined (wave-frequency zy +low-frequency) analysis. The most accurate method for any stress time-history, such as output from most riser analyses, requires specialist software and uses more computer time than alternative methods, but is nowadays fairly widely used. Simpler methods may be better for rapid turn around of results; e.g. early feasibility checks or parameter studies. Assume a Rayleigh distribution for the stress peaks from a combined analysis. This overestimates fatigue damage significantly unless stress is highly narrow-banded. There are potential ambiguities in counting the cycles as the response becomes more wide-banded. Use a bimodal method. This still overestimates damage but less so than the Rayleigh assumption, and it is quicker than RFC. A method by Jiao and Moan (1990) is valid when bimodal peaks of the stress spectrum are distinct and well separated. The method
- 194. Drilling and Production Risers zyxwvutsr 84 z 1 z can be used under the right circumstances, but is cumbersome and requires continual checking of the spectrum. Separate wave-frequency and LF stresses. The damage for each frequency region can then be calculated assuming a Rayleigh distribution, and these are summed to get the total damage. This method usually underestimates damage, sometimes significantly. As in 4, but factor the result by (CS,)"/CSy, where SI are individual stress process rms's. Theoretically, this is a somewhat crude correction, but in practice it often works fairly well. However, no attempt is made to correct for the different upcrossing- rates of the different stress processes, which can lead to serious error. A number of investigators have developed correction factors to the Rayleigh approach; e.g. Wirsching and Light (1980), Ortiz and Chen (1987), Lutes and Larsen (1990, 1991). The most accurate and most easily applied of these methods is the single moment method of Lutes and Larsen (note, however, that the spectrum of stress is required, which may require specialist post-processing software, depending on the riser analysis program which has been used). One view on the order of preference is (i) RFC, (ii) Rayleigh or other method with similar or better accuracy, (iii) Lutes-Larsen. In a single moment method of Lutes and Larsen (1990, 1991) the fatigue damage expres- sion given involves one moment of the spectral density function and can be written as follows: (9.40) where Tis duration, and K and zyxwv rn are the parameters of the S-N curve defined by equation (9.40). The single moment in the fatigue damage equation is CT; h2/m= 1 w2/" .G(w) .dw (9.41) where zyxwvut G(o)is the spectral density function of stress-range and w is frequency in rad/s. This method requires no more effort than the Rayleigh method, but the results are generally more accurate, approaching the accuracy of direct RFC for practical purposes. It is recognised that many factors influence the selection of a method, including the domain and format of riser stress data. available software, available time, the relative importance of different terms and the required accuracy at a particular stage in a partic- ular project. However, as a design moves in to final detailed design, there will be a strong expectation that RFC will be used, unless comfortable margins of safety are demonstrated.
- 195. 842 zyxwvutsrqpo Chapter 9 z The use of combined stresses; Le. LF and wave-frequency components calculated in the same dynamic analysis, is preferred, and the level of accuracy should be commented on in all cases. Other methods are possible. For example, regular wave analysis may be sufficient in some cases, especially where fatigue is not a governing criterion; it may also enable more rapid design evolution. Similarly, although time-domain analysis is generally regarded as essential for extreme and confirmatory assessment of riser fatigue, enhanced frequency domain analysis may have a part to play in feasibility studies, parameter studies and fatigue estimation. This is especiallytrue for deepwater risers, where large regions are not subject to grossly non-linear structural response and where accurate random time-domain analysis can be time- consuming. In these cases RFC is not applicable and the Lutes-Larsen method may see greater use. For fatigue analysis it is usually assumed that the riser is installed and operating. Fatigue life is influenced by many factors, and the designer has many techniques at his disposal, for example: zyxwvu 0 zyxwvutsrqp e 0 e 0 0 e 0 0 0 Use of thick-end forging (increased thickness and better S-N curve) Use of project-specific S-N curves, generated by a dedicated test program Refinement of current profiles through further analysis or site measurements Inclusion of wave spreading Increased wall thickness in TDZ Use of auxiliary buoyancy in TDZ Optimisation of hang-off angle Use of lazy-wave rather than free-hanging configuration Review and refinement of inertia coefficient (e.g. if straked pipe is used) Review of structural damping coefficient used in analysis The relative importance of the parameters varied depends on numerous factors, including geographical location and vessel type. 9.7.1 First- and Second-Order Fatigue There are a number of methods available for conducting fatigue analysis of SCRs and the more reliable methods require more computational time and effort. The most important considerations are to include all the relevant sources of fatigue loading and to account correctly for the interaction of first- and second-order contributions. Two example methods for dealing with first- and second-order fatigues are discussed below. The second approach, rainflow counting applied to the combined response, is probably the most accurate. Selectingwhich method to use depends on a number of factors, such as the required level of detail, design stage, type of vessel, and whether or not a wave scatter diagram is available. Other approaches and variations are possible, including cruder but quicker regular wave analysis.
- 196. Drilling and Production Risers zyxwvutsr 843 The earlier discussion on preferred methods for estimating the statistics in specific sea-states provides input to the example methods below. Methodology I : Add Separately Calculated First and Second Order Random Fatigue Damages First-Order Fatigue zyxwvu 0 0 Discretise wave scatter diagram into linearisation windows, as in fig. 9.68. Select sea-state from each window, to give equal or greater damage than for original sea-state. Use selected sea-states in non-linear time-domain analysis, with associated mean offset. Combine tension and bending to obtain total stress. Fourier analysis to get stress RAOs around circumference for each window, as in fig. 9.68. Apply statistics (e.g. Rayleigh distribution) to obtain damage due to each sea-state in window. Multiply damage by probability of occurrence and sum for all sea-states in window. Repeat for each window. Repeat for other loading directions and the sum for total damage. Second-Order Fatigue Discretise scatter diagram into windows or analyse every sea-state, depending on required level of detail. 0 Conduct quasi-static riser analysis using second-order vessel motions. Determine RMS stress response in each case. 0 Apply statistics (e.g. Rayleigh distribution) to obtain damage due to each sea-state. Multiply damage by probability of occurrence and sum damage for all sea-states. Repeat for required number of loading directions and sum for total damage. Combining First- and Second-Order Fatigue 0 Sum the first- and second-order damages at each point on circumference and along the riser length. A variation on this approach, which allows greater flexibility to use the methods already discussed is to calculate the total (first- plus second-order) damage in each sea-state before applying the probabilities. When the preferred approach (RFC) is used in conjunction with this variation, the analysis is essentially the same as the second example methodology, given below. Methodology 2: Apply Rainfow Counting to a Combined First- and Second-Order Random Response If necessary, condense the scatter diagram to manageable number of “bins” (say, 10-20). For each bin, apply mean offset and conduct non-linear time-domain analysis with vessel second-order motions included.
- 197. 844 zyxwvutsrqpon Chapter 9 z rirt zyx A9.126 I zyxwv Figure 9.68 Example windowing and sea-state selection of long-term scatter diagram
- 198. Driiliiig zyxwvutsrqpon and Production Risers zyxwvutsr 845 z 60 zyxwvuts 40 zyxwvuts Total Stress, 20 0 -20 -40 MPa -601 0 100 200 300 400 500 600 700 800 Time, sec zyxwv Figure 9.69 Combination of HF and LF narrow-banded Gaussian processes Combine tension and bending to obtain total stress (fig. 9.69). Rainflow count total stress time traces to get fatigue damage due to each bin at points around the circumference and along the riser length. Multiply damage by probability of occurrence of bin and sum over bins. Repeat for required number of loading directions and sum for total damage. As for all random analyses, convergence of statistics needs to be understood and checked. In this method, use of a minimum of ten low-frequency cycles to achieve meaningful results is one rule of thumb, though this should not be taken as a substitute for checking. 9.7.2 Fatigue Due to Riser VIV To estimate long-term riser VIV fatigue damage: Establish current data. Normally, at least ten profiles are required, and directional variations should be included. If available, concurrent data; Le. actual profiles, are preferable to exceedance profiles. Conduct VIV analysis using a suitable VIV analysis tool. The nominal (or neutral) riser configuration may be used, but this is not essential. Factor calculated fatigue damage in each current according to the probability of current and sum of all such damages to obtain total damage and hence predict fatigue life. Sensitivity analyses may be conducted in which currents and riser configuration are varied. Justification and a methodology for spreading (or smearing) fatigue damage in the TDZ, based on the fact that the TDP and riser system properties will vary over time, is given in Section 9.2.5.2.2. VIV fatigue in risers is commonly assessed using dedicated software such as SHEAR7 or VIVA. It should be noted that there are other prediction tools available, such as
- 199. 846 zyxwvutsrqponm Chapter z 9 z VIVANA and Orcaflex. The tools chosen for discussion in this section should not be taken as any form of recommendation, rather as typical examples. Most VIV programs allow input of only 2D current, although advances are expected in this area. As a general rule, for a SCR, resolution of velocity on to planes parallel with and perpendicular to the plane is acceptable. It is assumed that an initial VIV fatigue analysis is performed (e.g. a modal analysis using SHEAR7) where the vessel is in the neutral position. Apart from the VIV, no dynamic forces or motions are accounted for in this initial analysis. Under these assumptions it is found that the predicted fatigue damage in the TDZ peaks sharply at anti-nodes of the calculated mode-shapes, where curvature and bending stress peak. This results in large fluctuations in overall predicted fatigue life between anti-nodes, the extent of this effect depending on which modes, and how many, are mobilised. In reality, riser system properties and boundary conditions will vary continuously. The TDP will move under the influences of vessel motion and direct hydrodynamic loading on the riser, and the riser mass will change for various reasons over various time-scales. This means that mode-shapes will also be continuously changing, and the locations on the riser of the modal anti-nodes may move around significantly. The effect of this will be to tend to even out peaks and troughs in the calculated damage curve. If this region governs the fatigue, then the true life of the riser will be greater than that predicted by the “constant riser system” assumed in the initial VIV analysis. Reasons for variation of riser (e.g. SCR) system properties and boundary conditions are numerous, and include both short-term and long-term effects over the design life; e.g. 1. 2. 3. 4. 5. 6. Wind, second-order wave loads and varying current introduce low-frequency vessel motion and affect mean vessel offset, causing the location of the TDP to change. Variation of current force applied directly to the riser will also move the TDP. Vessel draught and tidal variations will move the TDP. Depending on the field development plan, vessel drilling offsets may be applied over a substantial period, and risers phased in at a later stage may impose incremental offsets. Density of riser contents may vary. Short-term density variations in production risers may arise from variable well fluids and conditions. There may also be long-term variations as a reservoir becomes depleted and the composition of both produced and exported fluids changes. Even if these variations are small, they may be sufficient to shift natural modes and frequencies enough to have an important effect on fatigue peak locations. Riser mass may depend on several long-term effects; e.g. corrosion and water absorption in auxiliary buoyancy. It is emphasised that this list is not exhaustive and that not all of these effects need to be considered in every analysis.
- 200. Drilling and Production Risers zyxwvuts 847 Fatigue Damage P TPD zyx Ls zyxw Figure 9.70 Approximate fatigue calculation zyxw Effects on SCR TDP boundary conditions and response may also arise from trenching, suction and other soil-related phenomena; and the way the TDP is modelled in the VIV analysis can be crucial. However, whilst important, these are considered to be the aspects of detailed structural modelling which should be addressed elsewhere. Comprehensive statistical treatment of all influences on the fatigue damage distribution is possible but will normally be unnecessary. The excess conservatism of an assumed constant riser system should be avoided, however, although it is possible to make reasonable allowances without performing an unduly complex analysis. The preferred approach depends on specific risers, field development plans, available software and individual company practice. In some cases it may be considered necessary to perform a separate VIV fatigue analysis for numerous variations from the neutral configuration, to cover all scenarios in 1-6, above. In general, however, the depth of analysis required to get the right balance of accuracy, conservatism and economy will vary. One simple approximation, which may either be useful as a preliminary check or give sufficient confidence in itself, is: (a) Determine a characteristic movement along the riser of the anti-node nearest to the TDP, allowing for all effects, such as those stated above. This is denoted Ls. (b) For each point P in the region of the TDP, take the fatigue damage as being that calculated from initial VIV analysis, averaged over a distance Ls, centred on P. This may be described as a “moving average” calculation. It applies to all points around the circumference, although the averaging is performed in the lengthwise sense, only. The essentials of this calculation are illustrated in fig. 9.70. Ls can be determined from, zyxwvu Lz, = zyxwv cL;,j (9.42)
- 201. 848 zyxwvutsrqpon Chupter zy 9 z where Ls,i is a characteristic movement of the anti-node due to the ith effect acting in isolation, and it is assumed that all effects are uncorrelated. There is some freedom in the choice of the Ls,i,each of which is some representative value of a random variable. But it is suggested that a value of two standard deviations of the amplitude of movement will ensure that benefits are realised, whilst a degree of conservatism is maintained. Correlation between the various effects and use of more realistic distributions can be incorporated into the analysis, if the information required to do this is available. However, this may complicate the analysis considerably without yielding a great improvement in results. One relatively simple adjustment which could be reasonable in some cases is to assume Gaussian behaviour and weight the initial fatigue damage distribution accordingly (Le. instead of using the uniform distribution implied by step (b) above) but this approach is not assumed here. It is possible that only a single value of Ls will be required, applicable across all initial VIV fatigue analyses. However, if currents from different directions make signi- ficant contributions to fatigue damage, it may be necessary to use more than one value of Ls - each in conjunction with results for the corresponding current direction and associated probability. Also, a situation may arise in which the initial VIV analysis is not performed for a single neutral position but for, say, two configurations, near and far. It is not possible to anticipate all such scenarios, and judgement and adjustment must be made on a case-by-case basis. Ultimately it is the responsibility of the contractor to identify key influences and account for them appropriately. In any event, it is recommended that sensitivity checks be performed to determine how much the anti-nodes of typically excited modes move under the influence of effects like 1-6 above. In addition to first- and second-order fatigues and riser VIV, other possible sources of fatigue damage are vessel VIV, vessel springing, and internal fluid effects, such as slugging and pressure surges. For example, vessels with cylindrical sections subjected to current loading may oscillate due to vortex shedding; e.g. spars (usually straked to reduce this effect) and other deep draft floaters. Fatigue also depends on riser/seabed interaction. Trenching, suction and seabed consolidation will also have an effect on fatigue. This topic has been the subject of several recent industry research initiatives. zyxwv 9.7.3 Fatigue Acceptance Criteria It is necessary to determine overall fatigue resistance, accounting for each relevant effect, which may include: Riser VIV Vessel VIV Issues to be addressed when combining fatigue damage are correlation, stress amplifica- tion, and interaction. Correlation refers to the fact that (for example) riser fatigue is due to First- and second-order loads and motions Other effects such as slugging, pressure surges
- 202. Drilling and Production Risers zyxwvutsr 849 z wind and wave effects may not be related to current induced fatigue, such as riser VIV. Fatigue due to slugging may occur at any time. Stress amplification refers to the effect of two or more loading regimes occurring in combination, for example, first-order wave loading and riser VIV. The resulting fatigue damage is greater than that calculated from treating the two separately and adding the damages. This effect is most significant when damage rates are of a similar order of magnitude. Interaction between loading mechanisms may reduce the effect of stress amplification; e.g. wave-induced riser response may disrupt riser VIV. With due consideration to these and other uncertainties inherent in riser fatigue prediction, the designer should select a safety factor to apply to fatigue life predictions. The choice of safety factor will depend on many factors. Typical ranges applied are from 3, for non- critical applications where in-service inspection is planned, to 20 to applications with increased uncertainty (e.g. VIV) where inspection is not possible. The choice of safety factor(s) should be made in conjunction with the end-user. The fatigue damage components predicted from all effects are accumulated to arrive at the total damage at each location on the riser, which must satisfy: 1/ SiDi > Design Life (9.43) where Si=safety factor and zyxwv Di =annual fatigue damage for the ith effect. The sum should include damage arising from all effects; e.g. first- and second-order, various types of VIV, installation and pressure surges. In calculating the D,, allowance should be made for the duration of each effect throughout the year. zyxwv 9.8 Fracture Mechanics Assessment Fracture mechanics (FM) analyses may be used to develop flaw acceptance criteria. The FM analysis is very useful not only in controlling fatigue limiting cracks, but also provides guidance for selecting appropriate weld inspection techniques, as well as reducing the number of welds needing to be cut-out and replaced. The fracture mechanics analysis usually consists of three steps, which are discussed below: 1. zyxwvutsr 2. Paris Law fatigue analysis 3. Acceptance criteria development Engineering Critical Assessment (ECA) of the riser body Development of stress histograms for input to FM analysis depends on data available from riser dynamic analysis, and may use a recognised cycle counting scheme [e.g. as in ASTM E1049-85 (1997)l or conservative distribution (e.g. Rayleigh curve, based on combined L F and HF dynamic analysis). This is analogous to determination of stress distributions for use in S-N fatigue analysis. For VIV fatigue a Rayleigh stress-range distribution is often considered suitable regardless of the number of modes responding.
- 203. 850 zyxwvutsrqpo Figure 9.71 Flaw characterisation zyxw Chapter 5 9.8.1 Engineering Critical Assessment zyxwvu The ECA is typically performed using industry accepted practices such as EPRI, CTOD method, or more rigorous analyses such as the R-6 method. SCRs to date have typically been assessed using PD-6493 (1991) or BS7910 (1999) methods. These methods allow for material behaviour ranging from brittle fracture to plastic collapse of the cross section. However, most modern materials with good ductility are often best characterised by nonlinear fracture mechanics, which is well treated using the Failure Assessment Diagram (FAD) approach. The analyses should result in an envelope of limiting crack sizes which cause failure under the expected extreme event (e.g. 100-yrreturn period hurricane) for a particular system (e.g. TLP, SPAR, etc.) and worst loading condition. Material and weld specific CTOD, measured at -10°C or lower, should be used, if available. Codified default values may be assumed. However, the designer should realise that these values might be far from representative depending on the weld process and inspection techniques employed. Material yield and tensile strength should be measured for the parent and weld metal, as well as, for the heat-affected zone. Conservative values should be used properly to account for the weld/parent metal mismatch. The BS7910:1999 Level 2 FAD is appropriate for an initial riser ECA. If material specific ductile tearing data is available, then the Level 3 approach zy (JR)may be used. Care should be taken with the Level 3 approach since very large limiting flaws may result. Cracks are usually assumed to be elliptical for analysis purposes. Surface breaking, buried, and interacting flaws should be considered. An idealisation of the elliptical surface and buried flaws is shown in fig. 9.71. Note that in some cases, the uncracked ligament of a buried flaw may be so close that it is re-characterised as a surface flaw. Refer to PD 6493 (1991) for guidance on values for zyxwv “x” in fig. 9.71. Stress intensity factors must be chosen so that the analytical solution accurately mimics the cracked pipe. In many cases, flat plate solutions provide sufficiently accurate results. However, for cases where the crack length and depth are not small with respect to the pipe circumference and wall thickness, the far-field uniform stress plate solutions may be
- 204. DriNing and Production Risers zyxwvuts 851 zy inaccurate. Moreover, thin shells with outer to inner radii greater than 0.8 need curvature correction factors [refer to PD6493 (1991) for guidance]. zyxw 9.8.2 Paris Law Fatigue Analysis The so-called Paris Law for fatigue is described using daldN zyxwvut = zyxwvu A(AK)” (9.44) where da:dN = crack growth rate of crack of depth a vs. the number of applied stress cycles N, AK = stress intensity factor range, while A and m are material specific constants. BS7910 (1999) provides recommended values for the Paris Law, which should be suitable for the fatigue analysis. Material specific data obtained from tests are relatively inexpensive and may be used in-lieu of codified data. If idealised stress intensity factor solutions are utilised (e.g. smooth plate solutions) in lieu of the finite element fracture mechanics analysis of the actual geometry, then relevant stress concentration factors should be applied to the stress range bins to account for increased applied stress due to local weld geometry, pipe mismatch, etc. 9.8.3 Acceptance Criteria The industry has typically followed an approach similar to the schematic in fig. 9.72. The approach has been to develop curves showing an envelope of elliptical cracks (edge or embedded), which may grow to the limiting flaw size (see above) in a specified time. The “specified time” is usually established as a safety factor multiplied by the design life. Deciding the safety factor is subjective, but must take into account the type of inspection used during weld fabrication. Additionally, the safety factor should reflect uncertainties in predicted loads (see, also, Section 9.7 on safety factors). 9.8.4 Other Factors To Consider Some of the other factors are listed below: Internal Contents: crack growth may be accelerated in H2S or other corrosive conditions Cathodic Protection: crack growth is dependent on the level of corrosion potential protection Hydrogen embrittlement from welding Plastic straining (for reeled risers) Internal pressure effects on crack growth 9.9 Reliability-Based Design Reliability-based design is becoming more common in pipeline engineering and other areas of the offshore industry. Its application to risers is limited at this time. Particularly,
- 205. 852 zyxwvutsrqpon Chapter 9 z Establish FM assessment parameters Calulate envelope zyxwv n,... -:-- is flaw which fails by oastic collaose durina the of limiting 4 a combination of fracture and Factor results zyx for acceptance criteria zyx Figure 9.72 Schematic of example acceptance criteria development procedure
- 206. Drilling and Production Risers zyxwvutsr 853 the deepwater SCR design for floating vessels is relatively a new technology. Hence it may be some time before the sufficient statistical data is available on SCR. However, procedures to determine component and system reliabilities have been investigated as part of the Integrated Mooring and Riser Design JIP, and are described in a Technical Bulletin (1990). zyxwvut A major step forward is also provided by DNV’s “Dynamic Risers” which provides an LRFD format with reliability-based calibration of partial safety factors. Development of long-term response distribution and comprehensive reliability assessment is possible but far from being standard analysis for risers. Nevertheless, limited methods and examples have been demonstrated for flexible risers [Farnes and Moan, 1993; Larsen and Olufsen, 1992;Trim, 19921and more recently for an SCR hung off a ship-shaped vessel in the GOM [Gopalkrishnan, et a1 19981. The key advantage of the reliabilityjlong-termmethods is their consistency; i.e. the fact that exceedance probabilities are used which account for all environmental conditions arising in the long-term. This is exemplified in Corr, et a1 (2000), which reports 100-yr responses 20% lower than those obtained using conventional combination of collinear 100-yr wind, wave and current conditions. In this method the joint statistics of environmental inputs were developed and combined with results of representative dynamic simulations to produce a response-surface (a response which is a function of several environmental variables). It should be cautioned that use of such methods cannot be assumed to always result in reduced response predictions, as that depends on the “conventional” methods to which they are compared. Nevertheless, their consistency and resulting high levels of confidence point the way to safer and more economic design. 9.10 Design Verification The purpose of design verification is to provide the designer with an independent review and confirmation that the design adequately addresses the key issues outlined below: Functional and operational requirements in the client’s specifications and documentation. Structural integrity. Stable overall configuration; no detrimental interference with adjacent risers, Resistance to fatigue and other forms of long-term degradation. umbilicals, moorings. Compatibility with fluids being transported The design verification process should also include riser appurtenances. In cases where the installation process results in significant effects on the riser; e.g. reeling, it will also be appropriate to include the installation operations, limits and contingency procedures in the scope of the review. The process is one of the confirmation for the client/designer and is
- 207. 8 zyxwvutsrqp 54 zyxwvutsrqponm Chapter zy 9 not intended to replace the more formal independent design review required by certifying authorities. It is appropriate when addressing the state of the art technology applied to critical equip- ment to consider two levels of design verification: (i) a systematic review of key documentation zyxwvut - specifications, design bases, design reports - to confirm the adequacy of the design process and documentation; (ii) an independent analysis of selected key load cases, preferably by a consultant with access to different analytical software to that used by the designer. Sources of uncertainty, as far as the current SCR design is concerned, include compression in the TDZ, riser/soil interaction and riser fatigue due to VIV. Model testing to confirm key design issues and assumptions may be considered as part of Design Verification, particularly where assumptions relate to safety-critical features of the design. It should be realised that the modelling process has fundamental shortcomings when used to address the behaviour of an integrated riser/host/mooring system in that the model scaling requirements for different parts of the system cannot be satisfied in a single model. zy 9.11 Design Codes The main design codes and standards used for riser design are: “Dynamic Risers”, DNV-OS-F201, 2001. “Recommended Practice for Design of Risers for Floating Production Systems and TLPs”, First Edition, API RP 2RD, June 1998. “Submarine Pipeline Systems”, DNV-OS-F101, 2000. “Recommended Practice for Design, Construction, Operation, and Maintenance of Offshore Hydrocarbon Pipelines (Limit State Design)”, API RP 1111, 3‘d Edition, July 1999. “Guidance on Methods for Assessing the Acceptability of Flaws in Fusion Welded Structures”, BS PD 6493, August 1991. “Fatigue Strength Analysis of Mobile Offshore Units”, DNV Classification Note No.30.2, August 1984. “Offshore Installations: Guidance on Design, Construction and Certification”, HSE Books, 1995 (supersedes same title from UK Dept. of Energy, HMSO, 1984 and takes precedence over “Code of Practice for Fatigue Design and Assessment of Steel Structures”, BS7608:1993, with or without amendment of February 1995). “Recommended Practice RP B401: Cathodic Protection Design: 1993”, DNV. Regarding the Fatigue Design Codes, the reader is referred to an industry design codes, which provide guidance on the appropriate selection of S-N curves to apply to girth welds and other components under cyclic fatigue loading. Factors, which the designer may need to consider are: Project-specific conditions (materials, production chemistry, welding procedures) etc. which may cause the riser fatigue performance to depart from published curves
- 208. DriNing and Production Risers 855 Compressive stress cycles Environmental conditions zyxwvu - in air, in seawater, in seawater with cathodic protection etc. Presence of mean stress for non-welded components Ovality and mis-match causing hi-lo conditions at the weld zyx Acknowledgement We acknowledge that Dr. A. D. Trim edited part of the material included in the section of Steel Catenary Riser of this chapter. References API RP 1111 (1999). “Design, construction, operation, and maintenance of offshore hydrocarbon pipelines (Limit State Design)”, (3rded.). July. API RP 164 (1993). “Recommended practice for design, selection, operation, operation and maintenance of marine drilling riser systems”, American Petroleum Inst., Wash, D.C. API RP 2A-LRFD (1993). “Recommended practice for planning, designing and constructing fixed offshore platforms - load and resistance factor design”, (lst ed.). July. API RP 2RD (1998). “Recommended practice for design of risers for floating production systems (FPSs) and Tension Leg Platforms (TLPs)”, (1” ed.). June. API (1994). “Bulletin on formulas and calculations for casing, rubing, drill pipe, and line pipe properties”, API Bulletin 5C3, (6th ed.). American Petroleum Inst., Wash, D.C., October. ASME 31.4. American Society Of Mechanical Engineers: Gas Transmission and Distribution Piping Systems. ASME 31.8. American Society Of Mechanical Engineers: Liquid Transportation Systems for Hydrocarbons, Liquid Petroleum Gas: Anhydrous Ammonia, and Alcohols. Baxter, C.: Pillai, S., and Hutt, G. (1997). “Advances in titanium risers for FPSOs”, OTC 8409. Bell, M. and Daly, R. (2001). “Reeled PIP SCR”, OTC 13184. May. Blevins, R. D. (1977). Flolk*-Znduced Vibrations, Van Nostrand Reinhold Company, New York, N.Y. Brekke, J. N., Chou, B., Virk, G. S., and Thompson, H. M. (1999). “Behavior of a drilling riser hung off in deep water”, Deep Offshore Technology Conference. Brooks, I. H. (1987). “A pragmatic approach to vortex-induced vibrations of a drilling riser”, Proceedings of the Offshore Technology Conference, OTC 5522, Houston, TX.
- 209. 856 zyxwvutsrqpo Chapter zy 9 BS7910:1999 (later BS version of above PD 6493). BS8010 (1973). “Code of practice for pipelines: Pt.3: Pipelines subsea: design, construction and installation”. BS7608 (1995). “Code of practice for fatigue design and assessment of steel structures”, 1993, with amendment, February. Buitrago, J. and Zettlemoyer, N. (1998). “Fatigue design of critical girth welds for deepwater applications”, OMAE 98-2004, Lisbon, Portugal, July. Campbell, M. zyxwvut (1999). “Complexities of fatigue analysis of deepwater risers”, Deepwater Pipeline Technology Conference, March. Chaudhury, G. and Kennefick, J. (1999). “Design, testing and installation of steel catenary risers”, Proceedings of the Offshore Technology Conference, OTC 10980, Houston, Texas, May 3-6. Clausen, T. and D’Souza, R. (2001). “Dynamic risers key component for deepwater drilling, floating production”, Offshore Magazine, pp. 89-93, May. Connelly, L. M. and Zettlemoyer, N. (1993). “Stress concentration at girth welds of tubulars with axial wall misalignment”, Proc. 5‘h Intl. Symp. Tubular Structures, Nottingham, UK, August. DNV Classification Note 30.2, (1984). “Fatigue strength snalysis of mobile offshore units”, August. DNV Classification Notes 30.5, Environmental Conditions and Environmental Loads. DNV (1981). “Rules for submarine pipeline systems”, H+vik. Norway. DNV (1996). “Rules for submarine pipeline systems”, December. DNV-OS-F101 (2000). “Submarine pipeline systems”. DNV-OS-F201 (2001). “Dynamic Risers”. Erb, P. R., Ma, Tien-Chi, and Stockinger, M. P. (1983). “Riser collapse - A unique problem in deep-water drilling”, IADCjSPE 11394, Society of Petroleum Engineers. Farnes, K. zyxwvut A. and Moan, T. (1993). “Extreme response of a flexible riser system using a complete long-term approach”, Proc. ISOPE. Finn, L. (1999). “Reliable riser systems for spars” Journal of Offshore Mechanics and Arctic Engineering, Vol. 121, pp. 201-206, November. Fumes, G. K., Hassanein, T., Halse, K. H., and Eriksen, M. (1998). “A field study of flow induced vibrations on a deepwater drilling riser”, Proceedings of the Offshore Technology Conference, OTC 8702, Houston, TX, pp. 199-208. Gardner, T. N. and Cole, M. W. (1982). “Deepwater drilling in high current environment”, Proceedings of the Offshore Technology Conference, OTC 4316, Houston, TX.
- 210. Drilling and Production zyxwvutsrq Risers zyxwvuts 857 z Garrett, D. L., Gu, G. Z., and Watters, A. J. (1995). “Frequency content selection for dynamic analysis of marine systems”, Proceedings OMAE, Copenhagen, Vol. 1-B, Gopalkrishnan, R., Kopp, F., Rao, V. zyxwv S., Swanson, R. C., Yu, X., Zhang, J. Q., Jones, W. T., and Zhang, H. (1998). “Development of the dynamic riser system for a ship-shaped production host in the deepwater Gulf of Mexico”, IOrh Deep Offshore Technology Conference, New Orleans, November. HSE Books (1995). “Offshore installations: guidance on design, construction and certification”, (supersedes same title from UK Dept. of Energy, HMSO, 1984). Integrated Mooring and Riser Design JIP (1999). “Technical bulletin: analysis methodology”, MCS International and Noble Denton Europe, September. IP6: Institute of Petroleum Pipeline Safety Code (Uk). Jiao, G. and Moan, T. (1990). “Probabilistic analysis of fatigue due to Gaussian load processes”, Probabilistic Engineering Mechanics, Vol. 5, No. 2. Johns, T. G., Meshlokh, R. E., and Sorenson, J. E. (1976). “Propagating buckle arrestors for offshore pipelines”, OTC 2680. Kavanagh, K. Dib, M., and Balch, E. (2004). “New revision of drilling riser recommended practice” (API RP 16Q), OTC 14263, May. Krolikowski, L. P. and Gay, T. A. (1980). “An improved linearization technique for frequency domain riser analysis”, Proceedings of the Offshore Technology Conference, OTC 3777, Houston, TX. Langner, C. G. (1975). “Arrest of propagating collapse failures in offshore pipelines”, Shell Deepwater Pipeline Feasibility Study. Larsen, C. E. and Lutes, L. D. (1991). “Predicting the fatigue life of offshore structures by the single-moment spectral method”, Probabilistic Engineering Mechanics, Vol. 6, No. 2, pp. 393-399. pp. 96-108. Larsen, C. M. and Olufsen, A. (1992). “Extreme response estimation of flexible risers by use of long term statistics”, Proc. ZSOPE. Lee, L., Allen D. W., Henning, D. L., and McMullen, D. (2004). “Damping character- istics of fairings for suppressing vortex-induced vibrations”, OMAE 2004-51209. Lutes, L. D. and Larsen, C. E. (1990). “Improved spectral method for variable amplitude fatigue prediction”, J. Struct. Engng., Vol. 116, No. 4, pp. 1149-1164, April. McIver, D. B. and Olson, R. J. (1981). “Riser effective tension zyx - now you see it, now you don’t!’’. zyxwvuts 371h Petroleum Mechanical Engineering Workshop and Conference, ASME, Dallas, TX, September. Mekha, B. B. (2001). “New frontiers in the design of steel catenary risers for floating production systems”, Journal of Offshore Mechanics and Arctic Engineering, Vol. 123, pp. 153-158, November.
- 211. 858 zyxwvutsrqpo Chapter 9 Miller, J. E. and Young, R. D. (1985). “Influence of mud column dynamics on top tension of suspended deepwater drilling risers”, Proceedings of the Offshore Technology Conference, OTC 5015, Houston, TX. Moe, G.,Teigen, T., Simantiras, P., and Willis, N. (2004). “Predictions and model tests of a SCR undergoing VIM in flow at oblique angles”, Proceedings of Offshore Mechanics and Arctic Engineering, OMAE2004-51563, Vancouver, Canada. Mork, K. J., Chen, M. Z., Spolton, S., and Baxter, C. (2001). “Collapse and buckling design aspects of titanium alloy pipes”, zyxwv 2OzhInt. Con$ OMAE, Rio de Janeiro, June. Murphy, C. E. and Langner, C. G. (1985). “Ultimate pipe strength under bending, collapse and fatigue”. NTNE Research Programme, “Handbook of hydrodynamic coefficients of flexible risers”, FPS 2000/Flexible Risers and Pipes,‘Report 2.1-16. Offshore magazine, “Drilling zyxwvu & production riser systems and components for floating units”, Poster distributed with May 2001. Ortiz, K. and Chen, N. K. (1987). “Fatigue damage prediction for stationary wideband stresses”, Jrh Int. Conf. on Application of Statistics and Probability in Soil and Structural Engng. Rothbart, Harold A. (Ed), (1964). Mechanical Design and Systems Handbook, McGraw- Hill Book Company. Siewert, T. A., Manahan, M. P., McCowan, C. N., Holt, J. M., Marsh, F. J., and Ruth, E. A. (1999). “The history and importance of impact testing”, Pendulum Impact Testing: A Century of Progress, ASTEM STP 1380,T. zyxw A. Siewert and M. P. Manahan, Sr., (Editors), American Society for Testing and Materials, West Conshohocken, PA. SHEAR7 User Manual for use with Version 4.1. Sparks, C.P. (1983). “The influence of tension, pressure and weight on pipe and riser deformations and stresses”, Proc. 2“dInt. OMAE Symposium, Houston. Spolton, S. and Trim, A. D. (2000). “Aspects of steel-titanium riser design”, JrhZnt. Con$ On Advances in Riser Technologies, Aberdeen, June. Stahl, M. J. (2000). “Controlling recoil in drilling risers following emergency disconnect”, Proceedings of ETCEIOMAE2000 Joint Conference, ETCE2000/DRILL-10105, New Orleans, LA. Timoshenko and Gere (1961). “Theory of elastic stability”, (2nded.). McGraw-Hill. Triantafyllou, M., Triantafyllou, G., Tien, D., and Ambrose, B. D. (1999). “Pragmatic riser VIV analysis”, Proceedings of the Offshore Technology Conference, OTC 10931, Houston, TX. Triantafyllou, M. S. (2001). “User guide for VIVA”, February. Trim, A. D. (1992). “Extreme responses of flexible risers”, Marine Structures: Special Issue on Flexible Risers (Part IIj, Vol. 5, No. 5.
- 212. Drilling und Pvoducrion Risers zyxwvutsr 859 Vandiver, J. K. (1998). “Research challenges in the vortex-induced vibration prediction of marine risers”, Proceedings of the Offshore Technologj Conference, OTC 8698, Houston, TX, pp. 155-159. Venkataraman, G. (2001). “Reeled risers: deepwater and dynamic considerations”, OTC 13016, May. Wirsching, P. and Light, M. C. (1980). “Fatigue under wide band random stresses”, Journal o f the Structural Division. ASCE, Vol. 106, No. ST7, pp. 1593-1607, July. See http::, www.matweb.com/,web site for material properties.
- 213. Handbook of Offshore Engineering zyxwvutsr S . Chakrabarti (Ed.) zyxwvutsrq C 2005 Elsevier Ltd. zyxwvutsrq All rights reserved zyxwvuts 861 Chapter 10 Topside Facilities Layout Development Kenneth E. Arnold and Demir I. Karsan Paragon Engineering Services Inc., Houston, TX, USA Subrata Chakrabarti Offshore Structure Analysis, Inc., Plainfield, IL, USA 10.1 Introduction The most important factors governing an offshore platform topside facilities layout and design are its purpose and whether it will be manned or unmanned. A manned facility will require accommodation quarters for the personnel and will be subject to additional safety requirements. A manned facility will also require special transportation, landing and evacuation facilities for the personnel. These requirements will necessitate additional deck space. Based on the equipment and personnel requirements for the topside facilities, first a deck layout plan should be developed. The layout plan is based on the operational workability and maintainability of the equipment and the health and safety requirements for the personnel who will operate it. The layout plan may be accommodated in a single deck level or may require multiple deck levels depending on the type of the offshore structure. For example, a Floating Production Storage and Offtake System (FPSO), which may be supported by a new built or converted ship shaped vessel, would normally have ample space available on its deck to accommodate most equipment and personnel on a single deck level. On the other hand, a fixed jacket, SPAR, or TLP topsides would have a smaller footprint and the production equipment may be laid in multiple levels. This chapter describes the general considerations for the layout and design of the topside facilities for offshore platforms. The effect of the environment on the deck design; the types of topside deck structures and the split of the construction, hookup and commissioning (HUC) activities between the onshore and offshore sites depending on the deck type; and the control and safety requirements, including fuel and ignition sources, firewall and fire equipment are presented. The practical limitations of the topside design are described. As examples, two different layout systems are compared and the topside design of the North
- 214. 862 zyxwvutsrqpo Chapter z 10 Sea Britannia platform is presented. Much of the material presented in this chapter has been derived from course notes prepared by Mr. Ken Arnold, CEO of the Paragon Engineering Services, Inc. zyxwvu 10.2 General Layout Considerations The following items require special attention in the topsides facilities design: zy e e e zyxwvutsrq 0 e e e e e 0 e e A Prevailing Wind Direction Firewalls and Barrier Walls Process Flow Safe Work Areas Storage Ventilation Escape Routes Fire Fighting Equipment Thermal Radiation Vapour Dispersion Future Expansion(s) Simultaneous Operations Provisions (such as producing while drilling or working over wells) number of offshore platform topside deck layouts have evolved in response to operational requirements and the fabrication infrastructure and installation equipment availability. Operational requirements dictate the general deck size and configuration (number of deck levels and their layout, etc.). For example, the need for a fully integrated drilling and production system would dictate vertical and horizontal layering of the deck structure in such a manner as to provide an efficient operation while also providing an acceptable level of human and environmental safety. If fabrication facilities and skilled labour are not available in the area; the economics may dictate building the deck in smaller pieces and modules and assembling them offshore using the low capacity offshore lifting equipment available. This approach may result in increased steel weight, and offshore construction time and cost, while avoiding the expense of investing in a major fabrication yard. Alternatively, the owner may design the deck as an integrated single piece structure or as a Module Support Frame (MSF) supporting a few large modules, which can be built at a location where fabrication infrastructure and equipment are readily available. The “integrated deck” may then be installed on site using high capacity lifting cranes, or if not available, a float-over deck installation approach. In a float-over deck installation approach, the fully integrated and pre-commissioned deck (or a large module) is loaded out onto a large transportation vessel(s) and transported to the installation site as a single piece. At the installation site, the deck is floated over and then lowered onto the support structure by either ballasting the vessel or using quick drop mechanisms. Alternatively, for the case of a floating support structure, the support structure may be de-ballasted to pick the deck up.
- 215. Topside Facilities Layout Development 863 z An integrated deck may be divided into a number of levels and areas depending on the functions they support. Typical levels are: Main (upper) deck, which supports the drillingiproduction systems and several modules (drilling, process, utilities, living quarters, compression, etc.) Cellar deck, which supports systems that need to be placed at a lower elevation and installed with the deck structures, such as pumps, some utilities, pig launchers/ receivers, Christmas trees, wellhead manifolds, piping, etc. Additional deck levels, if needed. For example, if simultaneous drilling and produc- tion operations are planned, some process equipment may be located in a mezzanine deck. An example of such a topside layout is the Diana SPAR design [Milburn and Williams, 20011. In the Diana Spar topsides design, the upper deck is called the “Drilling Deck”. and has the production and temporary quarters buildings, drill rig, chemical tote tank storage and communications and radar satellite dishes. The mid-level (mezzanine) deck is called the “Production Deck” and contains the majority of oil and gas separation, processing, treating, compression equipment, power generation equipment, the MCC/Control Room and many of the utilities. The lower “Cellar Deck” contains other utility systems (cooling water, fresh water, firewater, flare scrubber, etc.) as well as oil and gas sales meters, pipeline pig launchers and receivers, manifolds and shutdown valves. A subcellar Deck, which is a partial deck suspended below the cellar deck could also be installed to contain the gravity drain sumps and pumps. Because this deck is usually small it could be designed to withstand impact from the wave crest and transport the lateral loads to the rest of the structure. A modular deck may be divided into a number of pieces and modules depending on the functions they support and the installation equipment available. Typical modular deck components are: zyxwvut - Module Support Frame (MSF), which provides a space frame for supporting the modules and transferring their load to the jacket/tower structure. The MSF may also be designed to include a number of platform facilities, such as the storage tanks, pig launching and receiving systems, metering/proving devices and the associated piping systems, Modules. These provide a number of production and life support systems, such as the - Living Quarters Module (generally supporting a heliport, communication systems, hotel, messing, office and recreational facilities). Utilities Module (generally supporting power generation and electrical and produc- tion control systems, including a control room). Wellhead Module (generally supporting the wellheads, well test and control equipment). Drill Rig Module (containing the drill tower, draw-works, drillers and control rooms, drill pipe and casing storage racks and pipe handling systems). Drill rig module is located over and supported by the wellhead module. -
- 216. 864 zyxwvutsrqpon Chapter z 10 Production Module zyxwvu (containing the oil/gas/water separation and treatment systems, other piping, control systems and valves for safe production, metering and transfer of the produced liquids and gas to the offloading system). CompressionModule, if gas compression for injection to the formation and/or high- pressure gas pumping to the shore is needed. Since compression may be needed at later production stages, this module may be installed on the deck at a later date or on a nearby separate platform (generally bridge connected to the deck). Similarly, water injection and pumping modules may be added if these functions are needed at later field development stages. In general, integrated decks result in more efficient and lighter structural systems than modular decks, since additional module steel, which is only needed for installation reasons, is avoided. For demonstration purposes, the following paragraphs will elaborate only on the components and design of a mid-sized single deck structure. The design of MSF and modules follow similar design principles and methods. zyxw 10.2.1 General Requirements It is advantageous to design the topsides facilities using a Three (3)Dimensional Computer Aided Design (3D CAD) model. If the 3D model has a high degree of accuracy on all structural, piping and equipment layouts; the possibility of encountering “clashes” in fabrication between piping, fittings, structural members, instrumentation, electrical cable trays and conduits can be minimised. In addition, the use of virtual reality with a 3D model allows an operator to “move” around inside the deck structure and identify potential clashes, ensure the correct orientation of valves, study that ample access to equipment exists, and the equipment could be easily removed for maintenance or replaced. In general, there are two broad categories of equipment. One of these may be termed the “fuel sources” and the other the “ignition sources”. The primary goal in a deck layout should be to prevent hydrocarbon ignition and fire escalation by separating the fuel sources from the ignition sources. Any layout is a compromise that balances the probability of occurrence of these undesirable events against their consequences. Modern platform designs incorporate the learnings and recommendations from past disasters. Many of the decisions on selection and layout of the process and its control and safety systems are derived through quantified safety/risk analysis processes to ensure a low occurrence of accidental events, and in the event of an incident, to ensure safe evacuation of the personnel on board within acceptable risk levels. The cost of having to scaffold offshore for access to equipment can be very high. Therefore the designers should site equipment at deck level, wherever possible, or adjacent to access platforms. During design and development of the Process and Instrument Drawings (P&IDs) and plant layout [Croft-Bednarski and Johnston, 19991 the designer should ensure that the control valves are situated in easily accessible places so that start up, shutdown, isolation and maintenance can be carried out efficiently and safely. Another important aspect of the layout design is to identify areas of the plant, which would require frequent maintenance and ensure easy access to these areas.
- 217. Topside Facilities Layou! zyxwvutsrq Development zyxwvutsr 865 z The design of the control rooms in offshore platform topsides is very important. The control room operator must be able to control and manage safety and production critical emergencies efficiently and effectively to ensure that the platform can be shutdown and vented safely, ensure that fire control and mitigation efforts are initiated and that evacuation can be accomplished if necessary. On large platforms it may be necessary to have an emergency control room separate from the main control room to serve as a backup for these functions, if the main control room is not available. This room also duplicates as the emergency command centre. The layout and design of the personnel accommodation facilities is also very important. The operators must provide input to the layout of these facilities. Offshore personnel who normally work in areas such as the galley and sickbay should be brought in to work with the architects to achieve an optimal accommodation facilities layout. Designs should be based around natural colours and wherever possible, areas should give a feeling of comfort and security to the personnel who will be lodged in these facilities. zy 10.2.2 Deepwater Facility Considerations Deepwater floating facilities [Milburn and Williams, 20011 require a number of consid- erations during design that are not normally found in conventional fixed offshore structure. The motions of a floating facility must be taken into consideration during the design of its topside. Structural details that will be subjected to inertial loadings due to platform motions have to be checked for one or more load cases such as the operating, survival, transportation and installation conditions (for more details see Chapter 6, Fixed Offshore Platform Design). The sea state at which the facility should continue its normal operation must be deter- mined. The process vessels must be designed to meet these conditions. For a horizontal vessel, the motion effect might require special internal designs. Normally, the longer the vessel, the greater is the need for special care. Slugging and tight emulsions from subsea wells caused by the long vertical risers and cold sea temperatures must also be considered in the design of the production equipment. The expected motions in the design sea state should be supplied to all process vendors to insure their understanding of the operating conditions. Other systems that require careful consideration of motions are the drains and mechanical rotating equipment (turbines, compressors, generators, etc.). Deepwater subsea production presents a number of “flow assurance” problems for an offshore host facility. These include hydrates, wax, multi-phase flow, slugging and low temperatures. The host platform topside facilities may be required to provide for methanol storage or methanol recovery and regeneration for hydrate inhibition, equipment for heat- ing flowlines or for recirculating hot fluids, pig launchers and receivers for management of wax, slug catching capacity and valving for slugs and inlet heaters to increase temperature of the fluids for further processing. During design and fabrication, careful consideration must be given to regulatory authority requirements and the classing of the vessel. For SPARSin the US, the vessel is classed with the American Bureau of Shipping (ABS). The United States Coast Guard (USCG) and Mineral Management Service (MMS) are the principal regulatory authorities. Usually,
- 218. 866 zyxwvutsrqponm Chapter zy 10 z considerable USCG and ABS oversight occurs on topside systems. The systems which require such oversight include: primary topside structure, quarters and buildings, firewater system, life saving systems, compressed air supporting marine systems, diesel system, fuel gas systems supporting power generation, helicopter refuelling system, any systems for bulk storage and handling of liquids in the hull, deck drainage system, potable water system, sewage system and freshwater wash down system. In the USA, the USCG also reviews platform safety (including access/egress); lifesaving, fire protection, personal protection, ventilation and marine transfer facilities. Electrical systems and equipment include the alarm system, aids to navigation, communication system, area classification, power generation, emergency generator, electrical switchgear/MCC, lighting systems and fire detection. zyxwv 10.2.3 Prevailing Wind Direction In locating the equipment on the deck, it is important to consider the effect of the prevailing wind direction. An example “wind rose” summarising the wind data is shown in fig. 10.1. Certain equipment and elements should be placed upwind as much as possible. These components are quarters and control buildings, helidecks, air intakes of fired vessels, engines, turbines, air compressors and HVAC equipment. Similarly, certain components should be placed downwind, such as, vents, storage tanks, compressors, wellheads, etc. These will minimise the probability of escaping vapours being carried toward ignition sources and personnel. The main boat landing should be located on the leeward side, which will shelter the boat landing and keep vessels from hitting the platform. The main crane should be located on the boat landing side. Where high current is not aligned with the wind, the relative effects of each must be considered in the design of the component placement. N S Figure 10.1 Typical average yearly wind rose shown as percent occurrence per year
- 219. Topside Facilities Layout zyxwvutsrqp Deielopnient 867 z Gas Compressors The main escape areas (such as the safe gathering or mustering areas, helidecks, primary escape routes, stairs to boat landing, etc.) should be located upwind, wherever possible. However, rarely can it be guaranteed that “prevailing” wind conditions will occur at the time of the accident. Thus, secondary means of escape should be located downwind. Grinding Machinery zyx 10.2.4 Fuel and Ignition Sources Typical equipment found on the topside may be categorised as either “fuel” or “ignition” sources. Table 10.1below shows these two categories of equipment and other major topside fuel or ignition sources as listed by API 145 (API, September 1993). The desired topside locations for the fuel and ignition sources listed in table 10.1 are given in table 10.2. Liquid Hydrocarbon Pumps Table 10.1 Equipment zyxwvu - fuel sources vs. ignition sources (From API 145) Cutting Machinery or Torches I ~ Fuel Sources I Ignition Sources ~ Portable Fuel Tanks I IWellheads 1 Fired Vessels I 1Manifolds ICombustion Engines (including gas turbines) 1 1Separators and Scrubbers Electrical Equipment (including offices and buildings) 1Coalescers 1Flares I 1Oil Treaters ~ Welding Machines 1 1Hydrocarbon Storage Tanks IStatic Electricity 1Process Piping i Lightning I 1Gas-Metering Equipment 1Spark Producing Hand Tools I 1Risers and Pipelines 1Portable Computers I 1Vents 1 Cameras I ~ zyxwvutsrqp 1 1Pig Launchers and Receivers 1Cellular Phones I 1Drains 1Non-Intrinsically Safe Flashlights i IChemical Storage i 1Laboratory Gas Cylinders I I I sample Pots I I
- 220. 868 zyxwvutsrqpo Chapter z IO z Table 10.2 Recommended topside location objectives for fuel and ignition sources listed in table 10.1 (From API zyxwv 145 (API, September 1993)) Area Location Objective Example Equipment Source Typc Types Wellhead Minimise sources of ignition and fuel supply Wellheads, Chokes, Manifolds. Headers Fuel Protect from mechanical damage and exposure to fire Unfired Process Minimise sources of ignition Fuel Manifolds and Headers, Separators, Gas Sales Station, Pig Traps, Heat Exchangers, Water Treating Equipment, Pumps, Compressors, LACT Units Hydrocarbon Storage Minimise sources of ignition Fuel Ignition and fuel Storage Tanks, Gunbarrel Tanks, Sump Tanks, Produced Water Treating Tanks Fired Treaters, Line Heaters, Glycol Reboilers Direct Fired Process Minimise fuel supply Machinery Minimise fuel supply Generators, Electric Hoisting Equipment, Air Compressors, Engines, Turbines Office, Control Room, Switchgear/MCC, Warehouse, Maintenance Areas/Building Pig Launchers, Pig Traps, Valve Stations, Meter Stations Ignition Ignition Fuel Quarters/ Utilities Building Personnel safety Minimise sources of fuel Pipeline Minimise sources of ignition Protect from mechanical damage and exposure to fire Flares Minimise fuel sources Minimise ignition sources Vents
- 221. Topside Facilities Lajour Developmenr zyxwvuts 869 z 10.2.5 Control and Safety Systems zyxwvu The control and safety systems on platform facilities [Milburn and Williams, 20011 generally include: Either local or central operational control systems Data acquisition systems Manual operator interface Local equipment control and shutdown systems Well control and shut down systems Emergency Shut Down (ESD) System Fire detection systems Combustible gas monitoring systems 10.2.6 Firewalls, Barrier Walls and Blast Walls For safety reasons. adequate barrier and firewalls should be considered for areas where it is desirable to attempt to isolate certain areas where explosion, spillage or fire is possible. Barrier walls impede escaping gas or liquid leaks from entering an area with ignition sources. Firewalls provide a heat shield to allow personnel escape and protect potential fuel sources. Blast walls contain an overpressure from an explosion in a confined space from causing secondary damage on the other side of the wall. The disadvantages of firewalls, barrier walls and blast walls are that they restrict venti- lation, hamper escape and can, in themselves, help create overpressure in explosions. Thus, any decision to include one of these walls in the layout must balance the potential detriments against the potential benefits. Careful consideration is required for location of Shut Down Valves on lines that penetrate walls. 10.2.7 Fire Fighting Equipment Fire fighting equipment should be easily accessible in any location on the deck. This is particularly true for manned platforms. Hose stations should be located so that two hoses can reach any point of the deck. Firewater pumps, fire fighting chemicals and hose stations should be accessible and removed from locations where fire might occur. Ramps should be provided for wheeled chemical units. Spray systems should cover the entire area and point upwards at the wellheads, rather than downward. Automatic fire suppression systems can be considered for enclosures containing an ignition source, which cannot be isolated from a fuel source. In designing fire-fighting systems, consideration should be given to providing two separate pumps on opposite sides of the platforms so that damage to one would not likely cause the other to be inoperable. Firewater mains should be isolated, so that if a main is severed in an explosion, pressured water can still be delivered to the intact system. 10.2.8 Process Flow A well-developed process flow diagram is necessary to define the parameters for design of individual pieces of equipment. In laying out the equipment, a logical and orderly flow path is desirable from wellhead to sales meter. This minimises the required piping while reducing
- 222. 870 zyxwvutsrqpo Chapter z 10 the inventory of potential fuel, which could feed a fire. Locating the equipment solely to minimise piping for main process streams is often riot the best answer because of other safety issues involved. The needs to separate fuel and ignition sources, to consider prevail- ing wind, and to allow for ease of maintenance may be overriding. In addition, equipment items have many connections in addition to the main process flow and the most efficient piping arrangement may not necessarily follow the main flow pattern. The amount of high temperature piping should be minimised as well to reduce heat loss and insulation requirements. High-pressure piping should be kept away from high traffic areas and moving equipment. Long piping runs should be avoided where pressure drop is critical. The need for gravity flow may dictate relative vertical positioning. zyx 10.2.9 Maintenance of Equipment Provide adequate room for operations and maintenance. This includes the following: Pulling fire tubes from fired heaters Pulling tube bundles or plates from heat exchangers Removing compressor cylinders Replacing turbines, engines, generators, compressors and pumps Pulling vertical turbine or can-type pumps Removing plate packs from plate coalescers Pig insertion and removal Changing filter elements and filter media Removing and installing bulk storage containers Opening and removing inspection plates and manways Supplementary overhead cranes or lifting devices should be provided where necessary. Most injuries are due to falls from high places and handling of heavy loads. Layout should consider access to ladders and landings for maintenance purposes. 10.2.10 Safe Work Areas and Operations Provide safe welding and cutting areas for minor construction or routine maintenance. Floors should be solid and adequate ventilation and separation from fuel sources should be provided. Isolate the work areas from drains containing live hydrocarbons with liquid seals. Attention should be given to equipment handling requirements and weather protection. Operations should be planned for production, drilling, completion, wireline, pumpdown, snubbing unit work, construction activities, surface preparation and painting, removal or installation of wellhead equipment and installation of conductor pipe. Planning is required for equipment used in all phases of work anticipated. Adequate space and handling equipment are needed for consumables and support operations. 10.2.11 Storage Storage areas should be provided for diesel fuel, treating chemicals (e.g.corrosion inhibitors, demulsifiers, hydrate inhibition, glycols, biocides, etc.) and waste fluids. Storage for spare
- 223. Topside Faciliries Layout Development zyxwvutsr 871 z Figure 10.2 Example storage and maintenance area parts and solid consumables is normally provided in buildings, shops or warehouses. An example of a storage and maintenance area is shown in fig. 10.2. 10.2.12 Ventilation Due to the possibility of accidental flammable gas and flashing liquid discharges (leak, incorrect valve opening, sampling, etc.), adequate ventilation is a critical safety
- 224. 872 zyxwvutsrqpo Chapter z 10 consideration. Provide ventilation where necessary to disperse hazardous vapours and prevent their accumulation in gas traps. Enclosed buildings, which contain ignition sources, should be pressurised to prevent entry of flammable gas to the external atmosphere. Locate any air intakes in a safe area. Completely enclosed structures which house hydrocarbon fuel sources should have an air circulation and exhaust system to replace accumulated vapours with fresh air. Provide space for ducting. Fire and blast walls reduce natural ventilation. Try to keep at least two sides of the wellhead area open to natural ventilation. z 10.2.13 Escape Routes Provide two independent escape routes from each location. Maintain escape ways with a minimum clearance width of three feet, free of obstructions and with adequate headroom. Two stairs between all levels at opposite ends of a platform are preferred. Enclosed areas with fuel sources should have two exit doors, which open to independent escape routes. Evacuation can be provided through use of boat landing, survival craft and helideck (helideck may be difficult to use in case of a hydrocarbon related emergency). 10.3 Areas and Equipment The following are items to be considered in the design and layout of the deck for the different areas described in table 10.1 (Section 10.2.4). 10.3.1 Wellhead Areas Potential for uncontrolled flow and high pressures exists. The following considerations should be given to the wellhead areas: Provide adequate ventilation. Protect from sources of ignition, other large inventories of fuels, machinery and dropped objects and traffic. Protect equipment and instrumentation from drilling and completion fluid spillage. Provide unobstructed access to and egress from wellheads and separate them from the living quarters. For the wellhead area, size is a function of the drilling or work over rig and the number and spacing of the wells. Tight spacing makes access and escape paths difficult. On large platforms, the wells are usually isolated on one end of the structure. A firewall is sometimes placed to isolate the wellhead area from the production equipment. 10.3.2 Unfired Process Areas Unfired process areas are a potential source of fuel. Considerations should be given to the vertical placement of equipment. Liquid leaks from this area could ignite on hot surfaces or ignition sources below. Gas leaks could ignite on hot surfaces or ignition sources above. The unfired process area is usually located near the wellhead area. The area should be protected from dropped objects. An estimation of space required for these process vessels
- 225. Topside Fucrlitres Lujout Development zyxwvutsr 873 can be based on the following assumptions for piping twelve inches in diameter and smaller: Piping for horizontal vessels will take up an area of about four feet wider and two feet longer than the vessel itself, Piping for vertical vessels will take up an area of about two feet wider and four feet longer than the vessel diameter and Additional space is needed for walkways. 10.3.3 Hydrocarbon Storage Tanks Hydrocarbon storage tanks can provide a large inventory on the platform to feed a fire. There is a potential for a tank roof to fail, if subjected to overpressure. Also, tanks can be easily punctured. Therefore, careful considerations should be given to separate tanks from ignition sources and from other equipment, which can add fuel to a fire, such as, wellheads, pipelines and risers. Protect other equipment from liquids spilled from tanks and provide containment, where necessary. Protect from movement of equipment, which can puncture the tank. On offshore platforms, space can be saved using rectangular tanks. Locate oil tanks on the upper level, if possible, since it is very likely that the roof will fail if the liquid in the tank catches fire or if the tank is over pressured. Separate storage tanks for diesel or lube oil can be avoided by using the interiors of deck legs or crane pedestals. In sectionalised tanks, it is sometimes desirable to store non-flammable liquids between the stored fuel and the potential ignition sources to act as a safety buffer. Atmospheric tanks containing crude oil must be vented. If the vented gas is not to be recovered, it should be routed to a vent stack on the downwind side. Level gauges, controls and access will normally require about three feet on one side of a tank. On the sides of tanks without piping, only a walkway will be necessary. 10.3.4 Fired Process Equipment Direct fired process equipment is a source of ignition. If it contains flammable liquid (crude, gas, glycol), then it is also a potential fuel source. Air intakes should be from the perimeter of the platform on the upwind side to avoid sucking in hydrocarbon fuel with the air. The hot exhaust stack should be isolated from potential oil spills, since the pipe may be hot enough to ignite the spilled oil. Consider firewalls to protect surrounding equipment. Suggested clearance around fired process equipment is at least zyx 15 ft. Maintenance space must be provided for pulling the fire tube. Fire tube maintenance will also require lifting equipment. 10.3.5 Machinery Areas Be aware of oil leaks from above or gas leaks from below the sources of ignition (especially hot surfaces) in the machinery areas. Failure of mechanical seals or packing in compressors and pumps can provide fuel. The probability of failures of piping and connections in machinery areas are higher than normal due to vibration. Machinery areas, which do not contain flammable fluid, can be located near quarters or office/warehouse/auxiliaries as both are ignition sources. If flammables are present, then
- 226. 814 zyxwvutsrqpon Chapter 10 z the machinery area is a potential source of fire and should be separated from wellheads, pipelines, risers and tanks, which could escalate the fire, and separated from quarters due to hazard to personnel. Consider enclosing turbine and engine driven equipment and providing the enclosure with fire and gas detection and suppression equipment. zy A positive pressure could be maintained to exclude migration of gases into the area and disperse leaking gases. Isolate turbine inlets from ingesting gas with the air. AC motor driven pumps and compressors with proper electrical classification can be installed in process areas. DC motors are an ignition source. which when used on pumps or compressors should be in an enclosure with gas detectors and positive pressure from a safe intake source. Adequate space and hoisting capability should be provided. Generally, three feet of space on each side of a skid plus special clearances are needed. Heaviest and largest parts can be moved to an area accessible by crane. Noise should be taken into account as well. 10.3.6 Quarters and Utility Buildings Protection from external fires, noise and vibration is needed for these areas where there is a concentration of personnel. Consider fire resistant construction materials for the quarters. Potential sources of ignition from cooking, smoking and electrical equipment should be studied. Isolate quarters from potential gas leaks. Try to locate the quarters away from sources of noise and vibration. A firewall may be advantageous if the building cannot be safely located away from hazardous equipment. Minimise windows, which face the process area. Pay attention to escape routes and minimise exposure of personnel to radiation from potential flame sources. Try to locate utilities near the quarters building to minimise piping and conduit runs and minimise the external exposure to the quarters. In locating the quarters, consider the proximity to electrical generation, sewage treatment, heating, ventilation and air conditioning and potable water supply. 10.3.7 Pipelines Potential uncontrolled flows from pipeline risers and pig traps and launchers should be separated from quarters, control buildings and wellheads. Consider automatic Shut Down Valves (SDVs) and protect them from blast, fire or dropped objects by location of firewalls. Do not install instruments, vent valves or drain valves outside of SDVs. Risers should be protected from boat impact and dropped objects. Provide space for access to risers and space and work platforms for access to pig traps and launchers for pig removal. Consider the need for lift equipment for large diameter pigs. 10.3.8 Flares and Vents A vent is a potential gas fuel source and a potential liquid fuel source due to carryover from the vent. This is also true for a flare, if the pilot fails. A flare, or a vent, which has been ignited by lightning, is a potential ignition source. Liquid carryover from a flare is a potential ignition source as well. A flare may become a potential source of dangerously high SO2levels and a vent may become a potential source of dangerously high H2S levels, if H2S is present in the gas. If the pilot fails, a flare may become a potential source of dangerously high H2S levels.
- 227. Topside Facilities Layout Development zyxwvutsr 875 The potential danger from radiant heat from a flare as well as from accidental ignition of a vent should be taken into account for all potential wind directions. The normal flow from the flare provides continuous exposure to radiation, while a short-term radiation exposure is given from the emergency relief. Vents can give a short-term exposure from the normal flow if accidentally ignited. Dispersion of gases from vents must not create a problem with helicopter and boat approaches, air intakes for turbines and drilling derricks. An adequate scrubber should be provided for all flares and vents to minimise the possibility of liquid carryover. zyxwvuts 10.4 Deck Impact Loads The deck of an offshore structure is generally positioned at an elevation above the maximum water level that may be reached by a statistically probable wave crest, which may be experienced throughout the structure’s operation. This elevation is determined as part of the design process (see Chapters 3, and 6 of this handbook) from probability-based models aimed at predicting the largest wave in a particular return period. The maximum lateral pressure exerted on a structure by a wave occurs at or near a wave crest. The preference is to position the deck at an elevation sufficiently high to avoid an impact between the wave crest and a large area of the structure. For various reasons, situations may arise where the probability of a wave impact on at least a small portion of the deck is high enough that an estimate of the wave impact load on the deck is required (see Chapter 4). Offshore structure designers traditionally use the design wave method to establish the ultimate design load. The design wave approach considers the largest wave that appears in a random wave time series. Estimation of wave impact load is generally based on a numerical model, which is based on empirical factors. These factors are derived from scaled model tests in which deck structures are modeled. One such test set up is shown in fig. 10.3 in which the deck of a jacket structure subjected to a high wave in a random sea time series was floated on a load cell to measure this impact load. The jacket platform in the picture is the Vermilion 46A platform, located approximately 30 miles offshore South of Figure 10.3 Wave impact load on a jacket platform deck model
- 228. 816 zyxwvutsrqpon Chapter 10 z New Orleans. The model platform, which has 40 legs, consists of two identical 20-legjackets connected together at the base. The deck of each platform was separately instrumented to measure the two-component horizontal loads produced due to the wave impact. The associated errors in the empirical factors are difficult to quantify. That is why an attempt is made to position the horizontal members and floor beams which make up the lowest level of the deck above the maximum expected wave crest including wave run up on the vertical members of the structure supporting the deck. The distance between the design crest elevation and the lowest elevation of the significant area of horizontal steel is called the “air gap”. API recommends five-foot air gap for Gulf of Mexico Platforms to protect equipment from splash damage as well as provide a safety factor against the calculation of wave crest elevation. For further details see Section 6.2.3.2. zy 10.5 Deck Placement and Configuration Almost every offshore structure includes some type of a deck. The size and design of topside depends on the type of structure and its function. Fixed structures can be a jacket type structure, which is piled to the foundation with a deck on top. A typical fixed jacket platform and deck structure is shown in fig. 6.11. Platforms could also be steel or concrete gravity structures, which have a deck on top; or wood, concrete or steel slabs with piles and cap beams. Platforms can also be designed to be floated into place and “jacked up” as are mobile jack-up rigs. Other structure types, such as articulated towers, guyed towers, Semi Submersible Vessels (SSVs), SPARS or Tension Leg Platforms (TLPs) will have a deck which may either be an integrated part of the structure itself or installed on top, as with a fixed jacket. A ship shaped FPSO will have a module support frame, which supports the equipment 10 or more feet above the ship’s deck. The module support frame performs the same function as a deck. 10.5.1 Horizontal Placement of Equipment on Deck From a structural efficiency standpoint, it is beneficial to place heavy equipment near truss supports and to try to balance the vertical load on each leg. Adequate room for future equipment additions should be provided on the top deck and along the perimeter of the deck. Provide clearances for pad eyes and lifting slings. The need to keep deck equipment weight within capacities of lift barges may necessitate that some of the equipment be installed in separate lifts, require two or more decks side by side or necessitate a float over system (Sections 10.6.4 and 6.2.1.1). Rotating equipment should be oriented with its long axis transverse to the platform floor beams for increased stiffness. Allocate space on the top deck or around the perimeter for future equipment. 10.5.2 Vertical Placement of Equipment Allow adequate height for piping (e.g. relief valves, gas outlets) and maintenance. It may be necessary to have a tall piece of equipment penetrate the deck above due to its height. Place the equipment to take advantage of gravity flow. Pumps with a high negative suction head requirements may be located at a low elevation. Provide hatches in the upper decks or porches in lower decks for crane access. Locate heavy equipment as low as possible to
- 229. Topside Facilities Layout Development zyxwvuts Table 10.3 Typical load intensities on deck zyxw 877 z 1 Load Type I Item 1 Load Intensity 1 1Dead Load 1Floor beams and plate 150 psf I Derrick Load Live Loads for Floor Beam Design I IDeck Truss, Jacket and Piling 1Carry Down to Lower Levels 170% of live loads 1 lower the vertical centre of gravity, which will optimise stability, minimise dynamic response and aid in deck transportation. Open gravity drains must flow to a low point sump, which could be located in a subcellar deck. If specific loads are unknown, the deck should be designed based on the typical load intensities for the different types of loads shown in table 10.3. 10.5.3 Installation Considerations Installation procedures should consider the availability of lift equipment vs. minimising offshore hook-up time. Evaluation of the alternatives should take into account lift weight, available lift equipment and time required for the installation. Types of lifts that should be examined are single hook lift, lift with a spreader bar or a two-point lift (fig. 10.4). The deck on barges, concrete towers, etc. is normally installed as a unit. 10.5.4 Deck Installation Schemes There are several types of decks that may be installed on a structure based on their construction and transportation method. An Integrated Deck is one in which the equip- ment is pre-installed on the deck at an onshore yard. In this case, the deck structure supports the skids directly as well as provides any lateral support needed. A multi-piece Integrated Deck is used where the complete deck is too costly to lift in one piece and the platform has eight or more legs. A Modular Deck will have its equipment installed in modules. The module support frame may be an integral part of the tower or installed on the platform as a separate unit. Modular decks allow the fabrication workload to be spread out among several yards so that different yards may work on different modules in parallel, thus potentially decreasing the construction time. Lastly, Skidded Equipment containing some piping, valves and controls, but generally smaller than individual modules, may be used to minimise construction effort at the yard assembling an integrated deck or large module. Skids are generally transportable by truck and thus can be bid to a number of smaller fabricators for increased competition, lower project cost and shorter construction
- 230. 878 zyxwvutsrqpo Chapter z IO z SLlNQ zyxwvutsr SINGLE POINT 8PREABER zyx BAR SLING MooK3 6"""" SLINO TWO POINT Figure 10.4 Different arrangementsfor lifting and installing a deck on a platform time. Skids may be installed inside a module or an integrated deck. They do tend to increase the structural weight, however. The different installation schemes described above are depicted in fig. 10.5. The top figure shows integrated modules or deck installed on a steel tower with a few legs. On a shallow water concrete barge, the deck may be composed of skidded equipment lifted onto a steel or concrete deck or an integrated structure placed on steel or concrete columns. The bottom figure shows an integrated deck on a multi-piled platform. The cap beams are shown here. Various installation schemes on piled jacket structures are shown in fig. 10.6. In the top figures, modules are placed on the deck by using offshore lifting cranes. For this purpose a module support frame should be designed which accommodates the modules. In the bottom left figure the deck is brought on a barge, which is positioned in between the jacket legs. The deck structure is then lowered on the jacket legs and secured in place by use of lowering mechanisms or de-ballasting the barge. In the bottom right figure, two four pile integrated decks are installed on an eight-pile jacket with spanning integrated deck insert added next.
- 231. Topside Facilities Layout Development zyxwvutsr 879 z Figure 10.5 Deck installation scheme on piled jacket structures or floating barge Figure 10.7 shows the layout of a deck installed on an FPSO. In this case the helideck and the personnel quarters are separated from the process equipment by a sufficient distance. Such a separation is possible because the ship-shaped structure provides ample space on its deck. The flare is placed far away from the personnel and fuel sources. 10.6 Floatover Deck Installation Deck installation using the floatover deck concept [Salama, et a1 1999 and Section 6.1.1 of this handbook] in lieu of the traditional crane vessel lifting is a well-accepted method. This method provides an attractively cost effective way to install decks, especially when the deck weight exceeds available crane lifting capacity. The floatover method eliminates the use of
- 232. 880 zyxwvutsrqpo Chapter zy 10z MODULES zyxw MODULE 8UPPQRT FRAME MODULES AND SUPPORT FRAME zyxw DECK FLOAT BARGE THROUGH JACKET MULTI-PIECE DECK Figure 10.6 Several methods for installing decks on piled jacket structures heavy lift crane vessels since it uses the cargo vessel itself as an installation vessel. This method has been successfully used offshore in numerous deck installations. If environ- mental conditions are favourable and a protected deepwater site is available, a catamaran type floatover installation is also possible. In this method, the deck is transported with two vessels and lowered over the platform, which may be a fixed or floating structure. The floatover installation can be accomplished in two different ways. In the first method the installation barge enters inside the jacket (as shown earlier), moored to it and then lowered down by de-ballasting, gently transferring the deck load onto the jacket legs. The barge is then retrieved from under the installed deck. This method is called floatover installation by barge ballasting. The second method is similar to the first except that hydraulic jacks are used to lift the deck prior to entrance in the jacket and then used again to rapidly transfer the deck load to the piles. This method is called floatover installation by jacking system. This second case is mainly used when the deck is transported with a low centre of gravity above installation barge deck and is then raised prior to entrance in the jacket to maintain sufficient clearance for the entry. Variations of this method are required in areas such as West Africa where ocean swells could cause damaging barge impact loads on the jacket if the barge ballasting method is used.
- 233. Topside Facilities Layout Development zyxwvuts 881 z Figure 10.7 Module support frame and equipment installed on an FPSO deck zy The selection of the size of the installation barge so that its width would fit the space between the jacket legs while providing adequate stability during transportation, is a vital issue for the whole operation, since the whole philosophy of the installation is based on this selection. Additionally, a fendering/shock absorbing system between the barge and the jacket legs is required in order to prevent steel-to-steel contact at any stage of the installation. The selection of the floatover installation method will have a major impact on topside layout and design and the selection of the installation barge. The floatover method may look simple at a first glance, but would require considerable preparation for a successful operation. The Crane Vessel Lifting Method is compared against the Floatover method in table 10.4. 10.7 Helideck The helideck can be the roof of the quarters building. However, installing the helideck as a separate level over the roof of the quarters building at an additional expense has the advantage of isolating vibration. The size of the helideck is based on its intended use and
- 234. 882 Engineering analysis involved zyxwvuts Chapter 10 Requires several types of marine and structural analyses zyxwv Table 10.4 Comparison of floatover and crane vessel lifting methods Installation aids/equipment requirements I Parameter zyxwvu 1 Floatover Method Cargo barge rigged with installation equipment. Fendering and mooring system required Derrick Lifting Method Hookup and commissioning ICargo barge 1Required (generally larger) Single piece completed onshore. Efficient layout and piping runs. Required System requirements IHeavy lift crane vessel INot required Requires an accurate cargo barge ballasting system Required ISensitivity to weather More weather sensitive (using cheaper and more Droductive man-hours) Less weather sensitive Most done offshore (using expensive and less productive man-hours) Hook up and commissioning requirements Does not need a special jacket design. Jacket design governed by topside layout. Requires bigger and usually heavier jacket. Jacket design governed by the installation method. Only lift analysis required Weight limitations Cargo barge capacity is the limit Derrick lift barge capacity is the limit Tug boats requirements for the operation Requires 3 or more tugs Generally one tug needed Lifting gear and spreaders required May have to be installed in two or more pieces; significantly increases HUC time and cost. Deck strength may be governed by installation loads Deck structural strength requirements Deck strength generally not governed by installation loads May require a specific crane vessel. Installation dependent on vessel availability Less weather dependent Installation time Weather dependent. Installation of After deck installation secondary items inside the jacket perimeter Before deck installation Bumpers and guides required Fendering system requirements Three different types of fendering required for the deck and the jacket IRisks during installation ILow Relatively high
- 235. Topside Facilities Layout Developnient 883 z the type of helicopters that will be landing. The surface area of the helideck must exceed that of the helicopter’s rotor diameter for proper ground cushion effect. The perimeter safety shelf may be solid for increased ground cushion area or open netting. The landing/ departure paths for the helicopter should be provided. All tall objects should be marked with a contrasting paint scheme. Gas should not be vented near a helideck. Gas injected with ambient air can cause the helicopter turbines to overspeed. For further details and recommended practice, see API Recommended Practice 2L, Planning, Designing and Constructing Heliports for Fixed Offshore Platforms (API, January 1983). zy 10.8 Platform Crane The main function of the platform crane is to load and off-load material and supplies from boats. The crane is usually located on the top deck over the boat landing area. It is recommended that an open laydownistorage area be located near the crane on each deck level. Loading porches should be provided on the lower deck for easier access. Hatches may be required through the main deck to access equipment on lower levels. The crane is also used for routine equipment maintenance, including handling such items as compressor cylinders, pumps, generators and fire tubes in fired vessels. Localised hoists or monorails may be needed in an area not accessible with the platform crane. Two cranes may be required on large platforms or in areas with rough seas. For further guidance and details please refer to API Recommended Practices 2D “Operation and Maintenance of Offshore Cranes (API, March 1983 and June 2003)”. 10.9 Practical Limitations The layout of equipment, facility and operation is always a compromise as it is not possible to fully separate all equipment from each other. Trade-offs are required. The first step in laying out the equipment for a specific layout and installation concept is to draw a wind rose. The wellheads are then located. These may be pre-determined because they have to be in platform legs or must be accessible from a rig. If a platform rig is required, it is normally laid out next. As a guideline, the production equipment is laid out with preference given roughly to the following hierarchy: Isolate quarters and helideck on windward side. Place vent or flare on leeward side and locate cranes. Separate ignition sources from fuel sources where possible. Locate rotating machinery for access to cranes. Put utilities and water handling equipment near quarters. Optimise placement of equipment to minimise piping. 10.10 Analysis of Two Example Layouts Two different deck layouts are compared in this section. The examples are taken from an API Recommended Practice, which is no longer in print (API RP 2G 1974). The layout
- 236. 884 zyxwvutsrqpo Chapter z 10 z - i i is, S E P A ~ T ~ R 1 PROiECllVE WA SUBSURFACE CONTROL I I 1 GAS SALES STATON SKIMMER 81 WATER : i ~ PIPELINE sE PIPELINE zyxwvutsrqpon " a mklim TREATMEN1 W I I , TWO LCYEL BUILDING L W E R LNEL - COMPRESSOR UPPER LEVEL - HEAT RECOVERY (ENCLOSED) I REBOILERS 1 COOLERS I GAS CONTACTOR FIREWALL OIL ZRODUCTION FACILIN 70 X 148 FT PLATFORM zyxwvutsrq Figure 10.8First example layout, oil production facility (API, January 1974.Reproduced courtesy of the American Petroleum Institute). in the first example given in fig. 10.8 shows a preliminary oil production facility layout on a two level 70'x 148' platform deck. The prevailing wind direction is from the northwest. The following is a brief discussion of the positive points of this layout: 1. 2. Protective walls on both the cellar and main deck effectively isolate the hazardous wellhead area from the remainder of the platform. The quarters building is located as far away from the wellhead area as possible and is protected with firewalls on the sides facing the process equipment.
- 237. Topside Facilities Layout Development 885 3. zyxwvuts 4. zyxwvut The two escape routes of the main deck are located near the quarters building and are partially shielded by the building itself. The cellar deck is well laid out with the separators conveniently located near the wells. The skimmer and water cleaning skid, the oil automatic custody transfer (LACT) unit and the gas sales station are located between and isolate the fuel source of the separators from the potential ignition source of the engine driven pipeline pumps. The cellar deck firewater pump has a duplicate backup pump, which is located away from the main pump in the event of a local area problem. The enclosed compressor on the main deck is vertically isolated from the reboilers and heat recovery units by using a two level building. Good vertical isolation has been obtained for the major fuel sources of the main and cellar decks by placing the oil treaters and storage tank directly above the separators. 5. 6. 7. Some of the negative points concerning this layout are: 1. 2. 3. 4. zyxwvu 5. 6. zyxwvutsrq 7 . There is only one escape route around the protective wall on the main deck. The main deck is quite congested and access around the oil treaters is restricted. The large fuel sources represented by the oil treaters and the storage tank on the main deck are a major hazard to the quarters building if a fire should occur. The compressors are located adjacent to the quarters and the generators beneath the quarters presenting noise and vibration problems. The flare boom is located near the helideck. The following sources of potential high-pressure gas leaks are located near the quarters: contact tower, compressor and gas sales. The platform crane cannot be used to maintain either the compressor or generator. No provision is made for crane access to lower deck. The layout in the second example shown in fig. 10.9 shows another oil production facility layout on a two level zyxwvu 72'x 150' platform deck. A discussion of the positive points of this layout follows: 1. 2. 3. Quarters building is located as far from the wellhead area as possible. Quarters building is additionally protected by using firewalls on the inboard sides of the building. Locating the potable water compartment of the sectionalised tank adjacent to the quarters serves as a safety buffer between the personnel and a large concentration of clean oil. Oil treater and glycol reconcentrator utilise waste heat from compressors and are located directly above for compact, efficient arrangement. The cellar deck is well laid out with adequate space between skids. The platform crane can be used to maintain the compressor and aid in the maintenance of the generator. 4. 5. 6. Some of the negative points concerning this layout are: 1. No escape routes are shown off the cellar deck to the boat landing.
- 238. 886 zyxwvutsrqp h I 1 1 I I 1zyxw 1 - z Chapter zy 10 a 72'-0" SPACE FOR WORKOVER RIG 1 ELECTRICAL GENEQATORS 2 flRE FIGHTING EOUIPMENI I WORK BUllDlNG I uinnir IIPPFRzyxwvutsrqponmlkjih IMIC - . .. ._. . . COMPARIMENT ABOVE STORAGE, REMOTE FLAREzyxwvuts Z W O BBLS. UNDERWATERzyxwvutsrqp FMRL, OR FWRE BOOM ALSO ACCEPTABLE y WELL HEAD AREA hiGH & LOW PRESS. GAS SEPARATORSzyxwvutsrqponml g i B l j $ l a s@R 4 -+ OIL SEPARATORS HIGH zyxwvutsrqponmlkjihgfe II LOW PRLSS TEST SEPARATORS OIL SEPARATORS [1( S B U E M L L d % K ) @ WELL MANIFOLD E m 1 SURFACE^ SAFETY CONTROLS CELWR DECK OIL PRODUCTON FAC LlTy 77 X 150 FT PLATFORM - z c a h Figure 10.9 Second example layout oil production facility (API, January 1974. Reproduced courtesy of the American Petroleum Institute). 2. 3. The stair at the corner of the main deck should be rotated 180 degrees to put the top of the stair closer to the quarters building in case of an emergency. The large clean oil tank on the main deck has an exposed side near the workover rig area. There is a possibility that this tank may be punctured; a protective wall should be added to protect the tank.
- 239. Topside Facilities La)out zyxwvutsrq DeL'elopment zyxwvutsr 887 4. The gas sales station on the cellar deck is in an awkward location from a piping standpoint. In addition, the LACT unit and pipeline pumps are on opposite sides of the platform. No provision is made for crane access to lower deck to help in maintaining pumps, etc. The helideck is located close to the vertical flare tower. The compressors and generators are located close to the quarters presenting potential noise and vibration problems. The compressor is close to the quarters providing a potential source of high-pressure gas leaks. There may be insufficient vertical clearance for the gas contactor. zyx 5. 6. 7. 8. 9. zyxwvuts 10. Potable water in sectionalised tank is subject to possible contamination from diesel fuel in the adjacent compartment. zyxwv 10.11 Example North Sea Britannia Topside Facility The development of the North Sea SPAR Britannia topside facility was described by Garga (1999). Key factors that dictated the choice of facilities on the production platform were: Approximately 30 wells were drilled at the platform location in order to reach the vast aerial extent of the reservoir. The use of extended reach, near horizontal wells with measured depths of 31,000 ft (true vertical depths of 13,000 ft). Although 10 wells were to be pre-drilled, the remaining wells required a single, full drilling rig and services facility with a hook load of 510 ton, rotary torque of 60,000 lb ft 5000 psi surface rated equipment and mud circulation rates of 1500 usgpm at 7500 psi at 1.3SG, Basic separation of wet gas from the gas condensate reservoir into rich gas, condensate liquids and water. Produced gas to be dried to water dryness of 1 lb of water/MMSCF and to remain in dense phase at a pressure higher than 110bars at the on-shore terminal end of the gas export pipeline. The condensate to be stabilised to 125 TVP (true vapor pressure) and boosted to an inlet pressure of 180barg into the condensateexport pipeline, Production is intended from a large number of subsea wells. The control of production required a heating medium system, chemicals and their own dedicated test separator, Blast walls were used between hazardous and non-hazardous areas of the plant to control the effects of blast over-pressures and assist in the safe evacuation of personnel. The basic topside layout of the Britannia platform segregates the hazardous areas (con- taining hydrocarbons - process, well bay, gas compressor and condensate export) from the non-hazardous areas (containing no hydrocarbons - utilities, control room, accommoda- tion and lifeboats) by means of solid blast resisting walls and floors. The hazardous areas are themselves compartmentalised by blast walls in order to contain the consequences of likely events. This aspect of Britannia facilities is shown in fig. 10.10, in which the blast walls are shown as thick, black lines. Solid blast walls (shown as dark lines) separate hazardous process facilities from wellbay and from non-hazardous utilities.
- 240. 888 zyxwvutsrqpon Chapter z 10 z Figure 10.10 Side elevation of Britannia platform topside showing layout of facilities The safety design of the topside, apart from incorporating the usual active and passive fire protection systems, fire and gas detection systems and multiple safe evacuation devices, took a structured approach towards increasing safety with respect to fire and explosion events. This was achieved by incorporating zyxw - inherent safety via reduction of likelihood of leaks, measures to reduce ignition probabilities, improved ventilation and explosion venting, good detection of leaks and rapid isolation and blast protection for personnel, emergency equipment and critical plant. The notable design features that resulted from the above structured approach were: The knock-on benefit of minimum plant/minimum sparing in reducing hydrocarbon inventories in vessels, equipment and pipe work, Critical review and elimination of breaks or entries into pipework normally associated with maintenance or control needs. This review, undertaken with operations and maintenance personnel, was significantly aided by having all the key influencing parties in an alliance, working to common objectives, - open, high vertical height module spaces to provide good ventilation (over 100 air changes per hour) and explosion venting. Note that lowering module packing densities by enlarging roof heights goes against the philosophy of tighter packing to reduce weight and cost, but significantly reduces blast overpressures with its resultant benefits in lowering the weight and cost of blast resisting structures, Blast walls between process and wellbays (see fig. 10.10)designed to resist between 2 and 4 bars of overpressure. The blast wall between wellbay and utilities areas of double skin construction, also used as a potable water tank,
- 241. Topside Facilities Layout zyxwvutsrq Development zyxwvutsr 889 A large flare relief system of 830 MMSCFD, allowing blowdown to 7 barg in 13 min and atmospheric pressure in 30 min, High specification on cable entries and terminations into equipment in hazardous areas and isolation of power at 50% gas LEL (lower explosion limit), Extensive use of the latest technology in fire and gas detection and control systems, Use of electrically driven combined seawater and firewater duty pumps providing a more reliable, faster and high deluge water volume response, Critical systems (fire water, control cables) and their supports designed to survive high blast overpressures, zyxwvu - the adoption of self-verification methods, with their requirement to define “safety critical systems” and the attention to detail of the critical elements within those systems (in design, manufacture, testing, operations and maintenance). References API (January 1974).“Production facilities on offshore structures”,Recommended practice API-2G, (1“ ed.), American Petroleum Institute, Washington DC. API (January 1983). “Planning, designing and constructing heliports for fixed offshore platforms”, Recommended practice API-2L, (2nd ed.), American Petroleum Institute, Washington DC. API (March 1983). “Offshore cranes”, Specification API-2C, (3rd ed.), American Petroleum Institute, Washington DC. API (September 1993). “Recommended Practice for design and hazard analysis for offshore production facilities”, Recommended practice API-lU, (RP 14J), American Petroleum Institute, Washington DC. API (June 2003). “Operation and maintenance of offshore cranes”, Recommended practice API-2D, (5th ed.), American Petroleum Institute, Washington DC. Croft-Bednarski, S. and Johnston, K. (1999). “Britannia topsides: a low cost, safe and productive north sea facility”, Offshore Technology Conference, OTC 11018, Houston, Texas. Edwards, C. D., Cox, B. E., Geesling, J. F., Harris, C. T., Earles, G. A,, and Webre, Jr., D. (1997). “Design and construction of the mars TLP deck”, Offshore Technologj,Conference, OTC 8372, Houston, Texas, 5-8 May. Garga, P. K. (1999). “Britannia topsides: a safe and productive north sea facility at lowest cost”, Offshore Technology Conference, OTC 11015, Houston, Texas. Milburn, F. H. and Williams, R. H. (2001). “Hoover/Diana: topsides”, Offshore Technology Conference, OTC 13083, Houston, Texas. Salama, K. S., Suresh, P. K., and Gutierrez, E. C. (1999). “Deck installation by floatover method in the arabian Gulf”, Offshore Technology Conference, OTC 11026, Houston, Texas.
- 242. Handbook of Offshore Engineering zyxwvutsr S . Chakrabarti (Ed.) zyxwvutsrq C 2005 Elsevier Ltd. zyxwvutsr All rights reserved zyxwvut 89 z 1 Chapter 11 Design and Construction of Offshore Pipelines Andri: C. Nogueira and David S. Mckeehan INTEC Engineering, Houston, Texas zyxwv 11.1Introduction During the sixties, offshore pipeline design saw the vigour and strength for a young structured engineering field, as solutions to the practical problems demanded innovation and vision. Such initial vigour and strength is documented in numerous scientific papers and research reports of this era. For example, in the 1960s Shell Research and Development carefully studied and advanced the water depth of pipeline. Dixon and Rutledge (1968) published the stiffened catenary solution for offshore pipelines. In the mid 1960s, a straight stinger was used to lay pipe in the North Sea [Berry, 19681.A patent for the articulated stinger was issued in 1969 [Broussard, et a1 19691.The articulated stinger provided major technology advancement in the feasibility of laying pipe in ever-deeper waters. For the historically inclined reader, Timmermans (2000) presents an interesting overview of the development and achievements of the offshore pipeline design and construction discipline worldwide. The objective of this chapter is to serve as a reference and guide to the offshore pipeline engineer during the design process. The following aspects of offshore pipeline design are discussed: the establishment of a design basis, aspects of route selection, guidance in sizing the pipe diameter, wall thickness requirements, on-bottom pipeline stability, bottom roughness analysis, external corrosion protection, crossing design and construction feasibility. These topics encompass the majority of issues regarding offshore pipelines. Some issues not covered herein are expansion analysis, curve stability, risers and steel catenary risers (SCRs), analysis of the installation of in-line appurtenances, fracture analysis of weldments and subsea connections.
- 243. Chapter z 11 z 892 zyxwvutsrqp 11.2 Design Basis The first step in offshore pipeline design is establishing a concise design basis document (DBD). For consistency, every project requires, in its early phase, the establishment of the DBD. This is to be used as a reference by the design team for the different aspects of offshore pipeline design. The DBD provides basic project-dependent information, and enables consistency and correctness of project calculations, reports, bid specifications, contract documents, installation procedures, etc, with respect to the fundamental parameters of the project. For a major project, a DBD should include the following sections: zyx Development overview: describes the project location and basic layout. Reservoir and well information: provides reservoir characteristics, fluid rheology and production rates. Environmental defines geotechnical properties along the proposed route (shear strength, weight, etc.), meta-ocean data (waves and currents), and seawater temperatures/chemistry zyxwvu . Flow assurance: provides information on flowline parameters, e.g. operating pressures, temperatures and velocities; identify and address flow hazards such as hydrates, wax, scale, corrosion, slugging. Wellbore, drilling and completion information: provides safety valve philosophy, downhole chemical injection, completion design, downhole monitoring, general rig description, well servicing and intervention process, etc. Equipment design philosophy:describes the design and selection approach, standardisa- tion of components and interfaces, equipment design life, quality programme. Subsea trees and flowline]pipeline sleds: describes subsea trees, flowlineipipeline sled characteristics, tie-in jumpers, completion/workover system, equipment marking, corrosion protection. Production control system and umbilicals: establishes codes and standards, system overview, subsea instrumentation, redundancy, emergency shut-down valve (ESV) requirements, surface equipment, subsea equipment, power and hydraulic umbilical, methanol distribution umbilical (if required), intelligent well completions, metering. Pipelines: provides general characteristics (grade, size, water depth), route selection, applicable codes and regulation, system design requirements (design life, cathodic pro- tection system, etc.), risers and tie-ins, maximum shut-in pressure, corrosion allowance. Host facilities: gives general description of the host, process design, major equipment list, interface definitions. Operation and maintenance: outlines normal production parameters, start-up and shutdown procedures, routine testing requirements, pigging, system maintenance, abandonment philosophy. In general, the DBD is a living document that goes through several revisions during the course of a project. However, after the front end engineering design (FEED) phase of a project, the DBD must define the majority of the project requirements. After FEED,
- 244. Design zyxwvutsrqp and Construction of Offshore Pipelines zyxwvut 893 z a successful project will incorporate change only by an established “management of change” process, which provides evaluation of the proposed change, and its implication with regard to safety, cost and schedule. 11.3 Route Selection and Marine Survey In the FEED phase of a project, typically the seafloor bathymetry data is available to the so-called “regional survey” level. This means that coarse surface tow and swath bathymetry survey data are available for preliminary route selection, but not to a level of detail required for a finalisation. At this point, the pipeline lead engineer should select the base case route based on the regional survey data. If a challenging bathymetry is present, alternative routes should also be defined; environ- mental sensitivity zones should be avoided, as well as excessive span areas. During the detailed marine survey, a pipeline engineer should be on-board to perform a real-time bottom roughness analysis. Frequent communications should take place between the on-board pipeline engineer and the design office to assure a successful marine survey, which will suffice for purposes of supporting a final route selection as well as the required geohazard survey report. 11.4 Diameter Selection The selection of diameter is a process where the initial capital expenditure (capex) and operational expenditure (opex) are evaluated leading to an optimised design by minimising total cost through the life of the project. The main criterion for selection of pipeline diameter is the ability to carry fluids at the design flow rates, within the allowable pressure. Figure 11.1 depicts the processes and logic involved in the selection of an initial diameter, and the flow assurance work needed to guarantee operability of the system. To provide some reference point regarding diameter, flowrate and operating pressures, table 11.1 summarises the pipe diameters used in selected offshore developments. Key parameters are given, as well as references, for the interested reader to obtain more details regarding diameter selection and flow assurance. Diameters are provided as nominal outside diameter (OD). 11.4.1 Sizing Gas Lines The practice for selecting a pipe diameter is a detailed hydraulic analysis; especially for multi-phase flow with untreated gas. However, a quick way to estimate the size of dry, single phase, gas lines is to use the simplified equation (11.1) [McAllister, 19931. For small gathering lines, the answer will have an accuracy within 10% of that obtained by more complex formulas. 500 ID3@;-- P; zyxwv 1/z Q = where Q=cubic ft of gas per 24h, ZD=internal pipe diameter in starting point, P2zyxwvut =psia at ending point, L =length of line in miles. (11.1) inches, zy PI=psia at
- 245. 894 zyxwvutsrqpo [Project Canyon Express' Northstar Export Line2 zyxwvuts Chapter zy I 1 Contents and Maximum Pipeline Nominal OD peak flow rate operating pressure length Gas-condensate 4200 psig 55 miles 2 x 12 in. at 500 MMSCFD at 40°F 42" API Oil 1480 psig 6.0 miles 10 in. at 65,000 barrels at 100cFmax. zyx u Initial diameter selection Pressure data I Per day /Temperature/ . < I ~ ~ I data temperature zyxw Hydraulic analysis zyxw I I Gas-condensate at 600-1800 MMSCFD 1 1 1 1 1 Liquid flow Multi-phase Emulsions Gas flow Waxy crudes Pipeline type diameter selection and zyxwvuts 4 Chemical Iniection 3750 psig Figure 11.1 Diameter selection processes flow chart Table 11.1 Diameter for selected offshore projects Northstar Processed gas 1480 psig Gas Line2 100 MMSCFD 6.0 miles 10 in. l l 1 x 24 in. 1 x 36 in. 'Wallace, et a1 (2003) 2Lanan, et a1 (2001) 3Choate, et a1 (2002)
- 246. Design and Consirucrion zyxwvutsrq o f Offshore Pipelines zyxwvuts 895 For Throughputs (bbl per day) For example, given an 8 in. ID line, 9 miles long, if the pressure at the staring point is 485 psi and the pressure at the downstream termination is 283 psi, the total gas flow is estimated at: Use Pipe outside diameter (in.) Pressure drop (psi per mile) 500(8'):(485 zyxwvu +15)2-(285 + 15)2 = 34.1 million cubic ft per day (MMCFD) & zyx Q = 3000 to 7500 7500 to 16,500 zyxwvu 11.4.2 Sizing Oil Lines The sizing of oil lines is more complex than gas lines due to the different viscosities and specific gravities of crude oil. However, the table 11.2 from McAllister (1993) provides guidance in selecting line size, and pressure drop for oil of approximately 40" API gravity and 60 SSU viscosity. 6 % 16 8 % 10.5 11.5 Wall Thickness and Grade API 5L grade X-65 has become the steel grade of choice for deepwater offshore pipelines. The main reasons for this choice are cost-effectiveness and adequate welding technology. A lower grade, X-60, is typically used for SCRs, to ensure easier welding overmatch for these structures, and an improved fatigue life. For buried offshore pipelines in the Arctic, a more ductile, X-52 grade has proven the best choice for limit state design and the need for high toughness material that could sustain the high strain base design [Lanan, et a1 2000; Nogueira, et a1 2000; Lanan, et a1 20011. To calculate the required wall thickness for an offshore pipeline, three different failure modes must be assessed: Internal pressure containment (burst) during operation and hydrotest. Collapse due to external pressure. Local buckling due to bending and external pressure. Table 11.2 Crude oil sizing guidance 10 to 2000 13% 1 - 1 12000 to 3000 I4Y* I 32 I 116,500 to 23,500 110% I 8.5 1 123,500 to 40,000 112% I 7 1
- 247. 896 zyxwvutsrqpo Chapter z 11 z A fourth failure mode may be used to calculate the required wall thickness in deep water: Designing for each of these failure modes is discussed in each of the sub-sections below. A numerical design example, covering each failure mode, is given in Section 11.7. 11.5.1 Internal Pressure Containment (Burst) Pipelines to be installed in the Gulf of Mexico, or in any place within the jurisdiction of the Minerals Management Service of the United States, must comply with the appropriate Code of Federal Regulation (CFR). Three parts of these regulations are applicable for offshore pipelines: Buckle propagation and its arrest. Title 30, part 250 of the CFR [30 CFR 250, 20021 entitled “Oil and gas and sulphur operations in the outer continental shelf (OCS)”, and in particular subpart J entitled “Pipelines and Pipeline Rights-of-way”. This defines the so-called Department of Interior’s (DOI) jurisdiction, or a DO1 pipeline. Per 30 CFR 250.1001: “DO1 pipeline refers to a pipeline extending upstream from a point on the OCS where operating responsibility transfers from a producing operator to a transporting operator”. This is applicable to pipelines from wells to platforms. Pipelines in the OCS, which are not DO1 pipelines and are used in the transportation of hazardous liquids or carbon dioxide, must follow the Department of Transportation standards presented in 49 CFR 195 (2002), subpart A. These provisions are applicable for oil pipelines from platforms to shore, or other tie-in points into existing pipeline transportation systems. Pipelines in the OCS, which are not DO1 pipelines, and are used in the transportation of gas, must follow the Department of Transportation standards presented in 49 CFR 192 (2002), subpart A. These provisions are typically applicable for gas pipelines from platforms to shore, or other tie-in points into existing pipeline transportation systems. For simplicity, the following wall thickness design requirements are based on the provisions of 30 CFR 250.1002, entitled “Design requirements for DO1 pipelines”. The other two CFRs contain very similar design requirements. 30 CFR 250.1002 adopts an allowable stress design format. That is, the basic (burst) design equation sets the internal design pressure, zyxwv Pldrto a value such that the resulting hoop stress is a fraction of the pipeline yield stress. The relationship between P,J and the (nominal) wall thickness is given by: (11.2) Equation (11.2) above is given by 30 CFR 250.1002(a), which also defines the following terms (definition below are transcribed verbatim): Pid zyxwvuts =internal design pressure, t =nominal wall thickness, D =nominal outside diameter of pipe, S y =specified minimum yield stress,
- 248. Design zyxwvutsrqponm and Construction of Offshore Pipelines zyxwvuts 250 or less 897 1.000 zyxwv Table 11.3 Temperature de-rating factor, T,for steel pipe zyx 1Temperature (“F) 1Temperature De-rating Factor, T 1 , 1350 I0.933 I 1400 10.900 I450 10.867 1 F =construction design factor of 0.72 for the submerged component and 0.60 for the T =temperature de-rating factor obtained from Table 841.1C of ANSI B31.8, E =longitudinal joint factor. Obtained from Table 841.1B of ANSI B31.8 (see also According to 30 CFR 250, all pipelines should be hydrostatically tested with water at a stabilised pressure of at least 1.25 times the maximum allowable operating pressure (MAOP) for at least 8 h. The test pressure should not produce a stress in the pipeline in excess of 95% of the specified minimum-yield strength of the pipeline. The relationship between the maximum hydrotest pressure and the (nominal) wall thickness is similar to equation (11.2), and is given by: riser component, see table 11.3. Section 811.253(d)) - see table 11.4. (11.3) where Pm&,yd =maximum hydrostatic test pressure, and F= construction design factor, 0.95 for hydrotest. 11.5.2 Collapse Due to External Pressure During installation, offshore pipelines are typically subjected to conditions where the external pressure exceeds the internal pressure. The differential pressure acting on the pipe wall due to hydrostatic head may cause collapse of the pipe. Several design codes present formulation addressing the design against this failure mode. Amongst these codes, the most prominent are API RP 1111 (1999) and DNV OS-F101 (2000). The elastic collapse pressure [equation (11.5b)l and the plastic collapse pressure [equation (11.5c)l bound the problem. Timoshenko and Gere (1961) proposed a bi-linear transition between the two equations [see Timoshenko and Gere, 1961, figs. 7-91, which adequately bridges the two equations. API RP 1111 (1999) adopts a transition between the two equations [equation (11.5a)],which is simpler than the cubic interaction equation proposed by DNV. A comparison between the API and DNV collapse pressures, P,, normalised by the plastic collapse pressure, P,, is given in fig. 11.2. P, as calculated by equation (5.18) in the DNV OS-F101 (2000). uses ovalisation parameterf, = 1%, as defined in DNV OS-F101
- 249. 898 Furnace butt welded zyxw - continuous weld Seamless zyxwv Table 11.4 Longitudinal joint factor, zyxw E 0.6 1.o zy Chapter z 11 Electric fusion welded Spiral welded steel pipe Seamless Spec No. 0.8 0.8 1.o ASTM A 53 Electric flash welded Submerged arc welded ASTM A 106 1.o 1.o ASTM A 134 ASTM A 135 ASTM A 139 ASTM A 211 ASTM A 333 ASTM A 381 ASTM A 671 ASTM A 612 API 5L Pipe class 1E factor I Seamless 11.0 I Electric resistance welded 11.0 I Electric fusion arc welded 10.8 I Electric resistance welded 11.0 I Electric resistance welded 11.0 I Double submerged-arc-welded 11.0 I Electric fusion welded Classes 13, 23, 33, 43, 53 Classes 12, 22, 32, 42, 52 Electric fusion welded Classes 13, 23, 33, 43, 53 Classes 12, 22, 32, 42, 52 Seamless 11.0 I Electric resistance welded 11.0 1 Furnace butt welded 10.6 I (2000) equation (5.21). No factor of safety has been applied to either formulation. It can be seen that both codes yield very similar results for the collapse pressure. Due to their simplicity, the API equations are recommended for wall thickness design against collapse due to external pressure. Following API RP 1111 (1999), Section 4.3.2.1, the pipe collapse pressure zy P, (i.e. pipe collapse capacity) must be greater than the net external pressure (Le. effective applied external pressure) everywhere along the pipeline, as specified by equation (11.4) below: (Po- PJ if0 ' pc (11.4) where fo =safety factor: 0.7 for seamless or ERW pipe and 0.6 for cold expanded pipe, Po=external pressure and Pi=internal pressure.
- 250. Design and Construction zyxwvutsrq of Offshore Pipelines 899 z 1.o 0.8 0.6 0.4 DNV OS-F101 (2000) 0 0 zyxwvutsr 1 16 18 20 22 24 26 28 30 32 34 36 38 40 zy Dlt zyxwv Figure 11.2 Collapse pressure vs. D/t per API 1111 (1999) and DNV OS-F101 (2000) The collapse pressure is determined by equations (11.sa)-( 11.5~): collapse pressure 2E zyxwvu P - - ( L>i elastic collapse pressure 1 - v 2 D e - (1l.5a) (11.5b) t PJ= 2s) - D plastic collapse pressure (11.5c) where E =modulus of elasticity of steel, and u =Poisson’s ratio, 0.3 for steel. 11.5.3 Local Buckling Due to Bending and External Pressure This failure mode is typically most severe during installation when bending and external pressure effects are critical. However, local buckling also applies for the installed pipeline, in case of depressurisation. API RP 1111 (1999) and DNV OS-F101 (2000) have adequate formulations that address this failure mode, which are based exclusively on empirical data fitting. Once again, due to its simpler treatment of the subject, the API RP 1111 (1999) is the one presented herein. For Dit upto 50, the following interaction equation needs to be satisfied following API RP 1111 (1999), Section 4.3.2.2. (11.6)
- 251. 900 zyxwvutsrqpon Chapter zy I 1 z where zyxwvut E =critical strain (maximum compressive strain at onset of buckling) &b zyxwvutsrqp = &=critical strain under pure bending g(6)=(1+20 6)-’ =collapse reduction factor D m a x - D m i n D m a x +D m i n 6 = = ovality DmaX =maximum diameter at any given cross section D,,, =minimum diameter at any given cross section Equation 11.6 can be rewritten as: The bending strains shall be limited as follows: fie1 5 E h E 2 I E (11.7) (11.8a) (11.8b) where =maximum installation bending strain, ~2 =maximum in-place bending strain, f i =safety factor for installation bending plus external pressure, and f2 = safety factor for in-place bending plus external pressure. zyxwv A value of 2.0 for safety factorsf, andf2 is suggested by API RP 1111 (1999). Safety factor z fi may be larger than 2.0 for cases where installation bending strain could increase significantly due to off-normal conditions, or smaller than 2.0 for cases where bending strains are well defined (e.g. reeling). zyxwv 11.5.4 Rational Model for Collapse of Deepwater Pipelines The above API formulation is based on empirical data fitting [Murphey and Langner, 19851. Palmer (1994) pointed out that “it is surprising to discover that theoretical prediction (of tubular members collapse under combined loading) has lagged behind empirical prediction, and that many of the formula have no real theoretical backup beyond dimensional analysis”. Recently, this situation has changed with the rational model formulation presented by Nogueira and Lanan (2001). The rational model has been derived from first principles, e.g. equilibrium of forces and moments; and its predictions have been shown to correlate very well with test results. The cornerstone of the rational model is the recognition that when a pipe is subjected to bending moment, the longitudinal stresses generate transverse force components due to the pipe curvature. As a pipe bends, components of the longitudinal bending stresses act into the cross-section. This, in turn, generates a transverse moment, which ovalises the pipe cross section, or ring, until it collapses. A pipe under bending will collapse when its cross section (or ring) loses stiffness due to plastic hinges mechanism formation at the onset of local buckling. Therefore, when rings of the pipe lose their stiffness, the ovalisation
- 252. Design zyxwvutsrqponm and Construction of Offshore Pipelines zyxwvuts 90 z 1 (initially uniform along the pipe length) will concentrate at the weakest point along the pipe (e.g. a thinner ring) and a local buckle will form. If in addition to bending, pressure is applied, its effects are taken into account by noticing that it contributes to reduce the ring capacity to resist bending. This is due to the effects of the compressive hoop stress. Since this model has sound theoretical basis, it provides explanation to some intriguing issues in pipe collapse. For example, the rational model includes, by derivation, in its formulation the anisotropy ratio zyxwvu N = o o H / o o L , where zyxw o o ~ is the pipe yield stress in the hoop direction, and o o ~ is the pipe yield stress in the longitudinal direction. Tam, et a1 (1996) reported that when the anisotropy ratio is included in their model, its predictions fit more precisely the experimental results. However, Tam, et a1(1996) could not attribute a physical meaning to the ratio N. Following the rational model derivation, the explanation is that greater values of the longitudinal yield stress o o ~ (which generates the applied ring load) result in a greater applied transversal load, and lower values of the hoop yield stress o o ~ (which characterises the ring load capacity) result in a reduction in the ring capacity. Of course, this effect is numerically captured in the anisotropy ratio N, as defined above. The ratio N can be less than one especially for pipe manufactured by the UOE method. For explanations of other issues and for complete derivation of the equations of the rational model, see Nogueira and Lanan (2001). The model equations are given below. 8 - N P R A B & ~ +~ ~ ( 1 + 2 A B & p ) = ccO(l- P ~ ) ( 1 . 3 1 P ~ +1) Interaction equation T[ OoH Anisotropy ratio N = ~ OoL P Normalised pressure PR= - P I 2 o o H P -- ’- (D/t) Yield pressure f ( & T ) Ovality (due to bending and pressure) A ~ a p = ~ 1 -PlPc (11.9) (11.10) (11.11) (11. (11. 8N 1 n3&(D/t) Reference strain ETY = (11.14) (11.15) (11.16) Dmax - Dmin Dmax +Dmin Initial ovality AI = (11.17)
- 253. 902 zyxwvutsrqpo Chapter z 11 Hyperbolic ratio zyxwvut s zyx = Eco/&TY (11.18) 2&N 1.1N zyx E,, = - - __ Critical strain approximation - (D/O (11.19) In the above equations,p is the maximum applied external pressure, and ET is the rational model's critical strain. For example, given p , the solution of the above equations will yield ET. The data required to solve the above equations are: Pipe diameter zyxwvu (D) Pipe wall thickness (t) zyxwvu 0 In the case of external pressure only, the term with E= on the left hand-side of equation (11.9) vanishes. In order to obtain the correct results, equation (11.9) is re-arranged as shown by equation (11.20). This will lead to correct critical pressures for the perfect circular pipe. Pipe yield stress in the hoop direction ( C J ~ H ) Pipe yield stress in the longitudinal direction (0,~) (11.20) The collapse pressure predictions of equation (11.20) are shown in figs. 11.3 and 11.4, for pipe with different initial ovalities, compared to experimental results reported by Murphey and Langner (1985). The rational model shows slightly conservative results. The main I OD = 1 00 in, WT = 0 zyxwv 048inch, fso = 75 ksi ASTM 1015 Steel zyxwvu tube - D M= 20 8 o Experimental results (after Fig 13 Murphey & Langer 1985) -Nogueira and Lanan's Rational model predictions - -Rational model predictions with increased yield stress 1000 zyxwvuts 1 0 0 00 0 02 0 04 0 06 0 08 0 10 0 12 Initial ovality, A, Figure 11.3 Rational model prediction of collapse pressure vs. initial ovality, compared to experimental results for pipe with D/t =20.8
- 254. Design and Construction zyxwvuts o f Offshore Pipelines zyxwvut 903 z OD = 6 66 zyxwvu in, WT = zyxwv 0 190 inch, cro= 60 ksi 1600 1 reason for this conservatism is the model's elasto-perfectly-plastic material assumption, and that the yield stress given by the authors probably underestimates the actual yield plateau. It is common that the actual yield plateau to be 10-20 ksi higher than the yield stress. Therefore, the collapse pressures predictions are also given for higher yield stresses, as indicated in the figures. Predictions become closer to the experimental values, but are still conservative. The interaction equation (11.9) can be solved by means of spreadsheets. Results of critical pressure vs. critical strains predicted by the rational model are shown in fig.11.5, which also shows the interaction equation results of API RP 1111 (1999) and DNV OS-F101 (2000), for a pipe with diameter-to-thickness, D/t, ratio equal to 20. The API formulation is conservative for the most part of the interaction plot, except that the rational model is more conservative at very low levels of bending strain. The DNV formulation as well as the rational model produces somewhat similar predictions for higher bending strains, with the rational model formulation predicting the highest capacity for small levels of external pressure. Figure 11.6 shows rational model interaction equation predictions and compares them with experimental results presented in Table 5 of Fowler (1990). The rational model predictions are based on the average pipe properties. The rational model interaction equation predictions are very close to the experimental values. zyxw As a final comparison, results from the Oman-to-India pipe collapse programme [Stark and McKeehan, 19951are shown in fig. 11.7, together with the rational model predictions.
- 255. 904 zyxwvutsrqpon Chaprev 11 P lP , 0 0 2 0 4 0 6 zyxw 0 8 1 zyx 1 2 1 4 ET/G,DNVzyxwvu Figure 11.5 Nogueira and Lanan’s rational model pressure vs. bending strain prediction compared to API 1111 and DNV OS-F101 for pipe with D/t=20 4000 3500 3000 2500 2000 1500 1000 500 0 0.0 1 0 2.0 3.0 4.0 5.0 critical zyxwv strain,eT Figure 11.6 Rational model pressure vs. bending strain predictions (average pipe data used) vs. experimental results for pipes with D/t =23.9 (average)
- 256. Design and Construction of Offshore Pipelines zyxwvut 905 8000 zyxwvutsrq - zyxwv p zyxwvutsr CR 5000 - 4000 zyxwvutsr 1 +Experimental results, see Table 11 zyx 5 (Stark and zy -x- Rational Model predictions I McKeehan 1995) I 3000 zyxwvut 1 . . - __ . - - I 2000 1000 I I i 00% 02% 04% 06% 08% 10% 12% 14% 16% ET Figure 11.7 Rational model pressure vs. bending strain prediction vs. experimental results after Stark and McKeehan (1995) In this case, the individual pipe characteristics were taken into account, including hoop and longitudinal yield data (not shown). It can be seen that excellent agreement is also obtained for all test results. For completeness, the original data published by Stark and McKeehan (1995) is provided herein in table 11.5. The hoop yield stress zyx o , ~ was obtained by com- pressive uniaxial test, which is more difficult to obtain, as it requires stiff testing conditions, an accurately machined specimen, and alignment of the line of loading with the axis of the specimen. For this reason, an abnormal hoop stress result for specimen ZFV18 was discarded and substituted by the hoop stress average for all specimens. 11.6 Buckle Propagation In the unlikely event of a local buckling, the external pressure may cause a buckle to propagate along the pipeline. As long as the external pressure is less than the propagation pressure threshold, a buckle cannot propagate. A number of empirical relationships have been published for determining the minimum pressure, Pp, at which buckle propagation can occur for a given pipe diameter, wall thickness and steel grade. The mechanics of buckle propagation is explained by Nogueira (1998a,b,c). The APT RP 1111 (1999) equation for calculating the propagation pressure is as follows: Pp= 24S(T) tnom 2.4 (11.21)
- 257. W zy 0 z 01 Pipe designation Applied (live) critical strain zyxwvutsrqpo (%) zyxwvutsrqp Table 11.5 Test data and results for 26 in. OD, 1.625 in. WT, Grade X60, after Stark and McKeehan (1995) ZFV8 0.00 ZFV16 0.19 6700 71.5 80.4 26 1.620 16.05 0.13 WT (in.) ZFV21 0.2 zyxwvuts 1 6800 66.9 79.4 26 ___ _ _ 1.623 16.02 0.16 - zyxw 2 6 I 1.619 'The measured hoop strcss of specimen ZFVIII was reported to be 56.1 ksi, which was deemed low, due lo testing error. Thc average hoop strcss ol all spccirnens was used hercin. c c
- 258. Design and Construction zyxwvuts o fzyxwvuts Offshore zyxwvutsrqp Pipelines zyxwvuts 907 If the following equation is satisfied, with the buckle propagation safety factor. zy fp, of 0.80, then buckle arrestors are not required. Since, in this case a buckle cannot propagate along the pipeline. Po-P, ‘ f p . P p (11.22) In deepwaters, it is not economically feasible to have the wall thickness satisfy the buckle propagation criteria of equation (11.22). The wall thickness can be chosen to be less than the minimum calculated in equation (11.21), provided that buckle arrestors are recommended to mitigate the risk of buckle propagation. Buckle arrestors can be designed by the formulations presented by Park and Kyriakides (1997) or Langner (1999). zyxwvuts 11.7 Design Example This section provides an example of offshore pipeline design. Figure 11.8 shows schematically an example gas pipeline; corresponding design data is listed in table 11.6.The example gas pipeline runs from a subsea well at 8000ft water depth to a shallow water host platform at a water depth of 500ft. At the platform, a riser segment brings the gas to the topside piping. The pipeline is assumed to be 10.75 in. diameter and it has wall thickness (WT) break at 3000 ft water depth. 30 CFR 250 (2002) applies, and each wall thickness will be calculated, or verified, to prevent the three failure modes identified as internal pressure containment (burst) during operation and hydrotest, collapse due to external pressure, and local buckling due to bending and external pressure. Figure 11.8 Design example
- 259. 908 Pipe OD' Corrosion allowance, CA Specified minimum yield strength zyxw Table 11.6 Summarises the design data for this example 10.75in. 1/16in. 65,000 psi zyx Chapter z 11 Poisson's Ratio Pipeline content IParameter zyxwvu IValue 1 0.3 Natural gas 1Pipe material IAPI 5L X-65 seamless i Maximum Source Pressure (MSP) at well-pipeline interface Seawater density 6400 psig 64lb/ft3 IMinimum ultimate tensile strength 177,000psi I Maximum installation bending strain Maximum in-place bending strain 1Modulus of elasticity 129 x lo6 psi I 0.2% 0.2% 1Steel density 1490lb/ft3 1 1Gas density 1 14lb/ft3 I 1Gas maximum temperature I200"F I 1Gas minimum temperature I40°F I IMinimum internal pressure lo PSk I 1Maximum water depth 18000ft I 1Minimum water deDth 1 500ft 1 1 100-year ARP2 bottom current velocity I1.3ft/s I 'OD =outside diameter 'ARP =average return period 11.7.1 Preliminary Wall Thickness for Internal Pressure Containment (Burst) The first step to determine the wall thickness is to satisfy the required design pressure, Preq, at the shut-in condition. The general equation for the required design pressure is: Preq=MSP - Ppgas - Po (11.23) where MSP =maximum source pressure, which equals the wellhead pressure at the well- z Ppgas =internal gas weight (at shut-in condition) from source elevation to elevation of pipeline interface. interest.
- 260. Design zyxwvutsrqpon and Construction zyxwvutsrq of Offshore Pipelines Water depth (ft) 0 lo zyxwv 1Po(psig) 909 Ppgas (psig) Preq (psig) i 778 5622 1 I500 1222 1729 15449 1 18000 13556 ' 0 12844 1 Table 11.7gives the required design pressure, per equation (11.23), for the significant water depths assuming a seawater weight of 64 pcf. These values will be used to check that the required design pressure does not exceed the maximum allowable operating pressure (MAOP) of the pipeline system. Based on the numbers given on table 11.7, the minimum hydrostatic test pressure equals 1.25times the maximum required design pressure, or 1.25 x 5622 = 7028 psig. This assumes that the entire pipeline system will be subject to a single hydrostatic test. The nominal hydrostatic test pressure is set at a slightly higher value, as follows: The wall thickness must be selected such that equation (11.2) [given as a design inequality by equation (11.24)] and equation (11.3) [given as a design inequality by equation (1 1.25)] are satisfied at every point along the pipeline and riser system, as follows: Nominal hydrostatic test pressure, Pnom.hyd =7100 psig @ mean water level (MWL) Leading to 80% of hydrostatic test pressure = 5680 psig @ MWL (11.24) (11.25) Recall that F= construction design factor =0.6 for riser component, =0.72 for submerged component, and =0.95for hydrotest, E =longitudinaljoint factor, E = 1for API 5L seamless line pipe, and T= temperature de-rating factor, T= 1for maximum temperature of 200'F. Also recall that 30 CFR 250 defines that the maximum allowable operating pressure (MAOP) to be the least of the following: internal design pressure of the pipeline, valves, flanges and fittings, 80% of hydrostatic test pressure, MAOP of the receiving pipeline. The preliminary nominal wall thickness for each pipeline section is calculated to satisfy equations (11.24) and (11.25). The API 5L (2000) wall thickness selection is shown in table 11.8 (see also fig. 11.8). To verify that the design equations are satisfied, table 11.8 presentsthe numericalvalues ateachpertinent location alongthepipeline elevation.Note that: For each pipeline segment the controlling (higher) required design pressure, Preqr is at the shallower water depth.
- 261. 910 zyxwvutsrqpon Water depth (ft) zyxwvuts Chapter z 11 Wall Preq zyxwvu Pldl Pmax.hyd2 MAOP3 thickness (in.) ,(psig) (psig) (psig) (PSk) 500 10.719 15449 15716 3000 10.719 (4581 15716 lo 10.875 1.5622 15895 110052 15680 1 8260 5680 8260 5680 I500 10.875 I5449 15895 110052 15680 1 18000 10.625 12844 14897 17180 14897 Iz I p l d must be greater than preq [equation (11.24)] 'P,,,.hyd must be greater than Pno,,.hyd =7100 psig in this example [equation (11.25)] 'MAOP is the least of (a)0.8 x Pnom-hyd, (b) Psdr(c) internal design pressure of valves, flanges and fittings: which in this example will be assumed greater than the values shown and (d) MAOP of receiving pipelines. An API wall thickness thinner than 0.625 in., which is 0.562 in., would not be adequate since it would not satisfy equation (11.25),even though it would satisfy equation (11.24) for water depths greater than 3670ft. There is room to revise the wall thickness break to shallower water depth, thus making the 0.625 in. segment longer, which would lead to project savings. To calculate the water depth, X. at the wall thickness break, use equation (11.23) with Preq =Pid.In this example such an equation is: 6400 - X(64/144) - (8000 - X)(14/144) = 4897, which leads to X = 2088 ft. Of course the prudent offshore pipeline engineer will always give some allowance for installation tolerances, and the revised water depth for the WT break would be set at 2200 ft. However, for purposes of the ensuing numerical examples, the water depth of 3000 ft will be maintained. 11.7.2 Collapse Due to External Pressure Equations (11.5a-11.5~)yield the collapse pressure. The factored collapse pressure f o . z P, must exceed the net external pressure everywhere along the pipeline, as shown in equation (11.4). The collapse reduction factor in this example for seamless pipe isfo =0.7. The net external pressure can be determined as the hydrostatic water pressure at maximum water depth of each pipeline section assuming zero internal pressure. The net external pressures are all within the allowable collapse pressure as summarised in table 11.9. In this example the nominal wall thickness of each pipeline section is used. The authors judge that for collapse as well as local buckling, it is over-conservative to deduct the corrosion allowance in such calculations. While burst will occur at the maximum stress which occurs at the thinnest wall thickness (justifying the corrosion allowance deduc- tion on burst limit state check), collapse and local buckling involves the formation of a four-hinge collapse mechanism, with maximum bending moments at four hinges 90" apart around the pipeline cross section [see Nogueira and Lanan 2001; Timoshenko and Gere, 1961, Section 7.51. Given that the pipe mill's average wall thickness is
- 262. Design and Construction zyxwvuts o f Offshore Pipelines zyxwvuts Water depth Wall Collapse f o zyxw .P, (psig) External thickness (in.) pressure, P, (psig) pressure, Po (psig) (ft) 911 , O-500 ,0375 zyxwvu Table 11.9 Collapse pressure vs. external pressure 10113 7079 ,222 500-3000 0.719 7911 5538 1333 Table 11.10 Factored strain vs. limiting bending strain 3000-8000 Water depth (ft) 0-500 0.625 6471 4530 3556 Wall 1 Factored 1 Factored thickness installation in-place bending (in.) bending strain strain f l E l zyxwvuts (%I f i E 2 (%I 1 0.022 1 3.30 1 Critical 1 [equation (11.7)] Po - z P i - bending strain E (Yo) 0.719 ~ 0.4 ~ 0.4 1 0.168 ~ 2.22 1 0.875 0.625 1 0.4 1 0.4 1 0.550 1 0.82 1 0.4 0.4 typically 10% greater than the nominal wall thickness, around the line pipe circum- ference, the nominal wall thickness is recommended to be used in collapse and local buckling. Of course, this is a project decision that should be clearly stated within the project DBD. 11.7.3 Local Buckling Due to Bending and External Pressure Equation 11.6 gives the limiting bending strain, to avoid the local buckling limit state. By rearranging it, equation (11.7) is obtained, which yields the maximum bending strain. Therefore, the installation and in-place bending strains shall be limited per equations (11.8a) and (11.8b). In this design example, factors of safety adopted arefl =fi = 2.0. The ovality is conservatively set at 6 = 1%, which leads to g(6) = 0.833. The critical bending strains per equation (11.7) are shown in the right hand column of table 11.10 (with P,= 0) and are all greater than the factored bending strains. 11.7.4 Buckle Propagation The pipeline propagation pressure value, per API 1111 (1999), is given by equations (11.21) and (11.22). Assuming zero internal pressure the results are shown in table 11.11. From the results presented in table 11.11, buckle arrestors are required along the 0.625 in. WT segment, when the external water pressure is greater than 1352 psi. This corresponds to 3042 ft and greater water depth. Buckle arrestor design guidelines can be found in Park and Kyriakides (1997) and Langner (1999).
- 263. 912 zyxwvutsrqpon Water depth 500 3000 8000 (ft) zyxwvuts Chaptev z 11 Wall Propagation zyxwv f , . zyxw Ppy Net external Buckle thickness pressure (psig) pressure arrestor 0.875 3789 3031 222 not required 0.719 2365 1892 1333 not required 0.625 1690 1352 3556 reauired (in.) P p r (Psig) (psig) zyxw 11.8 On-Bottom Stability This section addresses stability analysis of offshore pipelines on the seabed under hydro- dynamic loads (wave and current). On-bottom stability is checked for the installation case with the pipe empty using the 1-yr return period condition and for lifetime using the 100-yr storm. Additionally, a minimum pipeline specific gravity of 1.20 during installation is desired. Hydrodynamic stability analysis involves the following steps: 1. Define environmental criteria for the 1-yr and 100-yr condition: Water depth Significant wave height ( H ) ,wave period ( T )and the angle of attack (p) Steady current velocity (U,) and angle of attack (p) Wave only particle velocity (Uw), maximum water particle velocity due to wave and current (UnJand steady current ratio (UR= U,/U,) Soil submerged weight (y), soil friction factor () or undrained shear strength (S,) Seabed slope (6) measured positive in downward loading Determine hydrodynamic coefficients: drag (CD), lift (C,) and inertia (CI).These may be adjusted for Reynolds number, Keulegan-Carpenter number, ratio of wave to steady current and embedment. Calculate hydrodynamic forces drag (FD),lift (F,) and inertia (FI). Perform static force balance at time step increments and assess stability and calculate concrete coating thickness for worst combination of lift, drag and inertial force. Hydrodynamic stability is determined using Morison’s equation, which relates hydraulic lift, drag and inertial forces to local water particle velocity and acceleration. The coefficients used, however, vary from one situation to another. For example, the lift and drag coefficients of 0.6 and 1.2, which is representative of a steady current, is not appro- priate for oscillating flow in a wave field. Additionally, these coefficients are reduced if the pipe is not fully exposed because of trenching or embedment. To determine wave particle velocity, the theory used depends on wave height, water depth and wave period. For most situations, linear theory is adequate as bottom velocities and accelerations do not vary significantly between theories. However, as the wave height to water depth ratio increases, Stoke’s fifth order theory becomes appropriate. For shallow 2. 3. 4.
- 264. Design and Construction of Offshore Pipelines zyxwvuts 913 z water or very high wave heights, a solitary theory should be used to predict particle velocity and accelerations [Sarpkaya and Isaacson, 19811. For breaking waves, or large diameter pipe that may affect the flow regime, other analysis methods may be appropriate. In general, pipelines should be trenched within the breaking wave (surf) zone. Experimental and theoretical researches [Ayers, et a1 1989; DNV RP E305, 19881 have shown that the traditional static analysis methods have been conservative in most cases. In the 1980s, two research groups developed theoretical and experimental models to assess pipe stability. Findings of these groups (American Gas Association in USA and PIPESTAB in Europe) resulted in the development of program LSTAB, which accounts for the effects of embedment. The commercially available computer program LSTAB, with the American Gas Association, is the state-of-art tool for assessing on-bottom stability of pipelines. It is comprehensive and easy to use. What follows is a summary of the most important factors for an on-bottom stability analysis and relevant references. zyxwvu 11.8.1 Soil Friction Factor The friction factor is defined as the ratio between the force required to move a section of pipe and the vertical contact force applied by the pipe on the seabed. This simplified model (Coulomb) is used to assess stability and requires an estimate of the friction factor, . Strictly speaking, the friction factor, ,depends on the type of soil, the pipe roughness, seabed slope and depth of burial; however, the pipe roughness is typically ignored. For stability analysis, a lower bound estimate for soil friction is conservatively assumed, whereas for pulling or towing analysis, an upper bound estimate would be appropriate. The following lateral friction factors [Lyons, 1973; Lambrakos, 19851 are given as a guideline for stability analysis in the absence of site-specific data: Loose sand: tan zyxwvu 4 (generally = 30") Soft clay: 0.7 Stiff clay: 0.4 Rock and gravel: 0.7 These coefficients are adequate for generalised soil types and do not include safety factors. Small-scale tests [Lyons, 19731 and offshore tests [Lambrakos, 19851 have shown that the starting friction factor in sand is about 30% less than the maximum value, which occurs after a very small displacement of the pipe builds a wedge of soil; past this point, the friction factor levels off. The values given above account for the build-up of this wedge of soil, which has been shown to take place. 11.8.2 Hydrodynamic Coefficient Selection Hydrodynamic coefficients have been the subject of numerous theoretical and experimental investigations and are often subject to controversy. Selection of CD, C, and C, are dependent on one of the following situations: Steady current only Steady current and waves Compact sand: tan 4 (generally 4 =35")
- 265. 914 zyxwvutsrqpon Chapter zy I 1 z For steady current conditions acting on a pipeline resting on the sea floor, CDx0.7 and CL zyxwvut FZ0.9. However, these coefficients are dependent on the Reynold’s number (Re = U,D/v, with v = 1.7 x ft2/s). and if more precision is warranted Jones (1976) may be consulted. For steady-current conditions, a conservative stability check may be performed by subtracting lift from the submerged weight, calculating the available friction force and verifying that the drag force is smaller than the available friction force. For waves and currents, these parameters are dependent on the Keulegan-Carpenter number (K, = U, T/D, where D =pipe outside diameter), pipe roughness and the steady current ratio. Bryndum, et a1 (1983, 1988) include guidelines in selecting these parameters. zyxwvut 11.8.3 Hydrodynamic Force Calculation The drag force, lift force and inertia force are given by the Morrison’s equations: 1 Drag force: FD = -CD~DU,IU,,,I Lift force: FL = , C L ~ D U , l 2 2 zyxwvutsr i Inertia force: FI = Crp 11.8.4 Stability Criteria The last step of the simplified on-bottom stability analysis consists in assessing stability using a simple lateral force equilibrium equation. In the following equation the symbols are as defined in Section 11.8 and W, is the pipeline submerged weight: zy p ( W,cos 6 - FL) ? ~ ( F D +FI + W,sin 6) (11.26) This formulation assumes a Coulomb friction model as described above and is over- conservative if the pipe is embedded. A preliminary conservative approach, however, is to consider no embedment. The drag force in equation (11.26) may include the effects of the angle of attack, in case that the design wave and current are not expected to be perpendicular to the pipeline alignment. The safety factor zyx (5)in equation (11.26)is desig- ned to account for uncertainties in: Actual soil friction factor Actual environmental data (wave, current) Actual particle velocity and acceleration Actual hydrodynamic coefficients Recommended safety factors are: 5= 1.05 for installation 5= 1.1 for operation 11.9 Bottom Roughness Analysis The objective of bottom roughness analysis is to identify possible free spans that exceed the maximum allowable span length that may occur during pipeline installation, hydrotest and
- 266. Design and Constmcrion zyxwvuts o fzyxwvuts Offshove Pipelines zyxwvuts Condition 915 Longitudinal stress Total Von-Mises stress zy 1 (% SMYS) (% SMYS) zyx Table 11.12 Allowable pipeline stresses Hydrotest' 95 95 I~nstallation~ 1 80 I 90 i 'Based on 30 CFR 250 stress limit during hydrotest 'ASME B31.4 and ASME B31.8 requirements 'Assumed identical to the ASME limits for the operating case operation. The bottom roughness analysis can be performed using computer software such as OFFPIPE, which is an industry recognised finite element tool used for the analysis of offshore pipelines (see www.offpipe.com for software information). The OFFPIPE model assumes a linear elastic foundation under the pipeline with supports at regular intervals. Due to this regular support interval, actual span lengths may be shorter than the calculated span lengths. One of the criteria to establish the maximum allowable span, is to limit the maximum pipeline stresses under static conditions. This is done by limiting both the total Von-Mises stress and the longitudinal stress as shown in table 11.12. In addition, the pipeline span lengths cannot exceed the maximum span lengths at which in-line vortex-induced-vibration (VIV) will occur. Therefore, the determination of the pipeline allowable span lengths must consider the following five criteria: onset of in-line VIV, onset of cross-flow VIV, maximum allowable equivalent stress, maximum allowable longitudinal stress, The spans from the first four criteria for each segment are considered when calculating the maximum allowable span, evaluating the bottom roughness analysis, or evaluating the pipeline crossing analysis. The last criterion involves performing a fatigue analysis to increase the span length due to in-line VIV as explained here. When free spans occur due to seabed irregularities or pipeline crossings along the pipeline route, the presence of bottom current may cause dynamic effects. The fluid interaction with the pipeline can cause the free span to oscillate due to vortex shedding. Two distinct forms of oscillation can be observed due to vortex shedding: in-line and cross-flow. In-line VIV occurs when the pipeline vibrates parallel to the direction of flow in a constant current. The amplitude of the in-line vibrations is typically less than 20% of the outside diameter of the pipe and is significantly smaller than (only about 10% of) the amplitudes for cross-flow vibrations. In-line VIV occurs at lower flow velocities and shorter spans than cross-flow VIV. It is the industry practice to allow span lengths to exceed the in-line VIV criteria, provided a fatigue analysis is done, that demonstrates adequate design life. fatigue life due to in-line VIV (optional criteria).
- 267. 916 zyxwvutsrqpon Chapter z I1 z DNV Guideline 14 (1998) presents a complete treatment of the subject of oscillations both in-line and cross-flow, including current and wave effects. What follows is a simplified approach for current-dominated oscillations. zyxw 11.9.1 Allowable Span Length on Current-Dominated Oscillations Several parameters are used to assess the allowable span length, for a given current velocity, that will lead to the onset of in-line VIV. For this analysis, the stability parameter z (KJ and the reduced velocity (V,) are used. The dimensionless stability parameter is calculated using equation (11.27). (11.27) where K, =stability parameter, Me=the effective mass, Me=Mp+M , +Ma, Mp=pipe mass, M , =mass of pipe contents, Ma=added mass, Ma = p7tD2/4, 6 =logarithmic decrement of structural damping, for steel pipe, 6 = 0.125 and p =mass density of the fluid around the pipe, for seawater p =2 slugs/ft3. The reduced velocity, V,, can be determined as a function of the stability parameter by: K, < 0.25,l 0.25 < K, 5 1.2, 0.188 +3.6K3- 1.6K,‘ [K, > 1.2, 2.2 V, = (11.28) The reduced velocity is then used to determine the critical frequency at which the onset of in-line VIV can occur. Calculation of the critical frequency is shown in equation (11.29). z V fcr = - VrD (11.29) where fcr =the critical frequency, and V = the current design velocity. To determine the span length at which the onset of in-line VIV can occur for the design current velocity, the natural frequency of the span is set equal to the critical frequency and solved for the corresponding span length. Equation (11.30)is used to calculate the natural frequency of the span. Equation (11.31) shows how the critical span length is calculated from the critical frequency. (11.30) (11.31) where f,=natural frequency of the span, C=the end condition constant (1.252~2 for pinned-fixed), E =the modulus of elasticity of the pipeline, I= the moment of inertia of the pipeline, L =a given span length and L,, =the critical span length. Cross-flow VIV occurs when the pipeline vibrates perpendicular to the direction of flow due to vortex shedding in a constant current. The response amplitudes for cross-flow VIV
- 268. Design zyxwvutsrqp and Construction of Offshore Pipelines zyxwvut 917 are much greater than for in-line VIV. Span lengths for the onset of cross-flow VIV are to be avoided. The parameters used to assess the potential of cross-flow VIV are the Reynolds Number, z Re,and the reduced velocity, V,. For cross-flow VIV, the reduced velocity can be estimated as a function of the Reynolds Number by: VD Re zyxwvuts = y (11.32) where Re=Reynolds Number, V = flow velocity, D =pipe outside diameter and z I)= kinematic viscosity of the fluid, for seawater zyxw u= 1.26 x ft2/s. R, < 5 x 104,5 5 x 104 < R, 5 3 x 106, c1 - C>R, +C 3 ~ ;+C 4 ~ :+c s ~ : Re > 3 x lo6, 3.87 (11.33) where V,=reduced velocity for onset of cross-flow VIV, cI =5.07148, c2= 1.61569 x c3=8.73792 x c4=2.11781 x and c5=1.89218 x Using the reduced velocity for cross-flow VIV, the critical frequency and critical span length are determined in the same manner as for in-line VIV. Note that a conventional riser along a platform (as shown in Fig.ll.8) must be designed such that in-line and cross-flow VIV does not occur. This check must consider wave and current. Clamps to the platform must be designed to avoid this critical design case in risers. Failure in risers in the Gulf of Mexico are rare, but have been reported during hurricanes, thus the extreme case combination for the environmental loads must be taken into account. zyxwvut 11.9.2 Design Example The design example (table 11.13) calculates the allowable span based on VIV criteria already described. From table 11.6, the bottom current used is 1.3 ft/s. The example assumes the pipeline is in the operating condition, i.e. the pipeline is filled with product. The allowable span lengths for both the in-linemotion and cross-flow motion are obtained. 11.10 External Corrosion Protection The external corrosion of the offshore pipelines is usually controlled by ways of an external corrosion coating and a sacrificial anode-cathodic protection system. The corrosion coating for the offshore pipelines is normally fusion-bonded epoxy (FBE) coating of about 16mil. The design of the sacrificial anode-cathodic protection system is typically performed using the design guidelines given by DNV RP B401 (1993). The surface areas to receive cathodic protection should be calculated separately for areas where the environmental conditions or the application of coatings imply different current requirements. All components to be connected to the system should be included in the surface area calculations. This may include various types of appurtenances or outfitting to be installed along the pipeline.
- 269. 918 zyxwvutsrqpon Step 1 zyxwvuts Chapter z 11 Value Unit zy Table 11.13 Allowable span design example. Segment 3 pipe: OD =10.75 in., WT =0.625 in. ~ Pipe mass, zyxwvuts Mp 2.103 Mass of contents, M , 0.214 sluglft slug/ft 1 Added mass, M , 1 1.254 1 slug/ft 1 ' Stability parameter, Ks ' 0.559 Step 3 Reynolds number, Re 92,427 1 Equivalent mass, Me 1 3.571 j slug/ft1 - - I I Step 5 Critical frequency for in-line VIV 0.853 ~ lis )Step4 I I 1 Reduced velocity, Vr, in-line 1 1.701 1 - 1 ~ Reduced velocity, Vr, cross-flow 1 4.93 1 - 1 I Critical span length for in-line VIV 1 104.4 I ft I IStep 6 ~ 1 I 1 Critical frequency for cross-flow VIV 1 0.294 I l/s I 1 Critical span length for cross-flow VIV 1 177.8 I ft I Surface area demand involves assumptions of coating breakdown factors. Offshore engineers designing pipelines in the Gulf of Mexico, typically use coating breakdown factors smaller and more realistic than those recommended by DNV RP B401 (1993). For example, see Britton (1999) who suggests initial coating breakdown factor of 3%, and final coating breakdown factor of 5% for a 20-yr design life. Thus, coating breakdown factor established for a project shall always be documented very clearly in the DBD, so that the project team consistently uses the project-specific values. 11.10.1 Current Demand Calculations The current demand I, to achieve polarisation during the initial and final lives of the cathodic protection system, and the average current demand to maintain cathodic protection throughout the design life should be calculated separately. The surface area A, to be cathodically protected should be multiplied with the relevant design current density zyxwvu i, and the coating breakdown factor fc: I, = A, . zyxwvu f,.i, (11.34) where I, =current demand for a specific surface area, i, =design current density, selected from tables 11.14 and 11.15, which follow guidance provided by DNV RP B401 (1993)
- 270. Design and Consrructiori zyxwvuts ofzyxwvutsrq Offshore zyxwvutsrqp Pipelines zyxwvuts 919 z Water depth (ft) Design current densities (initial/final) in A/ft2 Tropical Subtropical Temperate Arctic ( >20°C) (12-20°C) (7-1 2"C) ( <7 T ) 0-100 Initial Final Initial Final Initial Final Initial Final 0.0139 0.0084 0.0158 0.0102 0.1860 0.0121 0.0232 0.0158 z Table zyxwvut 11.15Average (Maintenance) design current densities for variousclimatic regions and depths - adapted from table 6.3.2 of DNV RP B401 (1993) > 100 0.0121 0.0074 0.0139 0.0084 0.0167 0.0102 0.0204 0.0121 10-100 10.0065 10.0074 10.0093 '0.011 1 Water depth (ft) I > 100 10.0056 10.0065 10.0074 10.0093 1 Design current densities (initial/final) in A/ft2 'Tropical Subtropical Temperate IArctic (>2O"C) (12-20°C) (7-12°C) '(<7"C) Section 11.3,fc =coating breakdown factors. See DNV RP B401 (1993) table 11.4.1 and Sections 6.5.3 and 6.5.4 for guidance on offshore pipelines. For items with major surface areas of bare metal, the current demands required for initial polarisation, zyxwvu Z ,(initial), and for re-polarisation at the end of the design life, Z ,(final), should be calculated, together with the average current demand I, (average) required to maintain cathodic protection throughout the design period. For pipelines and other items with high-quality coatings, the initial current demand can be deleted in the design calculations. 11.10.2 Selection of Anode Type and Dimensions The type of anode to be used is largely dependent on fabrication, installation and operational parameters. The anode type is determining for which anode resistance formulas and anode utilisation factors are used in further calculations. For pipeline bracelet anodes that are mounted flush with the coating, the thickness of the coating layer will be decisive to the anode dimensions. 11.10.3 Anode Mass Calculations The total net anode mass M (kg) required to maintain cathodic protection throughout the design life t, (yr) should be calculated from the average current demand I,: IC (average) t, .8760 u . ELT M = (11.35)
- 271. 920 zyxwvutsrqpo Chapter z 1 z I where 8760 is the number of hours per year, u is the utilisation factor, and zy ELT (A-h:lb) is the electrochemical efficiency of the anode material, which is 950 A hjlb for aluminum- based anode material type zyxwvu - see Section 6.6, DNV RP B401 (1993). zyx 11.10.4 Calculation of Number of Anodes For the anode type selected, the number of anodes, anode dimensions and anode net mass should be selected to meet the requirements for initial/final current output (A) and the current capacity ( A.h), which relate to the protection current demand of the protection object. The anode current output I, is calculated from Ohm's law: E,"- E," I, = ~ (11.36) Ra whereEt (V) is the design closed circuit potential of the anode, typically - 1.05V (relative to Ag/AgCl/seawater), see Section 6.65 of DNV RP B401 (1993). E: (V) is the design protective potential, which is chosen to be -0.80 V (relative to Ag/AgCl/seawater). z R, (ohm) is the anode resistance, is given by DNV RP B401 (1993), table 6.7.1; which for bracelet anode is: R = 0.315p/Z/;i (11.37) where A is the anode surface area, and p is the environmental resistivity; for which guidance can be found in Section 6.8 of DNV RP B401 (1993). For the Gulf of Mexico, typically, p = 30ohm-cm. Anode dimensions and net weight are to be selected to match all requirements for current output (initial/final) and current capacity for a specific number of anodes. This is an iterative process and a simple computer spreadsheet may be helpful. Calculations should be carried out to demonstrate that the following requirements are met: C, = n .c , I, .tr .a760 (11.38) (11.39) n .I, (initial/final) 2 I, (initial/final) To summarise, the cathodic protection design should optimise anode spacing and weight. The selected anode characteristics must meet two requirements: The anode mass must be sufficient to meet the current demand over its design life. The anode surface area at the end of its design life must be sufficient to provide the required current. At the end of its design life, the anode's surface area is assumed to be the product of the pipe circumference and the anode's length. 11.10.5 Design Example The following shows an example of the pipeline cathodic protection design. The data is from the deep segment (assumed 100,000 ft long) of the design example shown in Section 11.7. The rows on table 11.16are numbered so that the calculation may be easily followed, thus C1 in column B, row 3, refers to the numerical value shown in column C, row 1.Cross reference to the above said equations is also provided. Next Page
- 272. Design and Construction of zyxwvutsrq Offshore Pipelines zyxwvut 921 z 11.11 Pipeline Crossing Design Pipeline crossing design basically involved protecting the crossed pipeline using articulated concrete mattresses. Typically, in the US Gulf of Mexico (GOM), two 9 in. thick articulated concrete mats are used, for a 18 in. separation between the pipelines. The single lift equations (11.41)-(11.43) below [see Troitsky, 1982; Section 11.6.51 can be used to calculate the crossing loads, which were relatively small, so that the crossed pipeline could transfer such forces to the underlying seafloor. Typically, no intermediate supports for the crossing pipeline were needed. However, recent GOM Federal regulations require that crossing pipelines be covered by mats from touchdown to touchdown, for water depths less than 500ft. This leads to a crossing arrangement depicted in figs. 11.9 and 11.10, which shows with relative scale a 12 in. pipeline crossed by a 24 in. pipeline and 9 in. thick concrete mats. The capping mats impose a load on top of the crossing pipeline, which is transferred to the crossed pipeline as a concentrated load on a short ring of pipe (shown with a length zy L in fig. 11.10). Finally, the crossed pipeline transfers this load to the seafloor. Thus, crossing design needs to be evaluated as follows. The first step in the crossing analysis is to estimate the crossing load. Concrete mattresses' submerged weight is approximately 6000 Ib for a 9 in. thick mat with 8 ft by 20 ft dimensions, which leads to a submerged load of about zyxw M . 'zyxwvuts =38 psf. A 4.5 in. thick mat of same dimensions has a 3600 lb total submerged weight, for a load of about zyx M, =23 psf. The load imposed by the capping mattress is estimated by assuming an average drape angle of 30" (fig. 11.lo), and the corresponding maximum linear load, q, at the crown of the crossing is given by: 4 = W[2zyxwvu X 1.16(d zyxwvu 4 -ODTOP) +oD~op] (11.40) where 1.16 l/cos30, zyxwvu d is the crossing pipe prop height, which equals distance from adjacent seafloor (mudline) to bottom of crossing pipe, ODTop is the diameter of the crossing pipe. Given an existing pipeline with ODBOT=20.00in., WTBoT=0.812, embedded 3 in, two 9 in. separation mattresses, a crossing pipeline with 0 D ~ o p = 12.75 in., WTTop=0.750, and a 9 in. capping mattress; then: d=35 in. and q-391.2 plf. The crossing pipeline has water filled submerged weight of 83.7plf (= 96.2plf steel weight in air, 44.2plf water contents at 64pcf, minus 56.7plf buoyancy. When the mattress load is added, the total maximum pipeline load is qr= 475 plf, or 5.7 times that of the crossing pipeline during hydrotest. A pipeline on a prop with height d from the adjacent seafloor (see fig. 11.1l), with Young's modulus E, moment of inertia I and total submerged load q, will have a distance from the prop point to touchdown 1 given by: 72EId 4 1 =- (11.41) The total prop force Fp and maximum bending moment at the centre of the span, Mp, are given by: F p = 44113 (11.42) Mp = 4l2/6 (11.43) Previous Page
- 273. W zy N N zy Table 11.16 Cathodic protection design example Row number Source/Equation IA IB 1 IPipeline diameter IInput data 2 IPipeline length 1Input data 3 ITotal surfacc area In*Cl *c2/12 4 1Mean coating breakdown factor 1Project specific zyxwvut 5 1Final coating breakdown factor IProject specific zyxwvut 6 1Mean bare area Ic4*c3 zyxwvut 7 IFinal bare area Ic5*c3 8 1Required current density 1Project specific 9 1Required mean current 1C6*C8 IO IRequired final current 1C7*C8 I 1 IElectrochemical efficiency IInput data 12 (DesignLife IInput data 13 IAnode efficiency IInput data _ _ ~ Value C 10.75 ~~ 100,000 281,434 Isq ft 5 1% 9 IY" 25,329 Isq ft 0.00837 IA/ft2 % + 212 950 1A-h/lb 20 IY r 0.85 1
- 274. 14 118zyxwvutsrqpon 1Nominal nct anode weight IProject specified I130 IIbs 15 Weight required 16 Eq. 11.35: C9*C13* X760/(C10*C14) 17 Anode specificweight Anode thickness Average anode diamcter Estimated anode length Environmcntal resistivity Anode Area Anode resistance - -~ 180 lb/ft3 -~ Vendor specified Project defined I .5 in. zyx C1+c22 12.25 in. 1728*C19/(C2 zyxwvu 1*7c*C23*C2) 22 in. Project specified 30 ohm-cm n*C2I *c22 847 in.2 Equation (1 1.37): 0.1278 ohm -~ - ~ ~~ . . . 0.315*(C24/2.54)/C250.~ IINT(CyC16) +1 1.956 408 Anode current - --+rrent available Minimum anode weight ] C15/c17 A A 480 209 122.5 Ibs 19 20 . 21 22 23 24 25 126 IAnodc potential 1Project specified 1-1.05 IV 127 ICathode ootential IProiect wecificd 1-0.80 Iv W N w
- 275. 924 zyxwvutsrqpo Chapter z 11 Figure 11.9 Schematic of crossing arrangement zyxw - Side view of crossing pipeline Figure 11.10 Schematic of crossing arrangement- Side view of crossed pipeline zy In order to estimate the crossing load and make an initial assessment of the crossing integrity, a conservative analysis may be done as follows: Assume that the maximum load z qr is applied on the crossing pipeline along the entire crossing span. Calculate the prop force Fp as a function of the prop height d, for several values of the crossed pipeline additional embedment, A,. A simplified yet conservative reactive force can be calculated by assuming a soil reaction acting on the entire crossed pipeline outside diameter, thus leading to a soil reaction of qs=3.4 S , ODsoT. where S, is the undrained shear strength, and the
- 276. Design and Construction of zyxwvuts Offshore Pipelines zyxwvuts 925 Figure 11.11 Schematics of a propped pipeline Figure 11.12 Propping and reactive load for a pipeline crossing zyx factor 3.4 accounts for the round pipeline shape as a foundation. With this linear soil reaction, the same equations of pipeline on a prop can be used for the crossed pipeline to calculate the total reaction provided by the soil, FR, as a function of additional embedment z A,. Both forces Fp and FR can be plotted as a function of the additional embedment and their intersection will provide the crossing load and corresponding lengths to touchdown for each pipe. Such plots are shown in fig. 11.12 for the pipeline with the characteristics given in the example, and assuming that the crossed pipeline has an initial embedment (before crossing installation) of 3 in. In this case, the resultant crossing load is Fp zyxwv =FR=50.4 kip, at an additional embedment of 4.1 in. (for a total embedment of 7.1 in.). The constant soil reaction on the crossed pipeline is qR = 567 plf, as a result of the undrained shear strength value of 100 psf adopted. The prudent offshore engineer needs to adopt an upper bound for the shear strength, since this will lead to higher crossing loads. The distances to touchdown for the crossing (top) and crossed (bottom) pipeline are lrop =79.5ft and lBoT=66.6 ft. Therefore, the bending
- 277. 926 zyxwvutsrqpo Chapter z 11 moments at the crossing point applied at each pipelines are MTop zy =500.4 kip-ft and MBOT=418.9 kip-ft. Note that, if no capping mattresses were present, the crossing load would be about 14kip, leading to ZTOP = 127ft, and a smaller bending moment MTop=225 kip-ft would be present. Of course, the assumption of constant maximum load on the crossing pipeline results in an estimated crossing force higher than actual. Similarly, the maximum soil reaction acting along the entire crossed pipeline span results in the estimated reactive force higher than actual. A more precise finite element analysis considering the soil as hyperbolic non-linear springs and varying the applied load on the crossing pipeline may be performed using a finite element program, if warranted, thus leading to somewhat smaller crossing loads. With the applied crossing load and resultant bending moments, a checking against pipe capacity must be performed. Two checks are required: a longitudinal bending moment check as well as local collapse check (e.g. a ring of pipe being crushed, or excessive ovalised, at the crossing point). zyxwvu A limit state design is proposed where each failure mechanism is checked against the corresponding limit state. A factor of safety of 1.5 is suggested. The plastic bending moment capacity, Mp,is given by Mp= S y zy (D- zy t)2t.Adopting Sy= 65 ksi for both pipelines in the example above, Mp.~op=585 kip-ft and M~.BoT= 1619 kip-ft. Therefore, the crossing pipeline has a factor of safety, FS =585/500.4= 1.17, and does not meet the safety criteria proposed above. The crossed pipeline has a factor of safety, FS = 1619/418.9= 3.9 and it is adequate. For local collapse to occur, a three-hinge collapse mechanism must take place. This is shown schematically in fig. 11.13,where a free-body of a pipe ring at the crossing location is depicted. The effective ring of pipe has length L (see also fig. 11.10). The total soil reaction acting along the effective ring is 3.4 S, OD L, which is less than the crossing load. The shear at each end of the effectivering (shown in fig. 11.12as Vat each side of the ring) provides the force necessary for equilibrium. Collapse will occur when the total applied moment equals the total plastic capacity of the upper half of the ring [Baker and Heyman, 19691.The total applied moment on the upper half of a ring due to a load F a t 12 o'clock is = F OD/4 M, = M ; Figure 11.13 Free body of crossed pipeline at crossing point
- 278. Design and Construction of zyxwvuts Offshore zyxwvutsrqp Pipelines zyxwvuts 921 z M A = F 0 D / 4 (similar to a centred point load in a simply-supported beam). The total plastic moment capacity, Mc, is the summation of the plastic capacity of the hinges at 3 o’clock and 12 o’clock, which is Mc zyxwv =L&Syt2/4 [see Nogueira and Lanan, 2001 for derivation details]. The effective ring length is assumed to be equal to the pipeline diameter, L =OD; while this assumption is adequate for pipes with D/t < 25, it needs to be validated for thinner pipes. The inequality MA5 Mc leads to the local ring collapse, or denting load, FD, on a pipeline, as given in equation (11.44): Fo = 1.73S,t2 (11.44) The data of the example above leads to, for both pipes, FD-TOP = 1.73(65)(0.752)=63.3 kip and FD-BOT= 1.73(65)(0.8122)=74.1 kip. Therefore the crossing pipeline has a factor of safety, FS =63.3/50.4 = 1.26 and does not meet the safety criteria against local collapse. The crossed pipeline has a factor of safety, FS =74.1/50.4= 1.47,which also does not meet the safety criteria. Regarding the ring collapse limit state, changing the capping mats to 4.5 in. thick (which leads to a crossing load of FL=38.2 kip), would lead to a crossing design within the safety guidelines suggested herein. In this example, extra supports adjacent to the crossed pipeline are needed for the safety criteria suggested herein to be achieved. Note that both pipelines must be checked, since the free body shown in fig. 11.13 also applies to the crossing pipeline; except that the point load would be inverted: the higher force would be applied at 6 o’clock (the crossed pipeline reaction) and the smaller force at the 12 o’clock is due to the pipeline self-weight and capping mats. The above equations neglect the effects of external pressure. While this is an adequate assumption for water depths less than 500ft, such effects can be readily addressed by using the rational model for pipeline collapse. This way, crossing capacity in deep water can be more precisely estimated. Nogueira and Lanan (2001) showed that external pressure has the effect of adding to the total applied moments and also decreasing the ring collapse hinge capacity due to additional compressive hoop stress around the pipe. Thus, advantage can be taken of the terms given in equation (24) of Nogueira and Lanan (2001) to add the effects of external water pressure. The example above illustrates the fact that a new US GOM regulation, while with the goal of decreasing the risk of pipeline being dragged or damaged by fishing gear, has the effect of increasing substantially the stresses on pipeline crossings. Therefore, the offshore pipeline engineer needs to be aware that what used to be a traditionally trivial design matter, now requires renewed attention. Actually, any change on status quo in any area of engineering always needs to be carefully considered by knowledgeable and careful engineers, to assess all implications. zyxwv 11.12 Construction Feasibility Pipelines are installed on the seafloor by one of the four typical installation methods: J-lay, S-lay, Reel-lay and Tow. The J-lay and the S-lay method are shown schematically in figs. 11.14 and 11.15 (the shape each pipe assumes justifies the corresponding name). The reel-lay method includes one or more pipe spools on board the vessel, and the pipeline is
- 279. 928 zyxwvutsrqpo Chapter z 11 I zy SAGBEND REGION zyxwvutsr Figure 11.14 Schematic depiction of the J-lay installation method OVEREEND REGION Departure angle SLAY zyx METHOD SAGBEND REGION Figure 11.15 Schematic depiction of the S-lay installation method un-spooled during offshore works. It departs the vessel in a J-lay or S-lay configuration, depending on the vessel method employed. By J-lay mode it is meant a large departure angle, thus the J-lay tower can assume a large departure angle to the horizontal, leading the pipe to a single curvature, or J-shape. Conversely, the S-lay mode has a smaller departure angle and the pipe has a double curvature, or S-shape.
- 280. Design and Construction of zyxwvuts Offshove Pipelines zyxwvut 929 With the exception of the Tow method, all others use a self-contained laybarge to store pipe (with additional supply barges, as required). Some laybarges use anchored mooring system to mantain position, such as the Castor0 10 (at this time, owned by Saipem); others use thrusters in the dynamically positioned station-keeping mode. Station-keeping is very important during pipelay, since unexpected movement away from the planned laying route may severely bend the pipeline either in a sagbend or in an overbend, and the pipe may buckle or kink. The Allseas S-lay barge Lorelay (fig. 11.17) was first to apply dynamic positioning system to pipelay. The McDermott DB50 (J-lay) is also dynamically positioned. Both vessels use an integrated control system, which tracks the relative position of the touchdown point and the vessel. At the time of this writing, the Canyon Express flowlines in the Gulf of Mexico achieved the world's deepest pipeline installation. which took place in the Summer 2002 at a maximum water depth of 7300 ft or 2225 m [de Reals, et a1 2003; Nogueira and Stearns, 20031. This successful project consisted of 110 miles of 12 in. pipe, with a number of in-line structures, which had to land at precise locations on the seabed, with tight tolerances. Previously, the Blue Stream pipeline at 7050 ft (2150m) water depth was installed across the Black Sea [McKeehan and Kashunin, 19991consisting of about 390 km of 24 in. pipe. The installation contractor, Saipem, used the J-lay vessels during the installation of these projects: the Canyon Express project was installed with the vessel FDS and the Blue Stream project used the semi-submersible S-7000. The Horn Mountain 10-in. pipeline has been installed by Allseas, using the S-lay method with the vessel Solitaire, during the Winter 2002 in the Gulf of Mexico at a water depth of about 5500 ft. Allseas was able to achieve an impressive maximum lay rate of 5.6 miles in one day. The interested reader may consult Langner (2000) for a more comprehensive description of recent projects installed in the Gulf of Mexico. For the offshore pipeline engineer. it is interesting to know the availability of the vessel fleet, as well as its pipe storage capacity and lay rates. This information is important to help establish the potential cost of a project and, therefore, its feasibility. Table 11.17 lists all the major pipelay contractors and their addresses in Houston, Texas. This table will allow information to be obtained directly from the contractors, who frequently are upgrading their vessel fleet. For example, recent additions are the lay barges Deep Blue (Coflexip Stena reel ship) and the 4-4000 (Cal Dive). zyxw 11.12.1 zyxwvuts J -lay Installation Method The J-lay installation method is a relatively new type of installation method specifically aimed at deepwater and ultra-deepwater projects. This method is characterised by a steep ramp, typically 65" or higher departure angle, so that the pipe has a suspended J-shape. While fig. 11.14 depicts this schematically, fig. 11.16 shows the Balder J-laying pipe with the aid of a side tower. The stresses and strain close to the top are minimised, as well as the horizontal tension component at the top and the horizontal tension at the mudline [Langner and Ayers, 19851. The main advantages and disadvantages of the J-lay method are described in table 11.18. Typically, to assess the technical feasibility, analysis is performed using commercially available software packages, such as OFFPIPE. Alternatively, a simplified analysis may
- 281. 930 zyxwvutsrqpon ~ zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA Cofexip Stena Offshore zyxwvu www.technip-coflexip.com Chapter 11 7660 Woodway, Suite 390 Houston, Texas 77063 713-789-8540 zyxw Table 11.17 Major offshore pipeline installation contractors in Houston, Texas ~ Installation Contractor and website 1Address in Houston, Texas, USA ~ DSND Horizon www.dsnd.com www.subsea7.com 2500 City West Blvd Suite 300 Houston, Texas 77042 713-267-2246 Saipem Inc. www.saipem.it 15950 Park Row Houston TX 77084 281-552-5706 Stolt Comex Seaway www,stoltoffshore.com 900 Town & Country Lane Suite 400 Houston, Texas 77024 1 11911 FM 529 Houston Texas 77041 713-329-4500 CalDive International ~ www.caldive.com I ! 400 N. Sam Houston Parkway E. Suite 400 Houston, Texas 77060 281-618-0400 1 17154 Butte Creek, Suite 200 Houston, Texas 77090 281-880-1600 1 Allseas www.allseas.com 333 N. Sam Houston Pkwy. E. Suite 750 Houston, Texas 77060 281-999-3330 'J. Raj McDermott, Inc. 1www.jraymcdermott.com 200 WestLake Park Blvd. Houston, TX 77079-2663 281-870-5235 Global Industries ~ www.globalind.com ~ 5151 San Felipe, Suite 900 Houston, Texas 77056 713-479-7911 Torch Offshore Inc www.torchinc.com 11757 Katy Freeway, Suite 1300 Houston, Texas 77079 713-781-7990 be performed using the stiffened catenary equations, which can yield very accurate results for the J-lay configuration [Langner, 19841. Such analysis will provide top tension, bottom tension and pipeline stresses and strains along the suspended catenary. With these parameters, the pipeline wall thickness can be checked, as well as the required vessel forces.
- 282. Design and Construction of Offsshort. Pipelines zyxwvuts 931 z Figure 11.16 Heerema’s balder in J-lay mode zyxwv - Courtesy Heerema marine contractors Table 11.18 Advantages and disadvantages of J-lay 1Adv. !Adv. 1Adv. 1Adv. IAdv. Best suited for ultra deep-water pipeline installation. i Suited for all diameters. Smallest bottom tension of all methods, which leads to the smallest route radius, and allows more flexibility for route layout. This may be important in congested areas. Touchdown point is relatively closer to the vessel, thus easier to monitor and position. Can typically handle in-line appurtenances with relative ease, with respect to landing on the seafloor, but within the constraints of the J-lay tower. Some vessels require the use of J-lay collars to hold the pipe. If shallower water pipeline installation is required in the same route, the J-lay tower must be lowered to a less steep angle. Even then, depending on the water depth, it may be not feasible to J-lay the shallow end with a particular vessel and a dual (J-lay/S-lay) installation may be required. Such was the case of the Canyon Express project [de Reals, et a1 20031.
- 283. 932 Adv. All welds are done on horizontal position, making for efficient productivity of multiple stations (typically 5-6). zyx Figure 11.17 Allseas Lorelay S-lay vessel zyxw - Courtesy C. Langner zyx Chapter zy I 1 11.12.2 S-lay S-lay is utilised to install the vast majority of all offshore pipelines. Allseas have configured its flagship, the Solitaire, with a stinger that can reach very steep departure angles. As a result, it was able to install, a 10-in. pipeline at 5400ft water depth. S-lay is a very efficient lay method, since all welding is done with pipe in an horizontal position. The main advantages and disadvantages of the S-lay method are presented in table 11.19. Table 11.19 Advantages and disadvantages of S-lay Can typically handle smaller, more compact in-line appurtenances with ease, but larger in-line structures may be too large to go through the stinger. 1Disadv. Buckle arrestors will induce concentrated higher strains in their vicinity ,within the stinger. Disadv. Typically, pipeline will rotate axially during installation [Endal and Verley, 2000; Endal, et a1 19951. 1Disadv. IRequires a relatively high component of horizontal tension. I
- 284. Design and Construction of Offshore Pipelines zyxwvuts 'Disadv. Disadv. 933 Very high pipeline strains (of the order of 3-5%) are applied into the pipeline. Due to high strains, welding methods and acceptance criteria are more z Table 11.20 Advantages and disadvantages of reel-lay Disadv. zyxwvuts 1 Adv. IAlmost all welds are done on-shore. minimising offshore welding. I In-line structures are typically more difficult to handle and install. I Well suited for smaller diameter lines and smaller diameter-to-thickness ratios. can stored on-board, a very fast installation campaign is1 achieved, making this method very cost effective. 1 Disadv. If the route is too long or the diameter is relatively large, all the pipes may not be able to be stored on-board and a number of recharging trips to the spooling base may be necessary to re-load, thus offsetting the high lay rate. 1Disadv. IPipeline will rotate during installation and may coil on the seafloor. I 11.12.3 Reel-lay The reel method was patented in the USA by Gurtler (1968), who makes reference to a British Patent of 1948. The patent [Gurtler, 19681 has very detailed drawings of a horizontal reel, as well as a pre-bending apparatus and straightener. The main advantages and disadvantages of the reel-lay method are described in table 11.20. 11.12.4 Towed Pipelines In this installation method, the pipeline is constructed onshore and towed into place. There are different ways to tow the pipeline string to site: surface tow, mid-depth tow or bottom tow. In the surface tow the pipe is positively buoyant, towed to location on the surface, and sunk in position by flooding. Wave action is a factor; therefore this method is used typically where rough seas are not likely. In the mid-depth tow typically the pipe or pipe bundle is negatively buoyant, suspended above the seabed and towed by a lead tug with a tail tug at the end of the pipe string. If the pipe is positively buoyant, mid-depth tow may be achieved by incorporating the use of drag chains at specified intervals along the pipe string, so that the pipe string assumes an equilibrium position above the seabed. For the bottom tow method, the pipeline rests on the seabed, and a tug pulls it. The length of the towed string is limited to about ten miles in the most favourable conditions. The tow methods are challenging due to the effects of the environment such as waves action, oscillations during pull or abrasive effects of the seabed during bottom tow. However, the onshore construction may significantly reduce cost when compared to the installation methods described in the previous sections. Several failures of pipe bundles during tow attest to the precautions that the offshore pipeline engineer must take when using the tow method of installation.
- 285. 934 zyxwvutsrqpo Chapter z 11z References American Petroleum Institute zyxwv - Recommended Practice 1111 (1999). “Design, construc- tion, operation, and maintenance of offshore hydrocarbon pipelines (Limit State Design)”, (3rded.). American Petroleum Institute - Specification 5L (2000). “Specification for line pipe”, (42”ded.). American Society of Mechanical Engineers - B31.4 (2002). “Pipeline transportation systems for liquid hydrocarbons and other liquids”. American Society of Mechanical Engineers B31.8 (1999). “Gas transmission and distribution piping systems”. Ayers, R. R., et a1 (1989). “Submarine on-bottom stability: recent AGA research”, Eighth Znt. Conf. on Off: Mech. and Arctic Eng., March 19-23, The Hague. Baker, J. and Heyman, J. (1969). “Plastic design of frames”, Cambridge University Press, Cambridge, UK. Berry, W. H. (1968). “Pipelines from the North Sea block 49/26 to the Norfolk coast”. Journal o f Petroleum Inst., Vol. 54, No. 532, pp. 104106. Britton, J. (1999). “External corrosion control and inspection of deepwater pipelines”, Deepwater Pipeline Tech. Conf., organized by Pipes and Pipelines Int., New Orleans, Louisiana. Broussard, D. E., Barry, D. W., Kinzbach, R. B., and Kerschner, S. G. (1969). “Pipe laying barge with adjustable pipe discharge ramp”. U.S. Patent no. 3,438,213. Bryndum, M. B. et a1 (1983). “Hydrodynamic forces from wave and current loads on marine pipelines”, Off: Tech. Conf., Paper 4454, Houston, Texas. Bryndum, M. B., Jacobsen, V., and Tsahalis, D. T. (1988). “Hydrodynamic forces on pipelines: model tests”, Seventh Znt. Con$ on Off. Mech. and Arctic Eng. Choate, T. G. A,, Davis, H., and Gaber, M. (2002). Mediterranean zyx Ofl. Conf,Alexandria, Egypt. Code of Federal Regulations, Title 30, Part 250, Subpart J (2002). “Part 250 - oil and gas and sulphur operations in the outer continental shelf, subpart J - pipelines and pipeline rights-of-way”. 7-01-02 Ed., U.S. Gov. Printing Office, Washington, D.C. Code of Federal Regulations, Title 49, Part 192, Subpart A (2002). “Part 192 - transportation of natural and other gas by pipeline: minimum federal safety standards, subpart A - General”. 10-01-02Ed., U.S. Gov. Printing Office, Washington, D.C. Code of Federal Regulations, Title 49, Part 195, Subpart A (2002). “Part 195 - transportation of hazardous liquids by pipeline, subpart A - General”. 10-01-02Ed., U.S. Gov. Printing Office, Washington, D.C. DNV Guideline 14 (1998). “Free spanning pipelines”, Det Norske Veritas, Norway.
- 286. Design and Construction zyxwvuts of Offshore Pipelines zyxwvuts 935 DNV RP B401 (1993). “Cathodic protection design”, Det Norske Veritas, Norway. Det Norske Veritas RP E305 (1988). “On-bottom stability design of submarine pipelines”, Det Norske Veritas, Norway. Det Norske Veritas OS FlOl (2000). “Submarine pipeline systems”, Det Norske Veritas, Norway. Dixon, D. A. and Rutledge, D. R. (1968). “Stiffened catenary calculations in pipeline laying problem”. Journal of Engineeringfor Znd., Vol. 90, pp. 153-170. Endal, G., Ness, 0. B., Verley, R., Holthe, K., and Remseth, S. (1995). “Behavior of offshore pipelines subjected to residual curvature during laying”, zyx 14‘h Znt. Conf. of Off. Mech. and Arctic Eng. Endal, G. and Verley, R. (2000). “Cyclic roll of large diameter pipeline during laying”, 19‘hInt. Conf. of Off: Mech. and Arctic Eng., New Orleans, LA, USA. Fowler, J. R. (1990). “Large scale collapse testing”, Proceedings o f Collapse of Offshore Pipelines, Pipeline Research Committee zyxwv - American Gas Association, Houston, Texas. Gurtler, H. (1968). Method of Laying Pipeline, U.S. Patent Office, Patent number 3,372,461, patented March 12, 1968,New Orleans, LA, USA. Jones, W. T. (1976). “On-bottom pipeline stability in steady water currents”, Off. Tech. Conf., Paper 2598, Houston, Texas. Lambrakos, K. F. (1985). “Marine pipeline soil friction coefficients from in-situ testing”, Ocean Engineering, Vol. 12, No. 2, pp. 131-150. Lanan, G. A., Ennis, J. 0.S., Egger, P. S., and Yockey, K. E. (2001). “Northstar offshore Arctic pipeline design and construction”, Off. Tech. Conf., Paper 13133,Houston, Texas. Lanan, G. A., Nogueira, A. C, McShane, B. M., and Ennis, J. 0. (2000). “Northstar development project pipeline description and environmental loadings”, ASME, Znt. Pipeline Conf., Calgary, Canada. Langner, C. G. (1984). “Relationships for deepwater suspended pipe spans”, Third Int. Conf. on Off. Mech. and Arctic Eng., New Orleans, LA. Langner, C. G. (1999). “Buckle arrestors for deepwater pipelines”, Off. Tech. Conf., Paper 10711, Houston, Texas. Langner C. G. (2000). “Technical challenges for deepwater pipelines in the Gulf of Mexico - update 2000”, Marine Pipeline Engineering Course, Houston, Texas. Langner, C. G. and Ayers, R. R. (1985). “The feasibility of laying pipelines in deep water”, Fourth Int. Conf. on Off. Mech. and Arctic Eng. Lyons, C. G. (1973). “Soil resistance to lateral sliding of marine pipelines”, Off. Tech. Conf., Paper 1876,Houston, Texas. McAllister, E. W. (1993). “Pipe line rules of thumb handbook”, (3rded.), Gulf Publishing Co., Houston, Texas.
- 287. 936 zyxwvutsrqpo Chapter z 1 z I McKeehan, D. S. and Kashunin, K. A. (1999). “The blue stream project-A large diameter deepwater pipeline in the Black Sea”, Proc. of the Non-governmental Ecological Vernadsky Foundation, the Black Sea Regional Energy Centre, April 1999. Murphey, C. and Langner, C. (1985). “Ultimate pipe strength under bending, collapse and fatigue”, zyxwvuts 4lh Int. Conf. Off. Mech. and Arctic Eng. Nogueira, A. C. (1998a). “A new model for understanding buckle propagation: Link- beam”, Deepwater Pipeline Tech. Conf., organized by Pipes and Pipelines Int., New Orleans, LA. Nogueira, A. C. (1998b). “Link-beam model for pipeline buckle propagation”, zy Off. Tech. Conf., Paper 8673, Houston, Texas. Nogueira, A. C. (19984. “Link-beam model for dynamic buckle propagation in pipelines”, Eighth Int. Ofl.And Polar Eng. Conf. (ISOPE), Montreal, Canada. Nogueira, A. C., Lanan, G. A,, Even, T. M., Fowler, J. R., and Hormberg, B. A. (2000). “Northstar development: Limit state design and experimental program”, Int. Pipeline Conf., Calgary, Canada. Nogueira, A. C. and Lanan, G. A. (2000). “Rational modeling of ultimate pipe strength under bending and external pressure”, Int. Pipeline Conf., Calgary, Canada. Nogueira, A. C. and Lanan, G. A. (2001). “Application of a rational model for collapse of deepwater pipelines”, 4‘h Deepwater Pipeline and Riser Tech. Conf., Huston, Texas. Nogueira, A. C. and Stearns, J. P. (2003). “World record breaking: design and installation of the Canyon Express flowlines 7,300 foot deep”, Off.Pipeline Tech. Conf., Houston, Texas. Palmer, A. C. (1994). “Deepwater pipelines: improving the state-of-the art”, O f f .Tech. Con$, Paper 7541. Houston, Texas. Park, T. D. and Kyriakides, zyxwvu S. (1997). “On the performance of integral buckle arrestors for offshore pipelines”, International Journal of Mech. Sciences, Vol. 29, No. 6. de Reals, T. B., Lomenech, H., Nogueira, A. C., and Stearns,J. P. (2003). “Canyon express flowline system: design and installation”, Off. Tech. Conf., Paper 15096, Houston, Texas. Sarpkaya, T. and Isaacson, M. (1981). Mechanics of Wave Forces on Offshore Structures, Van Nostrand Reinhold Company, New York. Stark, P. R. and McKeehan, D. S. (1995). “Hydrostatic collapse research in support of the Oman India gas pipeline”, Off.Tech. Cony., Paper 7705, Houston, Texas. Tam, C., Raven, P., Robinson, R., Stensgaard, T., Al-Sharif, A. M., and Preston, R. (1996). “Oman India pipeline: development of design methods for hydrostatic collapse in deep water”, Off. Pipeline Tech. Conf., Amsterdam. Timmermans,W. J. (2000). “The past and future of offshore pipelines”. Off. Pipeline Tech. Conf., Oslo, Norway. Timoshenko and Gere (1961). Theory of Elastic Stability, McGraw-Hill, New York.
- 288. Design and Constrrrcrion zyxwvutsrq of Offshore Pipelines 931 z Troitsky, M. S. (1982). “Tubular steel structures -theory and design”, (2nded.).The James F. Lincolin Are Welding Foundation, Cleveland, Ohio., Wallace, B. K., Gudimetla, R., Nelson, S., and Hassold, T. A. (2003). “Canyon express subsea multiphase flow metering system: principles and experience”, Off. Tech. Con$, Paper 15098, Houston, Texas.
- 289. Handbook of Offshore Engineering zyxwvuts S. Chakrabarti (Ed.) zyxwvutsrq C 2005 Elsevier Ltd. All rights reserved 939 Chapter 12 Design for Reliability: Human and Organisational Factors Robert G. Bea zyxwvuts University of California,Berkeley, CA 12.1 Introduction Very advanced technology has been developed to assist offshore engineers in the design of platforms, floating structures, pipelines and ships. Those who have used and are using this technology have much to be proud of. Today there is a vast infrastructure of these structures located on the world’s continental shelves and slopes. In the main, this infrastructure has had a remarkable record of success. This chapter is about a part of this technology that is focused on people and their organisations and how to design offshore structures to achieve desirable reliability. The objective of this chapter is to provide the engineer zyxwvuts - designer-oriented guidelines to help reap success in the design of offshore structures. The application of these guidelines is illustrated with two examples: (1) design of a “minimum” offshore structures and zyxw (2) design of an innovative deepwater structure. This chapter will address the following topics: Recent experiences of designs gone bad Design objectives: life cycle quality, reliability and minimum costs Approaches to achieve successful designs Instruments to help achieve design success Example applications 12.2 Recent Experiences of Designs Gone Bad As a result of studying more than 600 “well documented” (these are difficult to find) major recent failures of offshore structures, some interesting insights have been
- 290. 940 zyxwvutsrqpo Chapter 12 z developed [Bea, 2000al: (1) Approximately 80% of the major failures (cost more than U.S. $ zyx 1 million) are directly due to human and organisational factors (HOF) and the malfunctions that develop as a result of these factors (e.g. platform fails due to explosion and fire). These causes will be identified as zyxwvu exhevent zyxwv causes. Only about 20% of these failures can be regarded as being natural or representing residual risk (e.g. platform fails due to hurricane forces). These causes will be identified as iizhevent causes. This finding is a tribute to the engineers and technology that has been used to design these structures. The primary causes of failures are not associated directly with the technology concerned with design for the conditions traditionally addressed by offshore engineers. Of the 80% of the major failures that are due to exherent causes, about 80% of these occur during operations and maintenance activities; frequently, the maintenance activities interact with the operations activities in an undesirable way. Frequently, the structure cannot be operated as intended and short-cuts and adaptations must be made in the field. It is important to define failure. In this chapter, failure is defined as realising undesirable and unanticipated compromises in the quality of the offshore structure. Quality is the result of four attributes: (1) serviceability (fitness for purpose), (2) safety (freedom from undue exposure to harm or injury), (3) durability (freedom from unanticipated degradation in the quality attributes), and (4) compatibility (meets business and social objectives - on time. on budget and happy customers, including the environment). The probability of failure is defined as the likelihood that the quality objectives are not realised during the life cycle of the offshore structure. Reliability is the likelihood that these quality objectives are realised. Of the failures due to the exherent causes that occur during operations and maintenance, more than half (500/,) of these can be traced back to seriously flawed engineering design; offshore structures may be designed according to the accepted industry standards and yet are seriously flawed due to limitations and imperfections that are embedded in the industry standards and/or how they are used. Offshore structures are designed that cannot be built, operated and maintained as originally intended; the structures cannot be built as intended and changes must be made during the construction process to allow the construction to proceed; flaws can be introduced by these changes or flaws can be introduced by the construction process itself. When the structure gets to the field, modifications are made in an attempt to make the structure workable or to facilitate the operations, and in the process additional flaws can be introduced. Thus, during the operations and the maintenance phases, operations personnel are faced with a seriously deficient or a defective structure that cannot be operated and maintained as intended. Of the 20% of failures that do not occur during the operations, the percentages of failures developing during the design and the construction phase are about equal. There are a large number of “quiet” failures that develop during these phases that are
- 291. Design for Reliability: Human and Organisutional Facrors zyxwvu 94 z 1 increasingly frequently ending up in legal proceedings. Recently, there have been several of these failures that have had costs exceeding U S . zyxw S 1 billion. Initially the causes of the failures were identified to be due to flaws in the engineering design processes. However, the causes of these failures were ultimately traced to flaws in EPCO (Engineering, Procurement, Construction, Operating) contracting, organisational and management processes. zyxwvut 0 The failure development process can be organised into three categories of events or stages: (1) initiating, (2) contributing and (3) propagating. The dominant initiating events are developed by “operators” performing erroneous acts of commission or interfacing with the system components that have “embedded pathogens” that are activated by such acts of commission (about 80%); the other initiating events are acts or developments involving acts of omissions (something important left out). The dominant contributing events are organisational; these contributors act directly to encourage or cause the initiating events. In the same way, the dominant propagating events are also organisational; these propagators are generally responsible for allowing the initiating events to unfold into a failure. zyxw A taxonomy (classification system) will be developed for these malfunctions later in this section. It is also important to note that these same organisational aspects very frequently are responsible for the development of “near-misses’’ that do not unfold into failures. It is important to define what constitutes an offshore structure “system”. In this work, a system has been defined as composed of seven primary components: (1) the structure (provides support for facilities and operations), zyxw (2) the hardware (facilities, control systems, life support), (3) the procedures (formal, informal, written, computer software), (4) the environments (external, internal, social), (5) the operators (those that interface directly with the system), (6) the organisations (institutional frameworks in which operations are conducted), and (7) the interfaces among the foregoing. Systems have a life cycle that consists of concept development, design, construction, operation, main- tenance and decommissioning. Failures must be examined in the framework of the components that comprise an offshore structure system and contexts of the life cycle activities. Most failures involve never being exactly repeated sequences of events and multiple breakdowns or malfunctions in the components that comprise an offshore structure system. These events are frequently dubbed “incredible” or “impossible”. After many of these failures, it is observed that if only one of the protective “barriers” had not been breached, then the accident would not have occurred. Experience has adequately shown that it is extremely difficult, if not impossible, to accurately recreate the time sequence of the events that actually took place during the period leading to the failure. Unknowable complexities generally pervade this process because a detailed information on the failure development is not available or is withheld. Hindsight and confirmational bias are common, as are distorted recollections. Stories told from a variety of viewpoints involved in the development of a failure seem to be the best way currently available to
- 292. 942 zyxwvutsrqpo Chapter z 12 capture the richness of the factors, elements and processes that unfold in the development of a failure. The discriminating difference between the “major” and the “not-so-major’’ failure involves the “energy” released by and/or expended on the failure. zy A not-so-major failure generally involves only a few people, only a few malfunctions or breakdowns, and only small amounts of energy that frequently is reflected in the not-so-majordirect and indirect, short-term and long-term “costs” associated with the failure. The major failures are characterised with the involvement of many people and their organisations, a multitude of malfunctions or breakdowns, and the release and/or expenditure of major amounts of energy; this seems to be because it is only through the organisation that so many individuals become involved and the access provided to the major sources of this energy (money is a form of energy). Frequently, the organisation will construct “barriers” to prevent the failure causation to be traced in this direction. In addition, until recently, the legal process has focused on the “proximate causes” in failures; there have been some major exceptions to this focus recently, and the major roles of organisational malfunctions in an accident causation have been recognised in court. It is important to realise that the not-so-major accidents, if repeated very frequently, can lead to major losses. To many engineers who design offshore structures, the human and organisational factor part of the challenge of designing high quality and reliability systems is “not an engineering problem”; frequently, this is believed to be a “management problem”. The case histories of these recent major failures clearly indicate that engineers have a critical role to play if the splendid histories of successes are to be maintained or improved. Engineers can learn how to use existing technology to reach such a goal. The challenge is to wisely apply what is known. To continue to ignore the human and organisational issues in design engineering of offshore structures is to continue to experience things that we do not want to happen and whose occurrence can be reduced. 0 An experience-based (heuristic) classification system (taxonomy) was developed to describe the causes of the recent failures (compromises in quality) that were studied [Bea, 2000al. The taxonomies go beyond human and organisational malfunctions (errors) [Reason, 1990, 19971 and include the structure-hardware malfunctions, the procedure malfunctions, and the environmental influences, This encourages examination of the “parts” in the context of the whole zyxwvut - the offshore structure “system”. The taxonomies define the hows of malfunctions; the generic categories of actions or activities that result in flaws and malfunctions. 12.2.1 Operator Malfunctions There are many different ways to define, classify and describe operator (individual) malfunctions that develop during design, construction, operation and maintenance of offshore structures [Wenk, 1986; Reason, 1990; Kirwan, 1994; Gertman and Blackman, 1994; Center for Chemical Processing Safety, 19941.Operator malfunctions can be defined as actions taken by individuals that can lead an activity to realise a lower quality and reliability than intended. These are malfunctions of commission. Operator malfunctions
- 293. Design for Reliability: Human and Organisational Factors zyxwvu 943 z Table 12.1 Taxonomy of operator malfunctions Communications zyxwvut - ineffective transmission of information Slips - accidental lapses Violations - intentional infringements or transgressions IIgnorance - unaware, unlearned 1 Planning and preparation - lack of sufficient program, procedures, readiness, and robustness Selection and training - not suited, educated or practiced for the activities Limitations and impairment - excessively fatigued, stressed and having diminished senses Mistakes - cognitive malfunctions of perception, interpretation, decision, discrimination, diagnosis and action also include actions not taken that can lead an activity to realise a lower quality than intended. These are malfunctions of omission. Operator malfunctions might best be described as action and inaction that result in lower than acceptable quality. Operator malfunctions also have been described as mis-administrations and unsafe actions. Frequently, the causes of failures are identified as the result of “human errors”. This identification is seriously flawed because errors are results, not causes [Woods, 1990; Reason, 19971.This is an important distinction if one is really interested in understanding how and why malfunctions develop. Operator malfunctions can be described by types of error mechanisms. These include slips or lapses, mistakes and circumventions. Slips and lapses lead to low-quality actions where the outcome of the action was not what was intended. Frequently, the significance of this type of malfunction is small because these actions are not easily recognised by the person involved and in most cases easily corrected. A taxonomy of operator malfunctions based on the study of the failures of offshore structures is given in table 12.1. Mistakes can develop where the action was intended, but the intention was wrong. Circumventions (violations, intentional shortcuts) are developed where a person decides to break some rule for what seems to be a good (or benign) reason to simplify or avoid a task. Mistakes are perhaps the most significant because the perpetrator has limited clues that there is a problem. Often, it takes an outsider to the situation to identify mistakes. A taxonomy of operator mistakes is given in table 12.2. It is important to note that the study of failures involving offshore structures clearly indicates that the single leading factor in operator malfunctions is communications. Communications can be very easily flawed by “transmission” problems and “reception” problems. Feedback, that is so important to validate the communications, frequently is not present or encouraged. Language, culture, societal, physical problems and environmental influences can make this a very malfunction-prone process. Also note the importance of violations, ignorance (failure to use the existing technology or develop the necessary
- 294. 944 zyxwvutsrqpo Table 12.2 Taxonomy of mistakes zyxw Chapter zy 12 1Perception - zyxw unaware, not knowing I 1Interpretation - improper evaluation and assessment of meaning 1 1Decision - incorrect choice between alternatives I Discrimination - not perceiving the distinguishing features Diagnosis - incorrect attribution of causes and or effects Action - improper or incorrect carrying out activities technology), planning and preparation, and selection and training. Engineers are frequently asked or required to do things that they are not sufficiently trained to do, and in some cases, are not capable of doing. But, they try. 12.2.2 Organisational Malfunctions The analysis of the history of failures of offshore structures provides many examples in which organisational malfunctions have been primarily responsible for the failures. Organisational malfunction is defined as a departure from the acceptable or the desirable practice on the part of a group of individuals that results in unacceptable or undesirable results. Based on the study of case histories regarding the failures of offshore structures, studies of High Reliability Organisations (HRO) [Roberts, 1989, 1993; Weick, 19991, and managing organisational risks [Reason, 1997; Haber, et a1 1991; Wu, et a1 19891, a classification of organisational malfunctions is given in table 12.3. Table 12.3 Taxonomy of organisational malfunctions lCimiuZations - ineffective transmission of information I 1 Culture - inappropriate goals, incentives, values and trust I Violations - intentional infringements or transgressions Ignorance - unaware, unlearned Planning and preparation - lack of sufficient program, procedures, readiness Structure and organisation - ineffective connectedness, interdependence, lateral and vertical integration, lack of sufficient robustness Monitoring and controlling - inappropriate awareness of critical developments and utilisation of ineffective corrective measures Mistakes - cognitive malfunctions of perception, interpretation, decision, discrimination, diagnosis and action
- 295. Design zyxwvutsrqpon for Reliability: Human and Organisational Factors zyxwvu 945 z Frequently, the organisation develops high rewards for maintaining and increasing production; meanwhile the organisation hopes for quality and reliability (rewarding “A” while hoping for “B’) [Roberts, 19931. The formal and informal rewards and incentives provided by an organisation have a major influence on the performance of operators and on the quality and reliability of offshore structures. In a very major way, the performance of people is influenced by the incentives, rewards, and disincentives provided by the organisation. Many of these aspects are embodied in the “culture“ (shared beliefs, artefacts) of an organisation. This culture largely results from the history (development and evolution) of the organisation. Cultures are extremely resistant to change; particularly if they have been “successful”. Several examples of organisational malfunctions recently have developed as a result of efforts to down-size and out-source as a part of re-engineering organisations [Bea, et a1 19961. The loss of corporate memories (leading to repetition of errors), creation of more difficult and intricate communications and organisational interfaces, degradation in morale, unwarranted reliance on the expertise of outside contractors, cut-backs in quality assurance and control, and provision of conflicting incentives (e.g. cut costs, yet maintain quality) are examples of activities that have lead to substantial compromises in the intended quality of systems. Much of the down-sizing (“right-sizing”) outsourcing (“hopeful thinking”) and repeated cost-cutting (“remove the fat until there is no muscle or bone”) seems to have its source in modern business consulting. While some of this thinking can help promote “increased efficiency” and maybe even lower CapEx (Capital Expenditures), the robustness (damage and defect tolerance) of the organisation and the systems it creates are greatly reduced. Higher OpEx (Operating Expenditures) and more “accidents” can be expected; particularly in the long-run - if there is one, before the system is scraped or sold. Experience indicates that one of the major factors in organisational malfunctions is the culture of the organisation. Organisational culture is reflected in how action, change, and innovation are viewed: the degree of external focus as contrasted with internal focus; incentives provided for risk-taking; the degree of lateral and vertical integration of the organisation; the effectiveness and honesty of communications; autonomy, responsibility, authority and decision making: trust; rewards and incentives; and the orientation towards the quality of performance contrasted with the quantity of production. In some organisations, the primary objective becomes “looking good”, not doing good. The culture of an organisation is embedded in its history. One of the major cultural elements is how managers in the organisation react to suggestions for a change in the management. Given the extreme importance of the organisation and its managers on quality and reliability, it is essential that these managers see suggestions for change (criticism?) in a positive manner. This is extremely difficult for some managers because they do not want to relinquish or change the strategies and processes that had made them managers. Another major cultural element is how organisations react to failures. Often the focus is on blame and shame; the author calls this focus “kill the victims”. Often the view is one that localises the failure; the fostered belief is that the failure was caused by a few misguided, poorly motivated or trained people. These reactions tend to stop the learning that can be
- 296. 946 zyxwvutsrqpon Chapter zy 12 z developed by truly understanding the factors and situations that result in failures. These reactions tend to suppress the early warning signs that developing failures can provide. z 12.2.3 Structure, Hardware, Equipment Malfunctions Human malfunctions can be initiated by or exacerbated by poorly designed offshore structures and procedures that invite errors. Such structures are difficult to construct, operate and maintain. Table 12.4 summarises a classification system for hardware- (equipment, structure) related malfunctions. New technologies compound the problems of latent system flaws (structural pathogens) [Reason, 19971. An excessively complex design, a close coupling (the failure of one component leads to the failure of other components) and severe performance demands on systems increase the difficulty in controlling the impact of human malfunctions, even in well operated systems. The field of ergonomics (people-hardware interfacing) has much to offer in helping create “people-friendly’’ offshore structures [ABS, 19981. Such structures are designed for what people will do, not what they should be able to do. Such structures facilitate construction (constructability), operations (operability), and maintenance (maintainability, repairability). The issues of offshore structure system robustness (defect or damage tolerance), design for constructability [Bea, 19921 and design for IMR (inspection, maintenance, repair) are critical aspects of engineering systems that will be able to deliver acceptable quality. The design of the system to assure robustness is intended to combine the beneficial aspects of configuration, ductility and excess capacity (it takes all three!) in those parts of the structure system that have high likelihoods and consequences associated with developing defects and damage. The result is a defect and damage tolerant system that is able to maintain its quality characteristics in the face of HOF. This has important ramifications with regard to engineering system design criteria and guidelines. A design for constructability is a design to facilitate construction, taking account of worker’s qualifications, capabilities, and safety, environmental conditions, and the interfaces between the equipment and workers. A design for IMR has similar objectives. A reliability-centered maintenance (RCM) has been developed to address some of these problems, and particularly the unknowable and the HOF aspects [Jones, 19951. Table 12.4 Taxonomy of structure and equipment malfunctions Serviceability - inability to satisfy purposes for intended conditions Safety - excessive threat of harm to life and the environment, demands exceed capacities Durability - occurrence of unexpected maintenance and less than expected useful life; unexpected degradation in other quality characteristics Compatibility - unacceptable and undesirable economic, schedule, environmental and aesthetic characteristics 1
- 297. Design zyxwvutsr for Reliabiiify: Human and Organisational Factors zyxwvu 941 z It is becoming painfully clear that our engineering design guidelines for the creation of sufficient robustness zyxwvu - damage - defect tolerance in offshore structure systems is not sufficient. Our thinking about sufficient damage stability and damage tolerance needs rethinking. Our thinking about designing for the “maximum incredible” events needs more development. While two offshore structures can both be designed to “resist the 100-yr conditions” with exactly the same probabilities of failure, the two structures can have very different robustness or damage tolerance during the life cycle of the structures. The “minimum” CapEx offshore structure may not have a configuration, excess capacity or ductility to allow it to weather the inevitable defects and the damage that should be expected to develop during its life. Sufficient damage tolerance almost invariably results in increases in CapEx; the expectation and the frequent reality are that OpEx will be lowered. But, one must have a “long-term” view for this to be apparent. Recent work has clearly shown that the foregoing statements about structure and hardware robustness apply equally well to organisations and operating teams. Proper configuration, excess capacity and ductility play out in organisations and teams in the same way that they do in the structure and hardware [Bea, 2000a. b]. It is when the organisation or an operating team encounters defects and damage - and is under serious stress, that the benefits of robustness become evident. A robust organisation or an operating team is not a repeatedly downsized (lean and mean?), out-sourced and financially strangled organisation. A robust organisation is a Higher Reliability Organisation (HRO). zyx 12.2.4 Procedure and Software Malfunctions Based on the study of procedure and software-related problems that have resulted in failures of offshore structures, table 12.5 summarises a classification system for procedure or software malfunctions. These malfunctions can be embedded in engineering design guidelines and computer programs, construction specifications and operations manuals. They can be embedded in contracts (formal and informal) and subcontracts. They can be embedded in how people are taught to do things. With the advent of computers and their integration into many aspects of the design, construction, and operation of oil and gas structures, software errors are of particular concern because the computer is the ultimate fool [Knoll, 1986; Rochllin, 19971. Software errors in which incorrect and inaccurate algorithms were coded into computer programs have been the root cause of several recent failures of offshore structures Table 12.5 Taxonomy of procedure and software malfunctions 1 I Incorrect - faulty I Inaccurate - untrue Incomplete - lacking the necessary parts 1Excessive complexity - unnecessary intricacy 1 1Poor organisation - dysfunctional structure I Poor documentation - ineffective information transmission
- 298. 948 zyxwvutsrqpo Chapter zy 12 [Bea, 2000a, b]. Guidelines have been developed to address the quality of computer software for the performance of finite element analyses. Extensive software testing is required to assure that the software performs as it should and that the documentation is sufficient. Of particular importance is the provision of independent checking procedures that can be used to validate the results from analyses. High-quality procedures need to be verifiable based on first principles, results from testing and field experience. This has particular importance in the quality assurance and quality control (QAIQC) in design. Given the rapid pace at which significant industrial and technical developments have been taking place, there has been a tendency to make design guidelines, construction specifications and operating manuals more and more complex. Such a tendency can be seen in many current guidelines used for the design of offshore structures. In many cases, poor organisation and documentation of software and procedures has exacerbated the tendencies for humans to make errors [Rochlin, 19971. Simplicity, clarity, completeness, accuracy and good organisation are desirable attributes in procedures developed for the design, construction, maintenance and operation of offshore structures. zyx 12.2.5 Environmental Influences Environmental influences can have important effects on the performance characteristics of individuals, organisations, hardware and software. These include: All three of these environmental influences can have extremely important effects on human, operating team and organisational malfunctions. External (e.g. wind, temperature, rain, fog, time of day), Internal (lighting, ventilation, noise, motions) and Sociological factors (e.g. values, beliefs and morays). 12.3 Design Objectives: Life Cycle Quality, Reliability and Minimum Costs 12.3.1 Quality In this development, the term "quality" is defined as freedom from unanticipated defects in offshore structures. Quality is fitness for purpose. Quality is meeting the requirements of those that own, operate, design, construct and regulate offshore structures. These requirements include those of serviceability, safety, compatibility and durability [Matousek, zyxwvut 19901. Quality is freedom from unanticipated defects in the serviceability, safety, durability and compatibility of the offshore structure system. Serviceability is suitability for the proposed purposes, i.e. functionality. Serviceability is intended to guarantee the use of the system for the agreed purpose and under the agreed conditions of use. Safety is the freedom from excessive danger to human life, the environment and property damage. Safety is the state of being free of undesirable and hazardous situations. The capacity of a structure to perform acceptably during extreme demands and other hazards is directly related to and most often associated with safety. Compatibility assures that the structure does not have unnecessary or excessive negative
- 299. Design for Reliability: Human and Organisational Factors zyxwvu 949 z impacts on the environment and society during its life cycle. Compatibility also is the ability of the structure to meet economic, time, political, business and environmental requirements. Durability assures that serviceability, safety and environmental compatibility are maintained during the intended life of the structure. Durability is freedom from unanticipated maintenance problems and costs. Experience with offshore structures has shown that durability is one of the most important characteristics that must be achieved; if insufficient durability is developed, then there are unanticipated and often undetected degradations in the other quality characteristics, and many times these degradations have disastrous results. This is a holistic definition of the key objective of engineering design processes; to achieve adequate and acceptable quality [Hessami, 19991. In recent years, a wide variety of processes, procedures and philosophies intended to improve and achieve adequate quality in goods and services have been developed and implemented including Total Quality Management [Demming, 19821, QA/QC and the International Standards Organization Quality Standards [ISO, 1994a, 1994b, 1994~1. These components are the building blocks of a quality management system. Engineers have learned that it is important to recognise that these processes, procedures and philosophies are all related to the same objective; they represent complementary parts of activities that are intended to help achieve adequate and acceptable quality in offshore structures. The challenge has been to learn how to use these processes, procedures and philosophies wisely, effectively and efficiently. Also, it is important to note that the traditional “business” objectives (e.g. serviceability, compati- bility) have been merged with the traditional “safety” objectives; quality can be good for business and vice versa. zyxwvu 12.3.2 Reliability Reliability is defined as the probability (likelihood) that a given level of quality will be achieved during the design, construction and operating life cycle phases of an offshore structure. Reliability is the likelihood that the structure system will perform in an acceptable manner. Acceptable performance means that the structure system has desirable serviceability, safety, compatibility and durability. The reliability, zy Ps (likelihood of success), can be expressed analytically as: Ps = P[C D] (12.1) where P[] is read as the likelihood that [ 1. D is the demandjs imposed on the system, and C is the capacity/ies of the system to successfully withstand the imposed demand/s. The complement of reliability is the likelihood or probability of unacceptable quality; the probability of failure, PJ Ps = P[D2 zyxwv q = 1 - Ps (12.2) This definition has linked the concepts of probability, uncertainty and reliability with the holistic definition of quality to reflect upon the likelihoods of achieving acceptable quality in offshore structures.
- 300. 950 zyxwvutsrqponm Chapter 12 z Compromises in quality of a structure system can occur in the structure itself and/or in the facilities it supports. These failures can be rooted in malfunctions developed by individuals (operators) in design, construction, operation, and/or maintenance. Individuals, the people who design, construct, operate and maintain the systems have direct influence on malfunctions developed in these phases. However, the malfunctions developed by the individuals can be and often are caused (contributing factors) or compounded (propagating factors) by malfunction-inducing influences from organisations, hardware, software (procedures) and environment (external, internal). It is the combination of the individuals, organisations, procedures, environments, hardware, structure and interfaces between the foregoing that constitutes an offshore structure system. zyxw An offshore structure can only be understood realistically in the context of all of the components or elements that comprise the structure system and influence its life cycle performance characteristics. The calculation of reliability or its complement, the likelihood of failure can be done in a variety of ways. The most straightforward method is to numerically integrate two distributions: (12.3) where F, is the cumulative distribution for the capacities (probability that capacity is equal to or less than a given demand, d), f D is the density distribution for the demands (probability that the demand is in the interval zyxwv Ad, and Ad is the integration interval. This expression can be used for any form of the distributions of demands and capacities. This expression can incorporate dependency (correlation) between the demands and capacities (e.g. as demand increases, capacity decreases) through the means used to define the cumulative distribution for the capacities. Given that the distributions of demands and capacities can be reasonably characterised as normal (Gaussian) and independent, then P f can be computed directly from: (12.4) where zyxwvuts p is defined as the safety index, is the mean (average) capacity, zy B is the mean demand, oC is the standard deviation of the distribution of capacities, and oD is the standard deviation of the demands. If the demands and capacities are not independent, then the safety index can be determined from: (12.5) where pDc is the correlation coefficient (-1 5 zyx p 51)that recognises the dependency of the magnitudes of the demands and capacities. Given that the distributions of demands and capacities can be reasonably characterised as Lognormal (normal distribution of logarithms) and independent, then Pfcan be computed
- 301. Design zyxwvutsrq for Reliabilitj’: Human and Organisational Factors zyxwvu 951 z directly from: (12.6) where zyxwvuts p is defined as the safety index, C50is the median (50th percentile) capacity, D50 is the median demand, olnC is the standard deviation of the Lognormal distribution of capacities and olnD is the standard deviation of the Lognormal distribution demands. If the demands and capacities are not independent, then the safety index can be determined from: In(CSO/DSO) P = + of,^ - 2pDcolnColnD The likelihood of failure is determined from the safety index as: (12.7) P f = 1 - zyx cp (p) (12.8) where index. For values of P between 1 and 3: (p) is the standard cumulative normal distribution for the value of the safety Pf 0.475 exp - (p)’.6 (12.9) even more approximately: Pf =10-p (12.10) The safety index is like a factor-of-safety; as P gets larger, Pfgets smaller. Note in equations (12.6) and (12.7) the ratio C50,’D50is like the traditional factor-of-safety; it is the ratio of the median capacity to the median demand (load). This ratio is referred to as the median factor-of-safety. As the factor-of-safety gets larger, the safety index gets larger, and the likelihood of failure gets smaller. Also, as the uncertainty in the demand and capacity increases (reflected in the standard deviations), the safety index gets smaller, and the likelihood of failure gets larger. Greater uncertainties lead to greater likelihoods of failure. zyxwvuts A very useful “normalised” characterisation of the uncertainty characteristics is the coefficient of variation (COV, ratio of standard deviation to mean value of variable z X= Vx).The COV of a variable, X,is related to the standard deviation of the logarithm of the distribution of X as: for small values of V, (less than about 40%), olnX x V,. It is important to recognise that the variables used in designing offshore structures are often “conservative”. Thus, there can be a source or sources of “bias” that must be eliminated or
- 302. 952 zyxwvutsrqponm Chapter 12 z recognised quantitatively in analyses of zyxw Ps or Pf.The actual mean or median values of demands and capacities are required to develop realistic evaluations of Ps or Pf. Also, it is important to recognise that there are different types of uncertainties that determine the resultant uncertainties associated with demands and capacities. One type of uncertainty (Type 1) is natural or inherent; this type of uncertainty is “information insensitive and random”. zyxwvu A second type of uncertainty (Type 2) is associated with modelling, parametric and state uncertainties; this type of uncertainty is “information sensitive” and systematic. A third type (Type 3) of uncertainty is related to HOF. The focus of this chapter is on the third type of uncertainty. However, many of the thinking and analytical processes that have been used to address the Type 1 and the Type 2 uncertainties associated with designing offshore structures are adaptable to the Type 3 uncertainties. This adaptation will be illustrated later in this chapter. It is very important to properly identify and characterise the Type 2 uncertainties. One approach is to express the Type 2 uncertainties as a “Bias” where this term is defined as the ratio of the actual or true value of the variable to the predicted or nominal (design) value of the variable. A variety of methods can be used to characterise the bias including field test data, laboratory test data, numerical data, and “expert” judgement. Often it is not possible to develop unambiguous separations of the Type 1 and Type 2 uncertainties and it is important not to include them twice. There are “advanced” approaches to calculating Ps and P’that are “distribution free” in the sense that a particular type of likelihood distribution (e.g. Normal, Lognormal) does not have to be assumed. These approaches have been termed first order reliability methods (FORM) and second-order reliability methods (SORM). While they are more advanced, they require much more complicated numerical methods to perform the calculations, and they too involve approximations. Because of these properties, the author suggests the use of equations (12.3), (12.6) and (12.7) to perform the majority of reliability analyses. The Lognormal distributions can be “fitted” to the important parts of the parameter distributions of concern (this takes some knowledge) and develop results that are very close to those from the more advanced approaches. The advantage of this “simplified” approach is that it is relatively transparent compared with the advanced approaches (calculations can be readily performed) and it can be used by design engineers that have a knowledge of the fundamentals of statistics and probability. 12.3.3 Minimum Costs Providing quality in the design of an offshore structure can result in lower life cycle costs, be safer, and minimise unrealised expectations during the life cycle of the facility. Quality can result in significant benefits to minimise costs and increase income through maximised serviceability and availability (durability). In this development, the focus will be solely on costs; however, even greater benefits can be developed if the maximised serviceability and availability effects on income are recognised. Achieving adequate levels of quality and reliability is not quick, easy, or free. It can be costly in terms of the initial investments of manpower, time and other resources required to achieve it (fig. 12.1). But, if it is developed and maintained, it can result in significant savings in future costs. In addition, initial costs can be reduced by discarding ineffective
- 303. Design for Reliabilify: Human and Organisational Factors zyxwv fnzyxwvut + zyxwv B zyxwvu 953 4 Total zyxwv Costs I E 8 LEVEL OF QUALITY zyxw Figure 12.1 Consideration of initial and future costs associated with various levels of quality and inefficient programs that are currently in use. A basic objective is to find ways to reduce both initial and future costs and thereby provide both a short-term and long-term financial incentive to implement improved quality and reliability programs. The objective is to find the level or degree of quality that will minimise the total of initial and future costs. Different levels of quality are needed for different levels of criticality of elements in a system. If a system element or component is particularly critical to the quality and reliability of a system, then even though it may have identical initial costs, it may have very different future costs (fig. 12.2).Higher levels of quality and more intense QA/QC (Quality Assurance and Quality Control) measures should be relegated to those elements and components that have higher levels of criticality. The costs to correct insufficient quality are a function of when the deficiencies are detected and corrected (fig. 12.3). The earlier the deficiencies are caught and fixed, the lesser the costs. The most expensive time to fix quality deficiencies is after the system is placed in service. This places a large premium on early detection and correction of errors. Not only Future Costs + ’. 1’. I t/ best quality lev$ -~ *, LEVELOF for element #3 QUALITY Figure 12.2 Criticality should determine the level of quality
- 304. 954 zyxwvutsrqpo Chapter 12 z v) K zyxwv 0 K K w c 0 w K K 0 0 zyxw z v) c v) 0 0 Concept Preliminaly Detailed Construction zyxwv 8 In Service Design Design Design Commision LIFE CYCLE PHASE zyx Figure 12.3 Life cycle costs to correct errors are there large direct future costs associated with fixing errors, but also large indirect costs associated with loss of business and loss of image. The present value of the total life cycle cost, C,associated with the performance of a system can be expressed as: c zyxwv = c o +( C s t Cr+ CM+ C R ) = co +CF (12.12) where the subscript 0 refers to the initial cost, zyxw S refers to loss of serviceability cost, z Zrefers to inspection cost, M refers to structural maintenance costs, R refers to structural repair costs and F to the total future costs associated with the maintenance of the system. Assuming a continuous discounting, each of the individual costs can be expressed as: cX= CC, exp(-r T,) = C,(PVF,) (12.13) where the uppercase subscript X refers to a type of cost, the lowercase subscript zy x refers to the specific cost, the summation is taken over the occasions or time for the category of cost, Y is the net discount rate, Tis the time that the expense is incurred, and PVF is the resultant net present value function. All of these categories of costs are variable and uncertain. Likelihoods (or probabilities) can be entered into the process in several ways. A traditional approach has been to focus on expected costs in which the estimated cost is multiplied by the probability, P. of experiencing that cost: E [CX]= CXPX (12.14)
- 305. Design for Relrabilirq: Human and Organisational Factors zyxwvu The total expected cost can be written as: 955 z (12.15) The expected initial cost includes the costs associated with the system capacity, durability (degree of corrosion and fatigue protection provided including materials, redundancy and robustness integrated into the system), and construction (including degree of QA/QC provided). The expected future cost includes the costs associated with loss of serviceability of the system (Cs) and the costs associated with a given inspection, maintenance, repair (IMR) program (CZ,CM,CR). Often, it is useful to provide the decision making process with an expression of the uncertainties associated with the expected costs. The uncertainties associated with each of the cost and the probability variables can be estimated on the basis of analyses, data and experience. Based on a first-order, second moment approximation that utilises the mean values and coefficients of variation (COV = V= ratio of standard deviation to mean) for each of the cost and components (CX,Px, Vex, zyxw Vpx)the mean cost and coefficient of variation in that cost can be estimated as: _ _ (12.16) Vc = zyxwv 4 - (12.1 zy 7) The process of defining what constitutes desirable quality and reliability can be expressed as a utility maximisation process. The objective of the utility maximisation process can be expressed as an expected total cost minimisation: (12.18) The expected value costs associated with an alternative is the average monetary result per decision that would be realised if the decision makers accepted the alternative over a series of identical repeated trials. The expected value concept is a philosophy for consistent decision making, which if practiced consistently, can bring the sum total of the utilities of the decision to the highest possible level. The expected value is not an absolute measure of a monetary outcome. It is incorrect to believe that the expected value is the most probable result of selecting an alternative. If one wanted to determine the probabilities of different magnitudes of utilities, then likelihoods could be assigned to each of the cost elements and these likelihoods propagated through the cost and likelihood evaluations to develop probability distributions of the potential utilities. Given that the initial costs associated with a given quality alternative can be related linearly to the logarithm of PF:
- 306. 956 zyxwvutsrqpo Chapter 12 z Co is the initial cost versus PFintercept, ACo is the slope of the initial cost curve, and z Po is the probability that the estimated initial cost will be realised. Given that the inspection, maintenance, and repair costs do not vary significantly with PF,differentiating the sum of initial and future cost with respect to PF to find the point of zero slope gives the PF that produces the lowest total cost (Pro): 0.435 Pf zyxwvut -- zyx O - R,PVF (12.20) R,(cost ratio) is the ratio of the present valued future cost, C,, to the expected cost needed to decrease PF by a factor of 10, ACo: CF - ACo R -- (12.21) PVF is a present value discount function. Based on continuous discounting and replacement: PVF = [l - (1 + zyxwv Y)-"]/Y (12.22) For a continuous discount function and long-life system (life > 10 yr), PVF zy x Y-' where r is the monetary net discount rate (investment rate minus inflation rate). For short-life systems (life 5 5 yr), PVF x L, where L is the life in years. As shown in fig. 12.4, as the costs associated with the development of insufficient quality increases, the reliability must increase. As the initial costs to achieve quality increases, the optimum reliability decreases. The optimum reliability is based on the quality that will develop the lowest total initial and future costs. The marginal probability of insufficient quality is double the optimum quality probability. It is the quality in which the incremental investment to achieve quality equals the incremental future benefit (cost/benefit = 1 z .O). Reliability of a system element, component and system is a function of its criticality expressed by the product of the cost ratio and present value function. Figure 12.4 The economics and likelihood of insufficient quality
- 307. Design for Reliabilit?.: Human and Organisational zyxwvu Fuctors zyxwvuts 957 Quality can be a substantial competitive aspect in industrial activities. If a purchaser or user recognises the benefits of adequate quality and is able and willing to pay for it, then quality can be a competitive advantage. If a purchaser or user does not recognise the benefits of adequate quality or is unable or unwilling to pay for it, then quality can be a competitive disadvantage. Purchaser/owner quality goals must be carefully defined so that uniformity can be developed in the degrees of quality offered in a product or service sector. Once these goals have been defined, then the purchaseriowner must be willing to pay for the required quality. zyxwvut 12.4 Approaches to Achieve Successful Designs An important starting point in addressing HOF in the quality and reliability of offshore structures is to recognise that while human and organisational malfunctions and errors are inevitable, their occurrence can be reduced and their effects mitigated by improving how structures are designed, constructed, operated, maintained and decommissioned. Engineering can improve the processes and products of design, construction, operations, maintenance and decommissioning to reduce the malfunction promoting characteristics, and to increase malfunction detection and recovery characteristics. Engineering can help develop systems for what people will do, not for what they should do. Engineering also can have important influences on the organisation and management aspects of these systems. Engineering organisations have important and pervasive influences on the reliability of offshore structure systems. High reliability organisations (HROs) have been shown to be able to develop high reliability systems that operate relatively error free over long periods of time and in many cases, in very hazardous environments. The HROs go beyond Total Quality Management and the International Standards Organization certifications in their quest for quality and reliability. They have extensive process auditing procedures to help spot safety problems and have reward systems that encourage risk-mitigating behaviours, They have high quality standards and maintain their risk perception and awareness. Most importantly, such organisations maintain a strong command and control system that provides for organisational robustness or defect tolerance. There are three fundamental, complimentary and interactive approaches to achieving adequate and acceptable quality and reliability in offshore structures: Proactive (activities implemented before malfunctions occur), Reactive (activities implemented after malfunctions occur) and Interactive or real-time (activities implemented during occurrence of malfunctions). In the context of these three approaches, there are three primary strategies to be employed: Reduce incidence of malfunctions, Reduce effects of malfunctions. One approach frequently considered by engineers is to combat the potential effects of human and organisational malfunctions by increasing the structure’s factors-of-safety.This has not proven to be an effective approach because making the structure stronger for the Increase detection and correction of malfunctions and
- 308. 958 zyxwvutsrqpo Chapter z 12 design loadings does not necessarily make the structure more reliable for extrinsic hazards. It has proven to be much more effectiveto implement measures directed at the source of the unreliability zyxwvut - people and their organisations. zyxw 12.4.1 Proactive Approaches The proactive approach attempts to understand a system even before it fails (unacceptable quality) in an attempt to identify how it could fail in the future. Measures can then be put in place to prevent the failure or failures that have been anticipated. Proactive approaches include well-developed qualitative methods such as HazOp (Hazard Operability) and FMEA (Failure Mode and Effects Analyses) and quantitative methods such as SRA (Structural Reliability Analyses), PRA (Probabilistic Risk Analyses) and QRA (Quantified Risk Analyses) [Center for Chemical Process Safety, 1989; Spouge, 1999; Moan, 1997; Soares, 1998;Vinnem, 19981. Each of these methods have benefits and limitations [Groeneweg, 1994; Molak, 1997; Apostolakis, et a1 1990; Aven and Porn, 1998; Bier, 19991. Proactive approaches also include organisational - management improvements and strategies that are intended to develop higher reliability organisations (HROs). Such organisations are able to operate over long periods of time conducting relatively error free operations and to consistently make good decisions regarding quality and reliability. The creation of HROs is perhaps the most important proactive approach. Another proactive approach that has not received the attention that it deserves is the creation of “robust” offshore structures and similarly robust design organisations. Robustness is defined here as damage or defect tolerance. Robustness in a structure or an organisation means that it can continue to operate satisfactorily without surrendering fundamental quality and reliability objectives until repairs and/or modifications can be made. These are “human friendly” structures in the sense that they can tolerate high probability defects and damage that have sources in human and organisational malfunctions. Studies of robustness in offshore structures [Avigutero and Bea, 1998; Bea, 2000al have shown that it takes the combination of four attributes to create a robust structure system: configuration, ductility, excess capacity and appropriate correlation Configuration relates to the topology of the structure system; how elements and materials are arranged. Frequently, this has been called “redundancy”; referring to the degree of static indeterminancy. But, configuration goes beyond redundancy so that as elements or members are damaged or defective, that the structure system is still able to perform acceptably until repairs and modifications can be made. Ductility relates to the ability of the structure to shift the paths of demands or loads imposed on the elements and system. It relates to the ability of the structure materials and elements to deform nonlinearly without undue loss in capacity. Excess capacity relates to the ability of the structure system to carry normal loadings and over-loadings even though some of its elements may be damaged or
- 309. Design for Reliability: Human and Organisational Factors zyxwvu 959 defective. This means that some elements must be intentionally “over-designed’’ relative to the normal loading patterns and distributions so that these elements can carry the loadings that are transferred to them when other members or elements are damaged, defective or fail. Appropriate correlation refers to the dependence between the strengths of paired elements. In systems comprised of parallel elements, independence is desirable. In systems comprised of elements in series, dependence or high correlation is desirable. This is fail-safe or intrinsically safe design. Robust systems are not created by over zealous value improvement programs (VIP), excessive downsizing and outsourcing and excessive initial cost cutting (reduced CAPEX at the expense of future OPEX). A Recent work with an HRO has clearly shown that development of robustness in engineering organisations is a very desirable proactive measure. Such organisations can tolerate defects and damage and still perform acceptably. This work also has shown that it takes the same three fundamental attributes: configuration, ductility and excess capacity. Such organisations are not downsized, out-sourced or cost-cut to the point that the organisation cannot tolerate daily and abnormal demands. Some organisation “fat” is a good thing when it allows the organisation to perform well when distressed. The author has been an active protagonist and practitioner of the proactive reliability analysis-based approach to help improve the quality of offshore structures for more than three decades [Bea, 1974, 1975, 2000a; Marshall and Bea, 19761. He believed that this approach provided an ability to forecast how systems could go bad. Very sophisticated analytical models could be developed to help foster this belief. Results from these analyses seemed to have value and to enhance his abilities to address some types of variability and uncertainty. This approach was workable as long as he dealt with systems in which the interactions of people with the systems were minimal or minimised. However, the problem changed radically when people began to exert major influences on the quality of the systems and in many cases on the physical aspects of the systems [Bea, 1996a, b]. In this case, his lack of knowledge of the physics and mechanics of the complex behaviours of people that in the future would design, construct, operate and maintain the system defined an “unpredictable”, or certainly one with very limited predictability. The author’s analytical models addressed systems that were essentially static and mechanical. Yet the real systems were dynamic, constantly changing, and more organic than mechanical. The analytical models generally failed to capture the complex interactions between people and the systems that they designed, constructed, operated and maintained. The author found most data on the reliability of humans in performing tasks to be very limited [Kirwan, 1994; Gertman and Blackman, 1994; Dougherty and Fragola, 1986; Center for Chemical Process Safety, 19941. Existing databases failed to capture or adequately characterise influences that had major effects on human reliability [Wu, et a1 1989; Haber, et a1 19911. Yet, when the numbers were supplied to the very complex analytical models and the numbers were produced, the results were often mistaken for “reality”. There was no way to verify the numbers. If the results indicated that the system was “acceptable”: then nothing was done. If the results indicated that the system was “not acceptable”, then generally the equipment and the hardware fixes were studied in an attempt to define a fix or fixes that would make the system acceptable or ALARP zy (As Low As Reasonably Practicable) [Melchers, 19931. When the author went to the field to
- 310. 960 zyxwvutsrqpo Chapter z 12 compare his analytical models with what was really there, he found little resemblance between his models and what was in the field [Bea, 1996bl. The author does not advocate discarding the analytical-quantitative proactive approach. He advocates using different types of proactive approaches to gain insights into how systems might fail and what might be done to keep them from failing [Weick, 2000; Bea, 2000a, b]. The marked limitations of the analytical models and the quantitative methods must be recognised or major damage can be done to the cause of the quality and reliability of offshore structures. The potential for engineers to be “hyper rational” and attempt to extend the applicability of SRA/PRA/QRA methods beyond their limitations must be recognised and countered. On the other hand, qualitative methods (e.g. HazOp, FMEA), in the hands of qualified and properly motivated assessors (both internal and external) can do much to help the causes of quality and reliability [Center for Chemical Process Safety, 1989, 19941. Experience, judgement and intuition of the assessors needs to be properly recognised, respected and fully integrated into the proactive qualitative and quantitative approaches. Much headway has been made recently in combining the powers of qualitative methods with quantitative Risk Assessment and Management (RAM) methods [Bea, 2000a, b]. The qualitative methods are able to capture more fully the dynamic, changing, organic, complex interactions that cannot be analysed [Weick, 2000; Haber, et a1 1991; Groeneweg, 19941. Given an input from the qualitative methods, the quantitative methods are able to provide numbers that can be used to assist the development of judgements about when, where and how to better achieve quality and reliability in offshore structures. But, even at this level of development, the proactive RAM methods are very limited in their abilities to truly provide quality and reliability in offshore structures. Other methods (e.g. interactive RAM) must be used to address the unknowable and the unimaginable hazards. It is the author’s experience in working with and on offshore structure systems for more than four decades, that many if not most of the important proactive developments in the quality and reliability of these systems were originated in a cooperative, trust-based venture of knowledgeable “facilitators” working with seasoned veterans that have daily responsibilities for the quality of these systems. This cooperative venture includes design, construction/decommissioning, operations and maintenance/inspection personnel. Yet, it is also the author’s experience, that many engineering and many well-meaning reliability z - risk analysis “experts” are not developing a cooperative environment. This is very disturbing. The conduct of each operation during the life cycle of an engineered system should be regarded as the operations of “families”. Knowledgeable, trained, experienced and sensitive outsiders can help, encourage and assist “families” to become “better”. But, they cannot make the families better. Families can only be changed from within by the family members. Proactive measures based on casual or superficial knowledge of a system or of an operation of that system should be regarded as tinkering. And, tinkering can have some very undesirable effects and results [Wenk, 1986; Woods, 1990; Weick, 1995; Bea, 1996, 2001al. The crux of the problem with the proactive analytical approaches is with the severe limitations of such approaches in their abilities to reasonably characterise human and organisational factors and their effects on the performance of a system [Center for Chemical Process Safety, 1994;Reason, 1997;Groeneweg, 1994;Haber, et a1 1991;Wu, et a1
- 311. Design zyxwvutsrqponm for Reliability. Human and Organisutional Factors zyxwvu 961 1981; Rasmussen, et a1 19871. Quantitative analytical approaches rely on an underlying fundamental understanding of the physics and mechanics of the processes, elements and systems that are to be evaluated. Such understanding then allows the analyst to make projections into the future about the potential performance characteristics of the systems. And, it is here that the primary difficulties arise. There is no fundamental understanding of the physics and mechanics of the future performance zyxw - behaviour characteristics of the people that will come into contact with a system and even less understanding of the future organisational influences on this behaviour. One can provide very general projections of the performance of systems including the human and organisational aspects based on extensive assumptions about how things will be done, but little more. The problem is that engineers and managers start believing that the numbers represent reality. To the author, the true value of proactive approaches does not lie in their predictive abilities. The true value lies in the disciplined process such approaches can provide to examine the strengths and weaknesses in systems; the objective is detection and not prediction. The magnitudes of the quantitative results, if these results have been generated using reasonable models and input information, can provide insights into where and how one might implement effective processes to encourage development of acceptable quality and reliability. The primary problems that the author has with the quantitative reliability analysis proactive approach are with how this method is used and what it is used to do. Frequently the results from the approach are used to justify meeting or not meeting regu1atory;management targets and, in some cases not implementing clearly justified - needed improvements in the quality - reliability of an engineered system. Perhaps the most severe limitation to the proactive approaches regards “knowability”. One can only analyse what one can or does know. Predictability and knowability are the foundation blocks of the quantitative analytical models [Apostolakis, et a1 1990; Rasmussen, 1996; Center for Chemical Process Safety, 1989; Spouge, 19991. But, what about the unknowable and the unpredictable? Can we really convince ourselves that we can project into the future of offshore structure systems and perform analyses that can provide sufficient insights to enable us to implement the measures required to fully assure their quality and reliability? Or are some other processes and measures needed? This fundamental property of the unknowability has some extremely important ramifications with regard to application of the ALARP principle [Melchers, 1993; Hessami, 19991. We can ALARP only what we recognise and this has proven to be extremely limited when it comes to very low probability - high consequence events that have their sources in human and organisational factors. The author has concern for some proactive reliability based analyses that have been and are being used to define IMR (Inspection, Maintenance, Repair) programs for offshore structures [Bea, 19921. Such analyses can only address the knowable and predictable aspects that influence IMR programs (e.g. fatigue damage at brace joints). Such analyses are frequently used to justify reductions in IMR program frequencies, intensities and costs [Faber, 1997; Soares, 1998; Marine Technology Directorate, 1989, 19921. But what about the unknowable and the unpredictable elements that influence the IMR programs? We look for cracks where we do not find them and we find them where we do not look for them [Bucknell, 20001. What about the host of major “biases” (differences between reality and the calculated results) that exert major influences on the results that come from such
- 312. Chapter 12 z 962 analyses [Xu, et a1 1999]? These elements are frequently referred to as being founded in “gross errors” [Marine Technology Directorate, 1989; Bea, 19921. Experience has adequately demonstrated that a very large amount, if not the majority of the defects and damages we encounter in offshore structures are not in any reasonable or practical sense “predictable” [Marine Technology Directorate, 1994; Winkworth and Fisher, 1992; Bucknell, 2000; De Leon and Heredia-Zavoni, 20011. Other approaches (e.g. inductive information based) must be used to address the unknowable zyx - unpredictable aspecrs that still must be managed in the operations of offshore structures. Studies of the HROs (Higher Reliability Organisations) have shed some light on the factors that contribute to errors made by organisations and risk mitigation in HRO. The HROs are those organisations that have operated nearly error-free over long periods of time. z A wide variety of HROs have been studied over long periods of time. The HRO research has been directed to define what these organisations do to reduce the probabilities of serious errors. The work has shown that the reduction in error occurrence is accomplished by the following [Roberts, 1989, 1993; Weick, 1995; Weick, et a1 19991: (1) command by exception or negation, (2) redundancy (robustness - defect and damage tolerance), (3) procedures and rules, (4) selection and training, zyxw (5) appropriate rewards and punishment and (6) ability of management to “see the big picture”. Command by exception (management by exception) refers to the management activity in which the authority is pushed to the lower levels of the organisation by managers who constantly monitor the behaviour of their subordinates. Decision-making responsibility is allowed to migrate to the persons with the most expertise to make the decision when unfamiliar situations arise (employee empowerment). Redundancy involves people, procedures and hardware. It involves numerous individuals who serve as redundant decision-makers. There are multiple hardware components that will permit the system to function when one of the components fails. The term redundancy is directed towards the identification of the need for organisational “robustness” - damage and defect tolerance that can be developed, given proper configuration (deployment), ductility - ability and willingness to shift demands, excess capacity (ability to carry temporary overloads) and appropriate correlation (low for parallel elements. high for series elements). Procedures that are correct, accurate, complete, well organised, well documented and are not excessively complex are an important part of an HRO. Adherence to the rules is emphasised as a way to prevent errors, unless the rules themselves contribute to error. The HROs develop constant and high quality programs of personnel selection and training. Personnel selection is intended to select people that have natural talents for performing the tasks that have to be performed. Training in the conduct of normal and abnormal activities is mandatory to avoid errors. Training in how to handle unpredictable and unimaginable unraveling of systems is also needed. Establishment of appropriate rewards and punishment that are consistent with the organisational goals is critical; incentives are a key to performance. An HRO’s organisational structure is defined as one that allows the key decision-makers to understand the big picture. These decision-makers with the big picture perceive the
- 313. Design for Reliabiiit): Human zyxwvutsr and Organisational Factors zyxwvu 963 important developing situations, properly integrate them, and then develop high reliability responses. In a recent organisational research performed by Libuser (1994), five prominent failures were addressed including the Chernobyl nuclear power plant, the grounding of the Exxon Valdez, the Bhopal chemical plant gas leak, the mis-grinding of the Hubble Telescope mirror and the explosion of the space shuttle, Challenger. These failures were evaluated in the context of five hypotheses that defined risk mitigating and non-risk mitigating organisations. The failures provided support for the following five hypotheses: Risk mitigating organisations will have extensive process auditing procedures. Process auditing is an established system for ongoing checks designed to spot expected as well as unexpected safety problems. Safety drills would be included in this category as would be equipment testing. Follow-ups on problems revealed in prior audits are a critical part of this function. Risk mitigating organisations will have reward systems that encourage risk mitigating behaviour on the part of the organisation, its members and constituents. The reward system is the payoff that an individual or organisation gets for behaving in one way or another. It is concerned with reducing risky behaviour. Risk mitigating organisations will have quality standards that exceed the referent standard of quality in the industry. Risk mitigating organisations will correctly assess the risk associated with the given problem or situation. Two elements of risk perception are involved. One is whether or not there was any knowledge that risks existed at all. The second is if there was knowledge that risk existed, the extent to which it was understood sufficiently. Risk mitigating organisations will have a strong command and control system consist- ing of five elements: (a) migrating decision making, (b) redundancy, (c) rules and proce- dures, (d) training and (e) senior management has the big picture. These concepts have been extended to characterise how organisations can organise to achieve high quality and reliability. Effective HROs are characterised by [Weick, et a1 1999; Weick and Quinn, 1999; Weick and Sutcliffe, 20011: Preoccupation with failure zyxwvu ~ any and all failures are regarded as insights on the health of a system, thorough analyses of near-failures, generalise (not localise) failures, encourage self-reporting of errors and understand the liabilities of successes. Reluctance to simplify interpretations ~ regard simplifications as potentially dangerous because they limit both the precautions people take and the number of undesired consequences they envision. respect what they do not know, match external complex- ities with internal complexities (requisite variety), diverse checks and balances, encourage a divergence in analytical perspectives among members of an organisation (it is the divergence, not the commonalties, that hold the key to detecting anomalies). Sensitivity to operations - construct and maintain a cognitive map that allows them to integrate diverse inputs into a single picture of the overall situation and status (situational awareness, “having the bubble”), people act thinkingly and with heed.
- 314. 964 zyxwvutsrqpo Chapter z 12 redundancy involving cross-checks, doubts that precautions are sufficient, and wariness about claimed levels of competence, exhibit extraordinary sensitivity to the incipient overloading of any one of it members, sensemaking. Commitment to resilience zyxwvu - capacity to cope with unanticipated dangers after they have become manifest, continuous management of fluctuations, prepare for inevitable surprises by expanding the general knowledge, technical facility, and command over resources, formal support for improvisation (capability to recombine actions in repertoire into novel successful combinations), and simultaneously believe and doubt their past experience. Under-specification of structures - avoid the adoption of orderly procedures to reduce error that often spreads them around, avoid higher level errors that tend to pick up and combine with lower level errors that make them harder to comprehend and more interactively complex, gain flexibility by enacting moments of organised anarchy, loosen specification of who is the important decision maker in order to allow decision making to migrate along with problems (migrating decision making), move in the direction of a garbage can structure in which problems, solutions, decision makers and choice opportunities are independent streams flowing through a system that become linked by their arrival and departure times and by any structural constraints that affect which problems, solutions and decision makers have access to which opportunities. The other side of this coin is LROs (Lower Reliability Organisations). The studies show that these non-HROs are characterised by a focus on success rather than failure, and efficiency rather than reliability [Weick, et a1 1999; Weick and Sutcliffe, 20011. In a non-HRO, the cognitive infrastructure is underdeveloped, failures are localised rather than generalised, and highly specified structures and processes are put in place that develop inertial blind spots that allow failures to cumulate and produce catastrophic outcomes. The LROs have little or no robustness, have little or no diversity; they have focused conformity. Efficient organisations practice stable activity patterns and unpredictable cognitive processes that often result in errors; they do the same things in the face of changing events, these changes go undetected because people are rushed, distracted, careless or ignorant [Weick and Quinn, 19991.In the non-HRO expensive and inefficient learning and diversity in problem solving are not welcomed. Information, particularly “bad” or “useless” information is not actively sought, failures are not taken as learning lessons and new ideas are rejected. Communications are regarded as wasteful and hence the sharing of information and interpretations between individuals is stymied. Divergent views are discouraged, so that there is a narrow set of assumptions that sensitise it to a narrow variety of inputs. In the non-HRO, success breeds confidence and fantasy, managers attribute success to themselves, rather than to luck, and they trust procedures to keep them appraised of developing problems. Under the assumption that success demonstrates competence, the non-HRO drifts into complacency, inattention, and habituated routines, which they often justify with the argument that they are eliminating unnecessary effort and redundancy. Often downsizing and out-sourcing are used to further the drives of efficiency and
- 315. Design for Reliability: Human and Organisational Factors zyxwvu 965 insensitivity is developed to overloading and its effects on judgement and performance. Redundancy (robustness or defect tolerance) is eliminated or reduced in the same drive resulting in the elimination of cross-checks, assumption that precautions and existing levels of training and experience are sufficient, and dependence on claimed levels of competence. With outsourcing, it is now the supplier, not the buyer that must become preoccupied with failure. But, the supplier is preoccupied with success, not failure, and because of low-bid contracting, often is concerned with the lowest possible cost success. The buyer now becomes more mindless and if novel forms of failure are possible, then the loss of a preoccupation with failure makes the buyer more vulnerable to failure. The non-HROs tend to lean towards anticipation of “expected surprises”, risk aversion and planned defences against foreseeable accidents and risks; unforeseeable accidents and risks are not recognised or believed. Reason (1997) in expanding his work from the individual [Reason, 19901 to the organisation, develops another series of important insights and findings. Reason observes that all technological organisations are governed by two primary processes: production and protection. Production produces the resources that make protection possible. Thus, the needs of production will generally have priority throughout most of an organisation‘s life, and consequently, most of those that manage the organisation will have skills in production, not protection. It is only after an accident or a near-miss that protection becomes for a short time period paramount in the minds of those that manage an organisation. Reason observes that production and protection are dependent on the same underlying organisational processes. If priority is given to production by management and the skills of the organisation are directed to maximising production, then unless other measures are implemented, one can expect an inevitable loss in protection until significant accidents cause an awakening of the need to implement protective measures. The organisation chooses to focus on problems that it always has (production) and not on problems it almost never has (major accidents). The organisation becomes “habituated” to the risks it faces and people forget to be afraid: “chronic worry is the price of quality and reliability” [Reason, 19971. zyxwvu 12.4.2 Reactive Approaches The reactive approach is based on the analysis of the failure or near failures (incidents, near-misses) of a system. An attempt is to made to understand the reasons for the failure or near-failures, and then to put measures in place to prevent future failures of the system. The field of worker safety has largely developed from the application of this approach. This attention to accidents, near-misses and incidents is clearly warranted. Studies have indicated that generally there are about 100+ incidents, 10-100 near-misses, to every accident [Hale, et a1 1997; Rassmussen, et a1 19871. The incidents and near-misses can give early warnings of potential degradation in the safety of the system. The incidents and near-misses, if well understood and communicated provide important clues as to how the system operators are able to rescue their systems, returning them to a safe state and to the potential degradation in the inherent safety characteristics of the system. We have come to understand that responses to accidents and incidents can reveal much more about maintaining adequate quality and reliability than responses associated with successes.
- 316. 966 zyxwvutsrqpo Chapter zy 12 Well-developed guidelines have been developed for investigating incidents and performing audits or assessments associated with near-misses and accidents [Center for Chemical Process Safety, 1992; Hale, et a1 19971. These guidelines indicate that the attitudes and beliefs of the involved organisations are critical in developing successful reactive processes and systems, particularly doing away with “blame and shame” cultures and practices. It is further observed that many if not most systems focus on “technical causes” including equipment and hardware. Human system failures are treated in a cursory manner and often from a safety engineering perspective that has a focus on outcomes of errors (e.g. inattention, lack of motivation) and statistical data (e.g. lost-time accidents) [Reason, 1997; Fischoff, 19751. Most important, most reactive processes completely ignore the organisational malfunctions that are critically important in contributing to and compounding the initiating events that lead to accidents [Reason, 19971.Finding “well-documented’’ failures is more the exception than the rule. Most accident investigation procedures and processes have been seriously flawed. The qualifications, experience and motivations of the accident assessors are critical; as are the processes that are used to investigate, assess and document the factors and events that developed during the accident. A wide variety of biases “infect” the investigation processes and investigators (e.g. confirmational bias, organisational bias, reductive bias) [Reason, 1997; Fischoff, 19751. In the author’s direct involvement with several major failures of offshore structures (casualties whose total cost exceeds U.S. $1 billion each), the most complete information develops during the legal, regulatory induced and insurance investigation proceedings. Many of these failures are “quiet”. Fires and explosions (e.g. Piper Alpha), sinkings (e.g. Petrobras P-36) and collisions/groundings (e.g. Exxon Valdez) are “noisy” and often attract media, regulatory and public attention. Quiet failures on the other hand are not noisy; in fact, many times overt attempts are made to “keep them quiet”. These quiet failures frequently are developed during the design and/or construction phases. These represent offshore structure “project failures”. The author recently has worked on two major quiet failures that involved the international EPC (Engineering, Procurement, Construction) offshore structure project failures that developed during construction. zyxwv A third major failure involved an EPCO (add Operation) project that failed when the system was not able to develop the quality and reliability that had been contracted for. In both these cases, the initial “knee jerk” reaction was to direct the blame at “engineering errors” and a contended “lack of meeting the engineering standard of practice”. Upon further extensive background development (taking 2 and 3 yr of legal proceedings), the issues shifted from the engineering “operating teams” to the “organisational and management” issues. Even though “partnering” was a primary theme of the formation of the contractors and contracting, in fact partnering was a myth. Even though IS0 certifications were required and provided, the IS0 QA/QC guidelines were not followed. The international organisations involved in the work developed severe “cultural conflicts” and communication breakdowns. Promises were made and not honoured; integrity was compromised. Experienced personnel were promised and not provided (“bait and switch”). There was a continually recurring theme of trying to get something: everything for nothing or next to nothing. As ultimately judged in the courts, these failures
- 317. Design for Reliability: Human and Organisationai Factors zyxwvu 961 z were firmly rooted in organisational malfunctions, not engineering malfunctions. The problem with most legal proceedings is that it is very rare that the results are made public. Thus, the insights important to the engineering profession is largely lost, and in some cases, seriously distorted. A primary objective of incident reporting systems is to identify recurring trends from the large numbers of incidents with relatively minor outcomes. The primary objective of near- miss systems is to learn lessons (good and bad) from operational experiences. Near-misses have the potential for providing more information about the causes of serious accidents than accident information systems. The near-misses potentially include information on how the human operators have successfully returned their systems to the safe states. These lessons and insights should be reinforced to better equip operators to maintain the quality of their systems in the face of unpredictable and unimaginable unraveling of their systems. A root cause analysis is generally interpreted to apply to systems that are concerned with the detailed investigations of accidents with major consequences. The author has a fundamental objection to root cause analysis because of the implication that there is a single cause at the root of the accident (reductive bias) [Center for Chemical Process Safety, 19941.This is rarely the case. This is an attempt to simplify what is generally a very complex set of interactions and factors, and in this attempt, the lessons that could be learned from the accident are frequently lost. Important elements in a root cause analysis include an investigation procedure based on a model of accident causation. A systematic framework is needed so that the right issues are addressed during the investigation [Hale, et zy a1 1997; Bea, et a1 19961. There are high priority requirements for comprehensiveness and consistency. The comprehensiveness needs to be based on a systems approach that includes error tendencies, error inducing environments, multiple causations, latent factors and causes and organisational influences. The focus should be on a model of the system factors so that error reduction measures and strategies can be identified. The requirement for consistency is particularly important if the results from multiple accident analyses are to be useful for evaluating trends in underlying causes over time. There is no shortage of methods to provide a basis for a detailed analysis and the reporting of incidents, near-misses and accidents. The primary challenge is to determine how such methods can be introduced into the life cycle risk assessment and management (RAM) of offshore structures and how their long-term support can be developed (business incentives). Inspections during construction, operation, and maintenance are a key element in reactive RAM approaches. Thus, development of IMR (Inspection, Maintenance, Repair) programs is a key element in the development of reactive management of the quality and reliability of offshore structures [Bea, 19921. Deductive methods involving mechanics- based SRA/PRA/QRA techniques have been highly developed [Faber, 1997; Spouge, 1999; Soares, 19981.These techniques focus on “predictable” damage that is focused primarily on durability; fatigue and corrosion degradations. The inductive methods involving discovery of defects and damage are focused primarily on “unpredictable” elements that are due primarily to unanticipated HOE such as weld flaws, fit-up or alignment defects, dropped objects, ineffective corrosion protection and collisions. Reliability centre maintenance (RCM) approaches have been developed and are continuing to be developed to help address both predictable and unpredictable damage and defects [Jones, 19951. Some very
- 318. 968 zyxwvutsrqpo Chapter zy 12 significant forward strides have been made in the development and implementation of life cycle IMR database analysis and communications systems. But, due to expense and cost concerns, and unwillingness or inability of the organisation to integrate such systems into their business systems, much of this progress has been short lived. The reactive approach has some important limitations. It is not often that one can truly understand the causes of accidents. If one does not understand the true causes, how can one expect to put the right measures in place to prevent future accidents? Further, if the causes of accidents represent an almost never-to-be repeated collusion of complex actions and events, then how can one expect to use this approach to prevent future accidents? Further, the usual reaction to accidents has been to attempt to put in place hardware and equipment that will help prevent the next accident. Attempts to use equipment and hardware to fix what are the basic HOF problems generally have not proven to be effective [Reason, 19971. It has been observed that progressive application of the reactive approach can lead to decreasing the accepted “safe” operating space for operating personnel through increased formal procedures to the point where the operators have to violate the formal procedures to operate the system. zyxwvu 12.4.3 Interactive Approaches Experience with developing acceptable and desirable quality and reliability of offshore structures indicates that there is a third important approach that needs to be recognised and further developed. Until recently, it was contended that there were only proactive and reactive approaches [Rasmussen, 1996: Rasmussen, et a1 19871. The third approach is interactive (real-time) engineering and management in which danger or hazards builds up in a system and it is necessary to actively intervene with the system to return it to an acceptable quality and reliability state. This approach is based on the contention that many aspects that influence or determine the failure of offshore structures in the future are fundamentally unpredictable and unknowable. These are the incredible, unbelievable, complex sequences of events and developments that unravel a system until it fails. We want to be able to assess and manage these evolving disintegrations. This approach is based on providing systems (including the human operators) that have enhanced abilities to rescue themselves. This approach is based on the observation that people more frequently return systems to safe states than they do to the unsafe states that result in accidents. Engineers can have important influences on the abilities of people to rescue systems and on the abilities of the systems to be rescued by providing adequate measures to support and protect the operating personnel and the system components that are essential to their operations. Quality assurance and quality control (QAiQC) is an example of the real-time approach [Matousek, 19901. QA is done before the activity, but QC is conducted during the activity. The objective of the QC is to assure that what was intended is actually being carried out. Two fundamental approaches to improving interactive performance are: (1) providing people support and (2) providing system support. People-support strategies include such things as selecting personnel well suited to address challenges to acceptable performance, and then training them so they possess the required skills and knowledge. Re-training
- 319. Design for Reliabilitj: Human and Organisational Factors zyxwvu 969 is important to maintain skills and achieve vigilance. The cognitive skills developed for interactive RAM degrade rapidly if they are not maintained and used [Weick, 1995; Klein, 1999; Knoll, 1986; Weick and Sutcilffe, 20011. Interactive teams should be developed that have the requisite variety to recognise and manage the challenges to quality and reliability and have developed teamwork processes so the necessary awareness, skills and knowledge are mobilised when they are needed. Auditing, training and re-training are needed to help maintain and hone skills, improve knowledge and maintain readiness [Center for Chemical Process Safety, 19931. The interactive RAM teams need to be trained in problem “divide and conquer” strategies that preserve situational awareness through organisation of strategic and tactical commands and utilisation of “expert task performance” (specialists) teams [Klein, 19991. Interactive teams need to be provided with practical and adaptable strategies and plans that can serve as useful “templates” in helping manage each unique crisis. These templates help reduce the amount and intensity of cognitive processing that is required to manage the challenges to quality and reliability. An improved system support includes factors such as improved maintenance of the necessary critical equipment and procedures so they are workable and available as the system developments unfold. Data systems and communications systems are needed to provide and maintain accurate, relevant and timely information in “chunks” that can be recognised, evaluated and managed. Adequate “safe haven” measures need to be provided to allow interactive RAM teams to recognise and manage the challenges without major concerns for their well being. Hardware and structure systems need to be provided to slow the escalation of the hazards, and re-stabilise the system. One would think that the improved interactive system support would be highly developed by engineers. This does not seem to be the case [Kletz, 19911.A few practitioners recognise its importance, but generally it has not been incorporated into general engineering practice or guidelines. Systems that are intentionally designed to be stabilising (when pushed to their limits, they tend to become more stable) and robust (sufficient damage and defect tolerance) are not usual. Some provisions have been made to develop systems that slow the progression of some system degradations. Effective early warning systems and “status” information and communication systems have not received the attention they deserve in providing system support for interactive RAM. Systems need to be designed to clearly and calmly indicate when they are nearing the edges of safe performance. Once these edges are passed, multiple barriers need to be in place to slow further degradation and there should be warnings of the breaching of these barriers. More work in this area is definitely needed. Reason (1997) suggested that latent problems with insufficient quality (failures, accidents) in technical systems are similar to diseases in the human body: “Latentfailures in technical systems are analogous to resident pathogens in the human body which combine with local triggering factors (i.e. life stresses, toxic chemicals and the like) to overcome the immune system and produce disease. Like cancers and cardiovascular disorders, accidents in defended systems do not arisefiom single causes. They occur because of the adverse conjunction of severalfactors, each one necessary but
- 320. 910 Chapter z 12 not sufficient to breach the defenses. zyxwvu As in the case of the human body, all technical systems will have some pathogens lying dormant M.ithin them”. zyx Reason developed eight assertions regarding the error tolerance in complex systems in the context of offshore structures: zyxwv e The likelihood of an accident is a function of the number of pathogens within the system. The more complex and opaque the system, the more pathogens it will contain. Simpler, less well-defended systems need fewer pathogens to bring about an accident. The higher a person’s position within the decision-making structure of the organisation, the greater is his or her potential for spawning pathogens. Local pathogens or accident triggers are hard to anticipate. Resident pathogens can be identified proactively, given adequate access and system knowledge. Efforts directed at identifying and neutralising pathogens are likely to have more safety benefits than those directed at minimising active failures. Establish diagnostic tests and signs, analogous to white cell counts and blood pressure, that give indications of the health or morbidity of a high hazard technical system. The single dominant cause of structure design-related failures has been errors committed, contributed, and/or compounded by the organisations that were involved in and with the designs. At the core of many of these organisation-based errors was a culture that did not promote quality and reliability in the design process. The culture and the organisations did not provide the incentives, values, standards, goals, resources and controls that were required to achieve adequate quality. Loss of corporate memory also has been involved in many cases of structure failures. The painful lessons of the past were lost and the lessons were repeated with generally even more painful results. Such loss of corporate memory are particularly probable in times of down-sizing, out-sourcing and mergers. The second leading cause of structure failures is associated with the individuals that comprise the design team. Errors of omission and commission, violations (circumventions), mistakes, rejection of information and incorrect transmission of information (commu- nications) have been the dominant causes of failures. Lack of adequate training, time and teamwork or back-up (insufficient redundancy) has been responsible for not catching and correcting many of these errors [Bea, 2000bl. The third leading cause of structure failures has been errors embedded in procedures. The traditional and established ways of doing things when applied to structures and systems that “push the envelope” have resulted in a multitude of structure failures. There are many cases where such errors have been embedded in design guidelines and codes and in computer software used in design. Newly developed, advanced and frequently very complex design technology applied in the development of design procedures and the design of offshore structures have not been sufficiently debugged and failures (compromises in quality) have resulted. Next Page
- 321. Design for Reiiabilitj: Human nnd Organisational Faciors zyxwvu 971 This insight indicates the priorities of where one should devote attention and resources if one is interested in improving and assuring sufficient quality in the design of offshore structures [Bea, 2000bl: (1) organisations (administrative and functional structures), (2) operating teams (the design teams), (3) procedures (the design processes and guidelines), (4) robust structures and (5) life cycle engineering of “human friendly” structures that facilitate construction, operation, maintenance and decommissioning. Formalised methods of QA/QC take into account the need to develop the full range of quality attributes in the offshore structure including serviceability, safety, durability and compatibility. QA is the proactive element in which planning is developed to help preserve desirable quality. QC is the interactive element in which planning is implemented and carried out. QA/QCmeasures are focused both on error prevention and error detection and correction [Harris and Chaney, 19691.There can be a real danger in excessively formalised QAiQC processes. If not properly managed, they can lead to a self-defeating generation of paperwork, waste of scarce resources that can be devoted to QAIQC and a minimum compliance mentality. In design. adequate QC (detection, correction) can play a vital role in assuring the desired quality is achieved in an offshore structure. Independent, third-party verification, if properly directed and motivated, can be extremely valuable in disclosing embedded errors committed during the design process. In many problems involving insufficient quality in offshore structures, these embedded errors have been centred in fundamental assumptions regarding the design conditions and constraints and in the determination of loadings or demands that will be placed on the structure. These embedded errors can be institutionalised in the form of design codes. guidelines and specifications. It takes an experienced outside viewpoint to detect and then urge the correction of such embedded errors [Klein, 19991. The design organisation must be such that identification of potential major problems is encouraged; the incentives and rewards for such detection need to be provided. It is important to understand that adequate correction does not always follow detection of an important or significant error in the design of a structure. Again, QA/QC processes need to adequately provide for correction after detection. Potential significant problems that can degrade the quality of a structure need to be recognised at the outset of the design process and measures provided to solve these problems if they occur. A study of the offshore structure design errors and the effectiveness of QAiQC activities in detecting and correcting such errors leads to the checking strategies summarised in table 12.6. The structure design checking studies performed by Knoll (1986), the series of studies performed by Stewart and Melchers (1988), and the studies performed during this research indicate that there is one part of the design process that is particularly prone to errors committed by the design team. That part of the process is the one that deals with the definition of design loadings that are imposed on and induced in the structure. This recognition has several implications with regard to managing HOF in design. The first implication regards the loading analysis procedures themselves. The second implication regard the education and training of structure design engineers in the development and performance of loading analyses. Given the complexities associated with performing Previous Page
- 322. 912 What to check? zyxwvut - zyxw high likelihood of error parts (e.g. assumptions, loadings, documentation) I- high consequence of error parts Chapter z 12 How zyxw to check? - direct towards the important parts of the structure (error intolerant) - be independent from circumstances which lead to generation of the Table 12.6 Structure design QA/QC When to check? - before design starts (verify process, qualify team) - during concept development - periodically during remainder - after design documentation completed of process design - use qualified and experienced engineers - provide sufficient QA/QC resources - assure constructability and IMR Who to check? ~ the organisations most prone to errors ~ the design teams most prone to errors - the individuals most prone to errors loading analyses, the complexities associated with the loading processes and conditions and the close coupling between the structure response and the loading environment, it is little wonder that loading analyses are probably the single largest source of structure design errors. What is somewhat disturbing is that many designers of offshore structures do not understand these complexities nor have been taught how to properly address them in structure design. The third implication regards the need for independent (of the situations that potentially create errors), third-party QA and QC checking measures that are an integral part of the offshore structure design process. This checking should start with the basic tools (guidelines, codes, programs) of the structure design process to assure that “standardised errors” have not been embedded in the design tools. The checking should extend through the major phases of the design process, with a particular attention given to the loading analysis portions of that process. Computer programs used to perform analyses for design of critical parts of the structure should be subjected to verifications and these analyses repeated using independently developed programs. The intensity and the extent of the design-checking process needs to be matched to the particular design situation. Repetitive designs that have been adequately tested in operations to demonstrate that they have the requisite quality do not need to be verified and checked as closely as those that are “first-offs” and “new designs” that may push the boundaries of current technology. The elements of organisational sensemaking are critical parts of an effective QA/QC process, and in particular, the needs for requisite variety and experience. There is a need for background and experience in those performing the QA/QC process that matches the complexity of the design being checked. Provision of adequate resources and motivations are also necessary, particularly the willingness of management and engineering to provide integrity to the process and to be prepared to deal adequately with “bad news”.
- 323. Design for Reliabilitj,’ Human and Organisational Factors zyxwvu 913 z 12.5 Instruments to Help Achieve Design Success zyxw Two instruments will be discussed in the remainder of this chapter that have been developed recently to help promote more effective application of proactive, reactive and interactive processes during the life cycle of offshore structures. A development of these two instruments have been concentrated on taking full advantage of the progress cited in this chapter while addressing some of the major limitations that have been recognised. The first instrument (computer program, application protocol) is identified as a Quality Management Assessment System (QMASO);this is fundamentally a qualitative process to help guide assessment teams to examine the important parts of offshore structure systems at different times during their life cycle. These assessment teams include members of the offshore structure design engineering team system being assessed. The instrument has been designed to elicit the insights and information that only these people can have. The second instrument (computer program, application protocol) is a System Risk Analysis System (SYRASO); this is a PRAiQRA/SRA instrument to help develop quantitative results that are often required by engineers and managers. Traditional event tree and fault tree analysis methods have been used in SYRAS. The analytical templates in SYRAS enable the analyses of each of the life cycles of an offshore structure and address each of the quality attributes. A “link” has been developed between the results from the QMAS and the input required for the SYRAS instrument. This link is based on translating the “grades” developed from application of QMAS to performance shaping factors (PSF) that are used to modify normal rates of human,/operator team malfunctions. The link has been developed, verified and calibrated from the QMAS-SYRAS analyses of failures and successes of offshore structure systems during their different life cycle phases [Bea, 2000a, b]. 12.5.1 Quality Management Assessment System The QMAS is a method that is intended to provide a level of detail between the qualitative/ less-detailed methods (e.g. HazOps, FMEA) and the highly quantitativeivery detailed methods (PRA, QRA). The QMAS encompasses two levels of safety assessment: coarse and detailed qualitative. The objective of the QMAS is with the least effort possible, to identify those factors that are not of concern relative to quality and reliability, to identify those mitigation measures that need to be implemented to improve quality and reliability and to identify those factors that are of concern that should be relegated to more detailed quantitative evaluations and analyses. Components The QMAS system is comprised of three primary components: (1) a laptop computer program and documentation that is used to help guide platform assessments and record their results, zyxwvuts (2) an assessor qualification protocol and training program, and (3) a three- stage assessment process that is started with information gathering and identification on factors of concern (FOC), then proceeds to observe operations, and is concluded with a final assessment and set of recommendations.
- 324. 914 zyxwvutsrqpon Chapter 12 z The surveying instrument is in the form of a laptop computer program that contains interactive algorithms to facilitate development of consistent and meaningful evaluations of existing facilities. The instrument includes evaluations of the categories of facility factors defined earlier: the operating personnel, organisations, hardware (equipment, structure), procedures (normal, emergency), environments, and interfaces between the categories of factors. A standardised and customised written, tabular and graphical output reporting and routines are provided. This instrument is intended to help identify alternatives for how a given facility might best be upgraded so that it can be fit for the intended purposes. The zyxwvutsr QMAS process has been developed so that it can be used effectivelyand efficiently by those that have daily involvement and responsibilities for the quality and reliability of offshore structures. The QMAS system is intended to help empower those that have such responsibilities to identify important potential quality and reliability degradation hazards, prioritise those hazards, and then define warranted or needed mitigation measures. Evaluation Steps There are five major steps in the QMAS. Step #1 is to select a system for assessment. This selection would be based on an evaluation of the history of quality and reliability degradation events and other types of high-consequence accidents involving comparable systems, and the general likelihood and consequences of potential quality and reliability degradations. Step #2 is to identify an assessment team. This team is comprised of qualified and trained QMAS assessors indicated as designated assessment representatives (DARs). These DARs normally come from the organisation/s and operation/s being assessed, regulatory or classification agencies, and/or consulting engineering service firms. DAR appointment is based on technical and operations experience. Integrity, credibility and deep knowledge are the key DAR qualification attributes. The DARs are qualified based on QMAS specific training and experience that includes development of in-depth knowledge of human and organisational factors and their potential influences on the quality and reliability of offshore structure systems. To avoid conflicts of interest, the DARs are allowed to request replacement by when such conflicts arise. It is desirable that the assessment teams include members of management and operations/engineering. The DAR teams include experienced “outsiders” (counsellors) who have extensive HOF background and QMAS applications experience. Step #3 consists of a coarse qualitative assessment of the seven categories of elements that comprise an offshore structure system. This assessment is based on the general history of similar types of facilities and operations and details on the specific system. These details would consist of current information on the structure, equipment, procedures (normal operations and maintenance, and emergency/crisis management), operating personnel (including contractors), and organisations/management. Discussions would be held with representatives of the operator/system organisation and the operating/engineering teams. The product of Step #3 is identification of the FOC that could lead to degradations in quality and reliability of an offshore structure. As a part of the assessment process that will be described later, the assessment team records the rationale for identification of the FOC. The assessment may at this stage also identify suggested mitigation. The results are
- 325. Design zyxwvutsrqponm for Reliabilifj: Human and Organisational Factors zyxwvu 915 z reported in user-selected standard textural and graphical formats and in user-defined textural and graphical formats (that can be stored in the computer or produced each time). For some systems, the information at this stage may be sufficient to allow the system to exit the QMAS with the implementation of the mitigation, recording the results and scheduling the next assessment. If it is deemed necessary, the QMAS proceeds to Step #4; development of scenario/s to express and evaluate the FOC. These scenarios or sequences of events are intended to capture the initiating, contributing and compounding events that could lead to degradations in quality and reliability. These scenarios help focus the attention of the assessors on specific elements that could pose high risks to the system. Based on the FOC and the associated scenarios, Step #5 proceeds with a detailed qualitative assessment. Additional information is developed to perform this assessment and includes more detailed information on the general history of the structure system, its details, results from previous studies, and management and operating personnel interviews. In recording results from the interviews, provisions are made for anonymous discussions and reporting. The product of Step #5 is a detailing of the mitigation measures suggested for mitigation of the FOC confirmed in Step #5. The rationale for the suggested mitigation are detailed together with projected beneficial effects on the FOC. As for the results of Step #3, the results of Step #4 are reported in standard and user-defined formats. At this point, the assessment team could elect to continue the QMAS in one of the two ways. The first option would be to return to the FOC stage and repeat Step #5-based “new” FOC and the associated scenarios. The second option would be to proceed wlth some of the FOC and the associated scenarios into coarse quantitative analyses and evaluations. If the assessment team is elected, the QMAS could be terminated at the end of Step zyx #5. The results would be recorded, and the next assessment scheduled. Evaluation Processes The QMASevaluation is organised into three sections or “Levels” (fig. 12.5). The first level identifies each of the seven structure system components: 1.0 zyx - operators, 2.0 - organisations, 3.0 - procedures, 4.0 - equipment, 5.0 - structure, 6.0 - environments and 7 -interfaces CTORS - graded 1.1 - communications 1.1.8 -feedback, 1.1.9 - no significant barriers zyxwv Figure 12.5 Quality components, factors and attributes
- 326. 916 zyxwvutsrqpon Chapter zy 12 z 7.0 zyxwvuts - interfaces. These seven components comprise “modules” in the QMAS computer program. The structure and equipment factors are modified to recognise the unique characteristics of different offshore structures. The second level identifies the factors that should be considered in developing assessments of the components. For example, for the operators (l.O), seven factors are identified: communications (1.l), selection (1.2), knowledge (1.3), training (1.4), skills (1.5), limitations/impairments (1.6) and organisation/coordination (1.7). If in the judgement of the assessment team, additional factors should be considered, then they can be added. Using a process that will be described later, the assessment team develops grades for each of these factors. The third level identifies attributes associated with each of the factors. These attributes are observable (behaviours) or measurable. These attributes provide the basis or rationale for grading the factors. For example, for the communications factor (1.1) six attributes are included: clarity (1.1.l), accuracy (1.1.2), frequency (1.1.3), openness/honesty (1.1,4), verifying or checking feedback (1.1.5) and encouraging (1.1.6). Again, if in the judgement of the assessment team, additional attributes are needed, they can be added to the QMAS. The factors and the attributes for each of the system components have been based on results from current research on these components with a particular focus on the HOF- related aspects. This approach avoids many of the problems associated with the traditional “question-based” instruments that frequently involve hundreds of questions that may be only tangentially applicable to the unique elements of a given structure system. Factors Grading The QMAS assessment team assigns grades for each component factor and attribute. Three grades are assigned: the most likely, the best, and the worst. These three grades help the assessors express the uncertainties associated with the gradings. Each of the attributes for a given factor are assessed based on a seven-point grading scale (fig. 12.6). An attribute or factor that is average in meeting referent standards and requirements is given a grade of 4. An attribute or factor that is outstanding and exceeds all referent standards and requirements is given a grade of 1. zyxwvu An attribute or factor that is very poor and does not meet any referent standards or requirements is given a grade of 7. Other grades are used to express characteristics that are intermediate to these. The reasons for the attribute and factors grades are recorded by the assessment team members. This process develops a consensus among the system or domain experts, allowing for expressions of dissenting opinions. The grades for the attributes are summed and divided by the number of attributes used to develop a resultant grade for the factor. Weightings of the factors and attributes can be introduced by the assessors. The assessors review the resultant grades and if they are acceptable, the grades are recorded. If it is not, they are revised and the reasons for the revisions noted. The uncertainties associated with the grades for the attributes are propagated using a first order statistical method. In the same manner, the grades for the factors are summed and divided by the number of factors to develop a resultant grade for the component. Again, the assessors review this resultant grade and if it is acceptable. the grade is recorded. If it is not, it is revised
- 327. Design for Reliability: Human and Organisaiional Factors zyxwvu 911 z - z Very poor, does not 7 meet any standards or requirements - 6 Poor 5 Below average 3 2 Excellent Very good >probable 4 meets most standards high bound zyx t most - Good, average, and requirements - - z low bound - Outstanding, 1 exceeds zyxwvu all standards and requirements - Figure 12.6 Scale for grading attributes, factors and components and reasons for the revision noted. The uncertainties associated with the grades for the factors are propagated using a first-order statistical method. A “Braille” chart is then developed that summarises the mean grades (and, if desired, their standard deviations) developed by the assessment team for each of the factors (fig. 12.7). The “high” grades (those above 4) indicate components and the associated factors that are candidates for attention and mitigation. Assessors The most important element in the QMAS system is the team of assessors. It does not matter how good the QMAS assessment instruments and procedures are, if the personnel Figure 12.7 Example component mean grading results
- 328. 978 zyxwvutsrqpo Chapter 12 using the instrument do not have the proper experience, training and motivations. The QMAS assessors must have experience with the system being assessed, quality auditing experience and training in human and organisational factors. The assessor team is comprised of members from the system (operators, engineers, managers, regulators) and z QMAS “counselors” who have extensiveexperiencewith the QMASsystem and operations z - facilities similar to those being assessed. An important aspect of the qualifications of assessors regards their aptitude, attitude, trust and motivation. It is very desirable that the assessors be highly motivated to learn about the human and organisational factors and safety assessment techniques, have high sensitivity to quality hazards (“perverse imaginations”), be observant and thoughtful, have good communication abilities and have a willingness to report “bad news” when it is warranted. It is vital that both the assessors and the QMAS counselor have the trust and respect of the system operators and managers. An assessor “just-in-time” training program has been developed as part of the QMAS instrument. This program includes training in human and organisational factors and the QMAS assessment process. Example applications are used to illustrate applications and to help reinforce the training. zyxwvu A final examination is used to help assure that the assessor has learned the course material and can apply the important concepts. The assessor training program has two parts: (1) informational and (2) practical exercises. The informational part contains background on the QMAS assessment process and computer instrument, failures involving offshore structures and other types of engineered structures, human and organisational performance factors and evaluations. The second part of training is the hands-on use of the computer software. Training exercises are performed to demonstrate the use of the QMAS instrument. Software demonstrations using offshore structures as case studies are walked through. Then the assessors assess another system on their own. Following this, the assessments are compared and evaluated. The assessors are asked for feedback on the QMAS. This approach is identified as a “participatory ergonomics“ approach. The people who participate in the daily activities associated with their portion of the life cycle of a system are directly involved in the evaluations and assessments of that system. These people know their system better than any outsider ever can. Yet, they need help to recognise the potential threats to the quality and reliability of their system. These people provide the memory of what should be done and how it should be done. These are the people who must change and must help their colleagues change so that desirable and acceptable system quality and reliability are developed. This is a job that outsiders can never do or should be expected to do. The QMAS has been applied to a wide variety of offshore structure systems including marine terminals, offshore platforms and ships. QMAS has been applied in proactive assessments (before operations conducted), in reactive assessments (after operations conducted), and in interactive assessments (during conduct of operations). Multiple assessment teams have been used to assess the same system; the results have shown a very high degree of consistency in identification of the primary factors of concern and potential mitigation measures. QMAS has proven to provide a much more complete and
- 329. Desigrijar. Reliability: Huntan and Orgnnisational Factors zyxwvu 919 realistic understanding of the human and organisational elements that comprise offshore structure systems than the traditional PRA,’QRA/SRA approaches [Weick, 20001. Frequently. RAM of an offshore structure system can be conducted solely on the basis of results developed from zyxwvu &MAS, factors important to quality and reliability can be defined and characterised sufficiently to enable effective actions to achieve these objectives. zyxwvuts 12.5.2 System Risk Assessment System The System Risk Assessment System (SYRAS)has been developed to assist engineers in the assessment of system failure probabilities based on the identification of the primary or major tasks that characterise a particular part of the life cycle (design, construction, maintenance, operation) of an offshore structure (fig. 12.8). This RAM instrument has been applied in the study of tradeoffs regarding “minimum” platforms. in quality assurance and quality control (QA,’QC) of the design of innovative deepwater structures, and the effects of Value Improvement Programs for several major offshore structures [Shetty, 2001; Bea, 2000al. The SYRAS instrument consists of a computer program and an applications protocol [Bea, 2000bI. Failures to achieve the desirable quality in an offshore structure can develop from intrinsic ( I ) or extrinsic ( E )causes. Intrinsic causes include factors such as extreme environmental conditions and other similar inherent, natural and professional uncertainties. Extrinsic causes are due to human and organisational factors zyxw - identified here as “human errors”. The probability of failure of the structure to develop quality attribute zy (i), P(Fs,), is (12.23) where (U) is the union of the failure events. The probability of failure of any one of the quality attributes (i) due to inherent randomness is P(Fsir).The probability of failure of any Figure 12.8 SYRAS components
- 330. 980 zyxwvutsrqponm Chapter zy 12 z Magnitude zyxwvu (Y) of Type (X) of HumanError zyxw Figure 12.9 Likelihood of unsatisfactory quality zyx one of the quality attributes zyxwv (i) due to the occurrence of human error is P(FsiE). The probability of human error in developing a quality attribute (i) in the structure is P(Esi). Then: P(Fsi)= zyxw P ( F S i l t E s i ) P ( E s i )+P(FsicI@si)P(gsi) +P(FsiEIEsi)P(Esi) (12.24) The first term addresses the likelihood of structure failure due to inherent causes given a human error (e.g. structure fails in a storm due to damage from a boat collision). The second term addresses the same likelihood given no human error. This is the term normally included in structural reliability analyses. The third term addresses the likelihood of structure failure directly due to human error (e.g. structure fails due to explosions and fire). The probability of failure given HOE, P(Fsla,characterises the "robustness" or defect and damage tolerance of the structure to human errors. The shape of the fragility curve (fig. 12.9)can be controlled by engineering. This is explicit design for robustness or defect (error) tolerance and fail-safe or intrinsically safe design. For the intensities (magnitude) and types of malfunctions that normally can be expected, the structure should be configured and designed so that it does not fail catastrophically (or have unacceptable quality) when these types and magnitude of malfunctions occur. The fragility curve for a particular system is determined using off-line analyses or experimental results and the results input to SYRAS. The probability of no human error is: P(@si)= 1 - P(Esi) (12.25) The probability of insufficient quality in the structure due to HOE, P(FsZE),can be evaluated in the (j)life cycle activities of design ( j = l), construction (j=2), operations (j=3), and maintenance (j=4) as (12.26)
- 331. Design for Reliability. Human and Organisarional Factors zyxwvu or zyxwvutsr 4 P(FsiE) zyxwvu = ~ ( ~ s i j i ~ s O ) ~ ( ~ s i J ) (12.27) j=1 Each of the life cycle activities ( j = 1 -4) can be organised into (n)parts zyx (k= 1 -n): 981 (12.28) This task-based formulation addresses the major functions that are involved in the principal activities that occur during the life cycle of an offshore platform. For example, the system design activity ( j= 1) can be organised into four parts (I? =4): configuration (k= l), system demand analyses (k=2), system capacity analyses (k=3) and documentation (k=4). The likelihood of insufficient quality in the system due to human error during the design activity is (12.29) If desirable, the primary functions or tasks can be decomposed into subtasks to provide additional essential details. The base rates of human errors of type “tn”, P(EkYkm), are based on the published information on human task performance reliability (fig. 12.10) [Center for Chemical Process Safety, 1994; Swain and Guttman, 1983; Kirwan, 1994; Gertman and Blackman. 1994; Kontogiannis and Lucas, 1990; Haber, et a1 1991). Performance Shaping Factors (PSF) are used to modify the base or “normal“ rates of human errors, P(Ekiikwl), to recognise the effects of organisations, structure, equipment, procedures, environments and interfaces: As discussed earlier, gradings from the QMAS component evaluations (Gejkm) are developed on a seven-point scale (fig. 12.6).The mean value and the coefficient of variation of each of the categories of PSF are developed based on an average of the mean values and coefficients of variation of each of the QMAS categories. Evaluation of each of the seven categories of PSF results in a final overall grading zyxw (GEjkm) and coefficient of variation ( V ~ ~ j k ~ ) on this grading that can be used to quantify a specified PSF. Each of the seven PSF (PSF,k,) can act to increase or decrease the base rates of human errors. SYRAS allows the user to specify the base rates and then scale the base rates by multiplying the base rates by the PSF identified by the user. The scales allow the base rates to be increased or decreased by three orders of magnitude. When quantification of the PSF is based on the use of the QMAS instrument and protocol, the PSF is computed from (fig. 12.11): __
- 332. 982 zyxwvutsrqpon Chapter zy 12 z unfamiiar task performedwith change system state without procedures without checking i s p e e d simple task performedwith speed or diverted attention E-2 change system with procedures with checking routine tasks trained, motivated respondto system commands with supervisorysystem zyxw Figure 12.10 Nominal human task performancereliability 1E-3 1E-2 zyxwvu m 1 2 3 4 5 6 7 IQMAS Grade I Figure 12.11 QMAS qualitative grading translation to quantitative PSF used in SYRAS
- 333. Design zyxwvutsr for Reliabilirj: Human zyxwvuts and Organisarional Factors zyxwvu 983 The resultant PSF that modifies the base rate of error is computed from the product of the seven mean PSF: (12.32) The resultant coefficient of variation of the PSF is computed from the square root of the sum of the squares of the PSF coefficients of variation: (12.33) The PSF provides the important link between the qualitative QMAS assessment process and the quantitative PRA-based SYRAS analysis process [Bea, 2000al. Results from QMAS are then “translated” to input that can be used in the traditional PRA/QRA approach embodied in SYRAS. The QMAS-SYRAS link has been based on a repetitive calibration process involving applications of QMAS and SYRAS to offshore structures that have failed (very high probabilities of failure) and succeeded (very low probabilities of failure) [Bea, 2000bl. As would be expected, due to the natural variability in human and organisational performance and the uncertainties associated with the evaluations of such a performance, the PSF have very large coefficients of variation (in range of 100-200%) [Bea, 2000a, 2000bl. The QMAS grades, FOC and system quality improvement recommendations are intended to help capture the processes that cannot be incorporated into highly structured quantitative analyses; these are the dynamic organic processes that characterise most real offshore structure systems. Frequently, the intensive application of the QMAS instrument and the underlying organisational philosophies provide the insights essential to help achieve desirable and acceptable quality and reliability. The coupling of the results from QMAS with the SYRAS probabilities are intended to provide engineers and managers with quantitative assessments of systems so that the effects of potential mitigation measures can be examined and the effects of VIP assessed. Of course, this means that potentially much of the richness of insights provided by QMAS can be lost or obscured by intense attention to the numerical results provided by SYRAS. The best experiences have been those in which both instruments are diligently applied; thus, capturing both qualitative and quantitative insights. Once the tasks are organised into the task structure for the life cycle phase, correlation among elements is assessed. In order to facilitate the calculation of the likelihood of failure, the elements can be designated as either perfectly correlated or perfectly independent. After determining the overall system task structures, the user has the option of analysing the effects of Quality Assurance and Quality Control (QA/QC) on the overall system probability. This is done in an “overlay edit-mode”. This means that the user is able to go back into the task structures and add in the QA/QC procedures as independent tasks with corresponding influences. The user is presented with both the original system Pf and the QA:QC-modified zyxwvu Pf.
- 334. 984 zyxwvutsrqpon Chapter zy 12 z Consequently, the next step in the SYRAS development addresses the HOF malfunction detection (D) and correction (C). This is an attempt to place parallel elements in the quality system so that failure of a component (assembly of elements) requires the failure of more than one weak link. Given the high positive correlation that could be expected in such a system, this would indicate that QA/QC efforts should be placed in those parts of the system that are most prone to error or likely to compromise the intended quality of the system. Conditional on the occurrence of type (m)of HOE, E,,,, the probability that the error gets through the QA,'QC system can be developed as follows: The probability of detection is P(D) and the probability of correction is zyxw P(C).The compliments of these probabilities (not detected and not corrected) are: P(P)= 1 - P(D), and P(a= 1 - P(C) (12.34) The undetected and uncorrected error event, UE,, associated with a human error of type m is: 8 UE, = U( E x zyxwv P , nzyxw G) (12.35) m=l The probability of the undetected and corrected HOE of type m is: (12.36) Assuming independent detection and correction activities or tasks, the probability of the undetected and corrected HOE of type m is The probability of error detection and the probability of error correction play important roles in reducing the likelihood of human malfunctions compromising the system quality. Introduction of QA/QC considerations into the developments into the earlier developments is accomplished by replacing P(Es,/k,) with P(UEsqkm)into the desirable parts of the SYRAS analysis. zyxwvut 12.6 Example Applications 12.6.1 Minimum Structures Results from a joint industry - government sponsored project that addressed the system reliability levels of three minimum structures compared with a standard four-pile jacket recently have become publicly available [Shetty, 20001. The study considered extreme storm, fatigue and ship collision conditions and considered the potential effects on
- 335. Design zyxwvutsr for Reliability Human and Organisational Factors zyxwvu 985 reliability from errors due to human and organisational factors that develop during design, construction and operation of such structures [Bea and Lawson, 19971. The structural concepts considered were a three-pile Monotower, Vierendeel Tower, Braced Caisson and a conventional four-pile Jacket (fig. 12.12). The structures were designed using a common design criteria, analysis and design procedure, and for operation at the same field and to support the same topside operations (RAMBIZILL, 1999). Key members were designed to have utilisation ratios close to 0.8 under the 100-yr return period environmental loading. Welded joints were designed to have minimum fatigue lives of five times the service life (20 yr) for the three minimum structures and three times the service life for the four-pilejacket. The in-place operational conditions, vortex shedding, and on-bottom stability require- ments were considered; reinforcements were made to joint cans and braces to ensure that the structures were able to fully mobilise their capacity during ship impacts. These “minimum” structures were designed to be much more robust (damage-defect tolerant) than their counterparts for the Gulf of Mexico [Bea, et a1 19981. The performance of the four structures under extreme conditions was studied by performing deterministic pushover and system reliability analyses [Gierlinski and Rozmarynowski, 1999; MSL Engineering, 19991(fig. 12.13). Based on the joint probability distributions of wave heights, periods and current parameters, and the ultimate capacity of the structures based on results from the pushover analyses, and accounting for the uncertainties in the calculated hydrodynamic loads and capacities of the four structures, system probabilities of failure were evaluated for each structure. Reliability characteristics for extreme storm conditions also were evaluated for other locations. The reliability under fatigue conditions were evaluated based on the failure of individual joints and the sequences of two, three and four joints assuming that the initialjoint failures were not detected and repaired. The impact of fatigue failure of joints was evaluated by calculating the conditional probability of collapse due to environmental overload given the initial failure of one or more joints by fatigue, and multiplying this with the probability of the fatigue failure sequence occurrence. Time domain, non-linear, ship-structure collision analyses were performed to study the performance of the structures against collisions from supply vessels [MSL Engineering, 19991. Analyses were carried out for a number of vessel mass and velocity combinations, which were considered as credible for operations in the Southern and Central North Sea fields. Following the impacts, a post-impact pushover analysis was performed to determine the reduction in capacity as a result of ship impact damage. A methodology to evaluate the potentials for and effects of human and organisational malfunctions were developed and implemented in the form of two computer programs z - instruments previously identified earlier in this chapter as the QMAS and the SYRAS. Based on the results from a questionnaire circulated to the operators of structures similar to those studied, a review of the world-wide accident database for marine structures, and reported incidents of damage to offshore structures in the North Sea, five error scenarios were identified [Bea and Lawson, 19991.These scenarios addressed errors that could develop during design (fatigue due to pile driving not considered), fabrication (fit-up, welding
- 336. 986 zyxwvutsrqp Figure 12.12 Structures studied (Shetty, 2001) zyxw Chapter 12
- 337. Design for Reliability: Human and Organisational Factors zyxwvu 987 z 0.0 0.5 zyxwvu 1 zyxwvut .o 1.5 2.0 2.5 3.0 Horizontal Deflection at Deck Level (m) zyx Figure 12.13 Results from static non-linear pushover analyses (Shetty, 2000) defects), installation (dented braces due to pile stabbing) and operation (dropped production package, supply boat collision) phase of the structures. For each scenario, the damage to the structures were determined and their reliabilities under the damaged condition were evaluated considering fatigue, extreme storm and ship collision conditions. An example of the evaluation process that was used in this study is that associated with the design phase and the omission of consideration of pile installation-driving-induced stresses. The source of the less-than-desired fatigue durability is due to pile driving stress- induced fatigue in the joints that connect the pile sleeves or guides to the structure. The stress is due to the difficulties associated with maintaining proper alignment of the piles in the underwater pile sleeves or in the caisson pile sleeve connections during installation of the piles. The structures were not designed to sustain the pile driving stresses nor were provisions developed to allow more precise alignment of the piles during pile driving. The structure is fabricated as specified. During installation of the platform, the pile driving- induced stresses cause fatigue cracking to be initiated in the joints of the vertical diagonal braces that connect the pile sleeves/guides to the primary structure elements. This damage leads to through-thickness cracking of several joints. In the case of the three- pile and four-pile monopods, through-thickness cracks are developed in the pile sleeves to vertical diagonal braces that connect to the central column. In the case of the braced caisson, fatigue cracks are developed at the connection between the caisson and the diagonal braces - piles that are driven through the connection. In the case of the four- leg jacket, the piles can be aligned in the legs and driven without imparting significant fatigue damage. The probabilities associated with each of the eight potential causes of this malfunction by the design team during this phase are: communications 5E-4, selection and training 2E-3,
- 338. 988 zyxwvutsrqponm Chapter 12 z planning and preparations 5E-4, limitations and impairments 6E-4, violations 1E-4, slips 1E-4,lack of knowledge 5E-3 and mistakes 1E-3.These probabilities reflect influences from the organisations (direction not provided by ownerjoperator, design contractor, regulatory), procedures (effectsnot included in design guidelines), hardware (no significant influences) and environments (no significant influences). The zyxwvutsr SYRAS analyses indicated that the probability of this HOE scenario is PE1.l =8.9E-3. The dominant causes of the potential malfunctions are ignorance (56%) and selection and training of the members of the design team (22%). The ignorance source error was influenced primarily by lack of organisational communications and defined design procedures to address this problem. There can be sources of correlation between the sources or causes of malfunctions. Such correlation can be developed through organisational influences that embed a specific “culture” in an organisation and result in “group think” biases. In the analysis of HOE, the SYRAS user is able to introduce correlation between the sources of HOE embedded in a task structure. In this case, for the case of perfect positive correlation between the sources or causes of HOE in the design process and team, the probability of the design error would be P~1.1 =5.OE-3. The values of P E ~ . I = 5.OE-3 and P E ~ . ~ =8.9E-3 could be viewed as “bounds” on the possible likelihoods of this specific HOE scenario. For this scenario, two QAjQC alternatives were considered. The first was QA/QC conducted during the design process. Two design process QA/QC alternatives were evaluated. One was a conventional checking of the design analysis calculations. The other was the verification of the design analysis processes by experienced “third-party” design and construction engineers. Based on the results of design process checking cited earlier, the probabilities of detection zyxwv (Po=0.10) and correction (Pc= 0.80) for the first alternative were determined to be P D ~ =0.08. The probability of not detecting and correcting the design HOE was therefore PNDC =0.92. In the second instance, the probabilities of detection (Po=0.80) and correction (Pc=0.90) were determined to be PDc= 0.72. Thus, the probability of not detecting and correction were determined to be PNDC =0.28. The resulting probabilities of design HOE with additional QA/QC measures were thus determined to be P E ~ . ~ A =4.6E-3 to 8.2E-3 for the first QA/QC alternative, and P E ~ I B = 1.4E-3 to 2.5E-3 for the second QAiQC alternative. Results for the QA/QC alternatives are summarised in table 2. Table 12.7 summarises the likelihoods associated with each of the five life cycle scenarios. The base rate likelihoods refer to the condition where the currently specified QA/QC measures were employed. Likelihoods were also developed for additional QA/QC measures representing significant (Alternative A) and major (Alternative B) improvements in these processes. The base rate likelihoods range from 1E-3to 9E-3. The ranges in the likelihoods represents the potential effects of “correlation” between the causes and tasks involved in the HOE scenarios (fig. 12.14). These likelihoods are in good agreement with the database results developed by the Marine Technology Directorate (1994) on platforms in the North Sea. In some cases, the additional QA,’QC measures are able to substantially reduce the base error rates, reducing them by a factor of 10 when the QA/QC measure is highly effective.
- 339. Design zyxwvutsr jov Reliability: Human and OrgaRisationalFactors zyxwvu 989 z Phase, HOE, scenario ID Base rate QAjQC alt. QA/QC alt. likelihood zyxw A likelihood B likelihood z 1 1Design, omit install fatigue 15.0-8.9E-3 1 4.6-8.2E-3 I 1.4-2.5E-3 1 Fabrication, fit-uplwelding defects Installation, dented braces Production, dropped package 0.7-2.1 E-3 1.0-2.8E-4 1.3-3.8E-4 I ! 2.0-3.7E-3 2.0-3.7E-4 4.0-7.4E-4 1.0-4.OE-3 1.4-5.7E-4 1.9-7.8E-4 ~ 1Production supply boat collision 1 3.0-8.7E-3 1 0.6-1.7E-3 I 0.9-2.6E-3 ~ Figure 12.14 Life cycle malfunctionsscenario likelihoods In some cases, the QAlQC measures are not very effective in reducing the likelihood of the HOE effects. The primary results from the system reliability analyses of the four structures are summarised in fig. 12.15. The intrinsic (error free) probabilities of system failure under extreme conditions and combined fatigue and extreme condition loadings are given in the first row of table 12.7. The probabilities of system failure as a result of extrinsic or HOE causes are added to the “error free” intrinsic probabilities of system failure to obtain the total probability of system failure. Based on the results from the analyses, the first two malfunction scenarios: (1) omission of pile driving stresses during design and not making adequate provisions for alignment of piles during driving and (2) fit-up and welding flaws introduced during fabrication, both of which affect fatigue strength, have the most significant influence in degrading the system reliability of the three-pile Monotower and Vierendeel Tower structures. These two structures are less robust under these HOE scenarios. The four-pile Jacket shows only a marginal influence due to HOE scenario (2) while the Braced Caisson shows practically no influence from these HOE scenarios. Under HOE scenario (4), only the three-pile
- 340. 990 zyxwvutsrqp Figure 12.15 Intrinsic and total annual probabilitiesof system failure zyx Chapter 12 Monotower shows a significant reduction in reliability as a result of ship impact damage, while the other three structures show high levels of robustness. The HOE scenarios (3) and z (5) involving damage to one of the braces do not show a significant impact on the system reliability of any of the four structures. The three-pile Monotower and the Vierendeel Tower structures were shown to be particularly susceptible to potential HOE, which affect the fatigue strength of critical welds. The implication is that effective QA/QC measures should be employed to safeguard these structures against such defects. In addition, designing the critical welds to longer fatigue lives and thorough inspection of welds at the fabrication yard and after installation are implicated to help minimise the risk of these scenarios actually developing. 12.6.2 Deepwater Structure Design Project A review has been performed of a deep-water structure design project that involved the use of very innovative design methods and technology. The assessment team was given full access to the design management organisation, the engineering organisation and the classification-verification organisation. This included reviews of design documentation, specifications and background information, and interviews zyxw - discussions with the members of each of the organisations. In an attempt to reduce initial costs, the design approach involved very advanced and innovative design procedures and technology. Specific “target” reliabilities were defined by the owner/operator for the structure. A Value Improvement Program (VIP) was instituted. The goal of the VIP was to reduce the initial cost of the project by 25%. At the time of this review, the design had been underway for two years. The work had included extensive analyses of alternatives, development of computer programs and performance of experimental work on several of the critical components. A leading classification society was involved in an on-going QA/’QC program that included a failure modes and effects analysis of the structure system.
- 341. Design for Reliabilitj. Human and Organisarional Factors zyxwvu 99 z 1 The assessment team included representatives of the management, engineering, and classification organisations (participatory ergonomics approach). They participated in training workshops that focused on the HOF aspects of the engineering design process and on the HOF considerations in developing successful platform life cycles. The consensus results from the first round of analyses indicated significant concerns for the design procedures, design personnel and management, technology and quality incentives. The concerns for: Design procedures were focused on the very sophisticated and complicated methods that were involved in the analysis of a very complex interaction of the structure, the foundation and the oceanographic environment. Design personnel and management were focused on the low level of experience of the lead design engineers and on their on-going debates with the project management’s requirements for verifications and validations of the results from the design analyses. Technology was focused in the first-time nature of the engineering methods and analytical tools being used in the design (based on limiting strains and deformations). Quality incentives regarded the VIP and the lack of specific guidelines on the effects of VIP on the quality and reliability of the structure. After the first round evaluations were completed, a second team of assessors was organised that included representatives of the design organisation’s management, engineering and verification teams. The results are summarised in fig. 12.16. The resultant uncertainties are indicated for each of the components (best estimate, zyxw &lo standard deviation). The high grades (indicating below average quality attributes) for interfaces, procedures and operators reflected the same primary issues indicated by the qualitative assessment. Examination of the factors and attributes associated with the gradings indicated that the primary reasons for the high grading of the interfaces referred to the lack of appropriate interfacing between the design and management teams. There was a contention between engineering and management. Engineering felt that once an analysis was completed and verified, then the results should be implemented in the design. Management felt that z Figure 12.16 Grades from interactive application of zyxwvu QMAS to design team, process, and organisation
- 342. 992 zyxwvutsrqponm Chnptev 12 z interpretation and judgement needed to be used as screens to assure that the results “made sense” before they were used. The reasons for the high grading of the procedures referred to the lack of first principles and experimental verifications of the computer programs that were being used in the design and the lack of any specific guidelines to determine the effects on the structure reliability of the VIP. The reason for the average grading of the operators (the design team) was the relatively low level of structure design experience in the design team and the lack of in-depth construction and operations experience in the team. The review included five recommendations to improve the QMAS gradings: Develop and implement definitive guidelines to evaluate the quantitative effects of the VIP alternatives and measures on life cycle costs and reliability of the structure (theseguidelineswould be consistent with the background that had been used to develop the reliability targets), Develop and implement a “challenge” process in the design procedures that would assure that all results from engineering analyses were validated by alternative analyses, experimental field data and experienced judgement (the ongoing QA/QC process would be replaced), Assign additional experienced structural design engineers to the design team (less experienced personnel would be assigned to other projects), Temporarily assign construction and operations engineering personnel to the design team to review the construction, operations and maintenance characteristics being used in the design (these personnel were representatives of the organisations that would build and operate the structure) and Develop a structural robustness program and design guidelines that would assure fail- safe design (intrinsic safety) for all of the critical structuraland equipment components through the life cycle of the structure (explicitdesign for damage and defect tolerance). The SYRAS instrument was used to evaluate the reliability and life cycle cost implications of the VIP alternatives [Bea, 20001a, 20011. The structural quality profiling instrument and the QMAS instrument proved to be effective and efficient. The recommendations developed during the assessment process were implemented by the management, engineering and classification-verification organisations. The recommendations proved to be practical and cost-effective. zyxwv 12.7 Summary and Conclusions Those responsible for the development and creation of offshore structures, the associated regulatory agencies, their engineers, managers and the operating staffs have much to be proud of. There is a vast international infrastructure of offshore structures that supply much needed goods and services to the societies they serve. This chapter addresses the issues associated with helping achieve desirable quality and reliability of offshore structures during their life cycles. The primary challenge that is addressed is not associated with the
- 343. Design for Reliability; Human and Organisational Factors zyxwvu 993 traditional engineering technologies that have been employed in the creation of these structures. History has shown that this is not the challenge. Rather, the primary challenge that is addressed is associated with the human and organisational aspects of these systems. A colleague recently stated: “most engineers want to believe that the planet is not inhabited”. It is clear that human and organisational factors are the primary challenges in developing offshore structure systems that have desirable and acceptable quality and reliability. Also, it is clear that there is a significant body of knowledge about how to address this challenge. The problem is wise implementation of this knowledge on a continuing basis. Two instruments have been advanced to enable improved recognition of HOF in the design of offshore structures. Qualitative insights into potential performance characteristics of offshore structures are provided by the QMAS instrument; the primary focus of this instrument is on the HOF that influences the quality and reliability of these structures. Quantitative insights are provided by the SYRAS instrument. zyxw A “calibrated link” has been developed to enable the insights developed with application of QMAS to be translated into “reasonable” quantitative results that include explicit analyses of HOF. The combination of QMAS and SYRAS have been applied in several industry projects that have studied the considerations associated with “minimum” offshore structures, and in a variety of operating settings including design QA/QC, construction and operations. It should be apparent to all engineers that HOF is of fundamental importance in the development of offshore structures that will have acceptable and desirable quality and reliability during their life cycles. Design engineers have a fundamental and primary responsibility in addressing HOF as an integral part of the design engineering process. It should also be apparent to all concerned with the quality and reliability of offshore structures that organisations (industrial and regulatory) have pervasive influences on the assessment and management of threats to the quality and reliability of offshore structures. Management’s drives for greater productivity and efficiency need to be tempered with the need to provide sufficient protections to assure adequate quality and reliability. The threats to adequate quality and reliability in offshore structures emerge slowly in the design office. It is this slow emergence that generally masks the development of the threats to words quality and reliability. Often, the participants do not recognise the emerging problems and hazards. They become risk habituated and loose their wariness. Often, emerging threats are not clearly recognised because the goals of quality and reliability are subjugated to the goals of production and profitability. This is a problem, because there must be profitability to have the necessary resources to achieve quality and reliability. Perhaps, with the present high costs of lack of quality and reliability, these two goals are not in conflict. Quality and reliability can help lead to production and profitability. One must adopt a long-term view to achieve the goals of quality and reliability, and one must wait for production and profitability to follow. However, often we are tempted for today, not tomorrow. The second important thing that we have learned about approaches to help achieve management desirable quality and reliability is organising the “right stuff’ for the
- 344. 994 zyxwvutsrqpo Chapter I2 “right job”. This is much more than job design. It is selecting those able to perform the daily tasks of the job within the daily organisation required to perform that job. Yet, these people must be able to re-organise and re-deploy themselves and their resources as the pace of the job changes from daily to unusual (it improves time). Given most systems, they must be team players. This is no place for “super stars” or “aces”. The demands for highly developed cognitive talents and skills is great for successful crisis management teams. In its elegant simplicity, Crew Resource Management has much to offer in helping identify, train and maintain the right stuff. If properly selected, trained and motivated, even “pick-up ball teams” can be successful design engineering teams. The final part of the 15-yr stream of research and development on which this chapter is based addresses the issues associated with implementation [Bea, 2000al. A case- based reasoning study of a dozen organisations that had tried the implementation for a significant period of time identified five key attributes associated with successful implementation: Cognisance zyxwvuts - of the threats to quality and reliability, Capabilities - to address the HOF and HRO aspects to improve quality and reliability, Commitment - to a continuing process of improvement of the HOF and HRO aspects, Culture - to bring into balance the pressures of productivity and protection and to realise trust and integrity, and Counting - financial and social, positive and negative, ongoing incentives to achieve adequate and desirable quality and reliability. It is interesting to note that of the seven organisations that tried implementation, only two succeeded. It is obvious that this is not an easy challenge, and that at the present time, failure is more the rule than success. It is also interesting to note that the two organisations that succeeded recently have shown signs of “backsliding”. Organisational-management evolution has resulted in a degradation in the awareness of what had been accomplished and why it had been accomplished. The pressures of doing something “new”, downsizing, outsourcing, merging, and other measures to achieve higher short-term profitability have resulted in cutbacks in the means and measures that had been successfully implemented to reduce the costs associated with lack of adequate and acceptable quality and reliability. Perhaps, all organisations are destined to continually struggle for the balance in production and protection, and accidents represent a map of that struggle to succeed and survive. zyxwv References American Bureau of Shipping (1998). The application of ergonomics to marine systems, Guidance Notes, Houston, Texas. Apostolakis, G. E., Mancini, G., van Otterloo, R. W., and Farmer, F. R. (Eds.) (1990). Reliability engineering and system safety, Elsevier, London. Aven, T. and Porn, K. (1998). Expressing and interpreting the results of quantitative risk analysis: review and discussion, Reliability Engineering and System Safety, Vol. 61, Elsevier Science Limited, London, UK, 1998.
- 345. Design zyxwvutsrqpon f o r Reliability: Human and Orgunisational Factors zyxwvu 995 Avigutero, T. and Bea, R. G. (1998). “Effects of damage and repairs on the lateral load capacity of a typical template-type offshore platform“, Proceedings International Offshore and Polar Engineering Conference, International Society of Offshore and Polar Engineer- ing, Golden, CO. Bea, R. G. (1974).Selectionof environmental criteria for offshore platform design,Journal of Petroleum Technology, Society of Petroleum Engineers, Richardson, Texas, pp. 1206-1214. Bea, R. G. (1975). Development of safe environmental criteria for offshore structures, Proceedings Oceanology International Conference, Brighton, UK. Bea, R. G. (1992). Marine structural integrity programs (MSIP), Ship Structure Committee, SSC-365, Washington, DC. Bea, R. G. (1996a). Human and organisation errors in reliability of offshore structures, Journal of Offshore Mechanics and Arctic Engineering, American Society of Mechanical Engineers, New York, Nov.-Dec. 1996. Bea, R. G. (1996b). Quantitative and qualitative risk analyses zyx - the safety of offshore platforms, Proceedings of the Offshore Technology Conference, OTC 8037, Society of Petroleum Engineers, Richardson, Texas. Bea, R. G. (2000a). Achieving step change in risk assessment and management (RAM): Centre for Oil and Gas Engineering, http www.oi1-gas.uwa.edu.au,University of Western Australia, Nedlands, WA. Bea, R. G. (2000b).Performance shaping factors in reliability analysis of design of offshore structures, Journal of Offshore Mechanics and Arctic Engineering, Vol. 122, American Society of Mechanical Engineers, New York: NY. Bea, R. G. and Lawson, R. B. (1997). Stage-11 analysis of human and organisational factors, Report to JIP on Comparative Evaluation of Minitnum Structures and Jackets, Marine Technology and Development Group, University of California at Berkeley. Bea, R. G.. Brandtzaeg, A.. and Craig, M. J. K. (1998). Life-cycle reliability characteristics of minimum structures, Journal of Offshore Mechanics and Arctic Engineering, Vol. 120, American Society of Mechanical Engineers, New York, NY. Bea, R. G., Holdsworth, R. D., and Smith, C. (Eds.) (1996). Proceedings 1996 International Workshop on Human Factors in Offshore Operations, American Bureau of Shipping, Houston, Texas. Bier. V. M. (1999). Challenges to the Acceptance of Probabilistic Risk Analysis, Risk Analysis, Vol. 19, No. 4. Bucknell, J. (2000). Defect assessment for existing marine pipelines and platforms, Proceedings APEC zyxwvut f Asia-Pacific Economic Cooperation) Workshop on Assessing and Maintaining the Integritj. of Existing Offshore Oil and Gas Facilities, Beijing, China, U.S. Minerals Management Service, Herndon, VA. Center for Chemical Process Safety (1989). Guidelines for technical management of chemical process safety, American Institute of Chemical Engineers, New York.
- 346. 996 zyxwvutsrqpo Chapter 12 Center for Chemical Process Safety (1992). Guidelines for investigating chemical process incidents, American Institute of Chemical Engineers, New York. Center for Chemical Process Safety (1993). Guidelines for auditing process safety management systems, American Institute of Chemical Engineers, New York. Center for Chemical Process Safety (1994). Guidelines for preventing human error in process safety, American Institute of Chemical Engineers, New York. De Leon, D. and Heredia-Zavoni, E. (2001). Probability distributions of mechanical damage for offshore marine platforms, zyxwv Proceedings of the Eleventh /2001) Iiiternational Offshore and Polar Engineering Conference, International Society of Offshore and Polar Engineers. Demming, W. E. (1982). Out of the crisis, Massachusetts Institute of Technology Center for Advanced Engineering Study, Cambridge, MA. Dougherty, E. M., Jr. and Fragola, J. R. (1986). Human reliability analysis, John Wiley and Sons, New York. Faber, M. H. (1997). Risk Based Structural Maintenance Planning, Probabilistic Methods for Structural Design, Soares, C. G. (Ed.), Kluwer Academic Publishers, Amsterdam, The Netherlands. Fischhoff, B. (1975). Hindsight does not equal foresight: the effect of outcome knowledge on judgment under uncertainty, Journal zyxwv o f Experimental Psychology, Human Perception, and Performance, Vol. 1, New York. Gertman, D. I. and Blackman, H. S. (1994). Human reliability and safety analysis data handbook, John Wiley andsons, New York. Gierlinski, J. T. and Rozmarynowski, B. (1999). Task 1.2 and 11.13: system reliability of intact and damaged structures under extreme environment and fatigue conditions, Report no. WSAiAM3681iTask 1.2, Report to JIP on Comparative Evaluation of Minimum Structures and Jackets, WS Atkins Consultants Ltd, Leatherhead, UK. Groenewg, J. (1994). Controlling the controllable, the management of safety, DSWO Press, Leiden University, The Netherlands. Haber, S. B., O’Brien, J. N., Metlay, D. S., and Crouch, D. A. (1991). Influence of Organisational Factors on Performance Reliability zyxw - Overview and Detailed Methodo- logical Development, U. S. Nuclear Regulatory Commission, NUREG/CR-5538, Washington, DC. Hale, A,, Wilpert, B., and Freitag, M. (1997). After the event, from accident to organisational learning, Pergamon Press, Elsevier Sciences Ltd., Oxford, UK. Harris, D. and Chaney, F. (1969). Human factors in quality assurance, John Wiley and Sons, New York. Hessami, zyxwvuts A. G. (1999). Risk management: a systems paradigm, Systems Engineering, John Wiley and Sons, London, UK, 1999.
- 347. Design for Reiiabiliry; Human and Ovganisational Factors zyxwvu 997 z International Standards Organisation (1994a). IS0 9000 Series, quality management and quality assurance standards, British Standards Inst. Publication, London, UK. International Standards Organisation (1994b). Quality systems zyx - model for quality assurance in design/development, production, installation, and servicing, IS0 9001, London, UK. International Standards Organisation (1994c). Health, safety, and environmental manage- ment systems, Technical Committee ISOITC 67, Materials, Equipment and Offshore Structures for Petroleum and Natural Gas Industries, Sub-committee SC 6, Processing Equipment and Systems, London, UK. Jones, R. B. (1995). Risk-based management - a reliability centered approach, Gulf Publishing Co., Houston, Texas. Kirwan, B. (1994). A guide to practical human reliability assessment, Taylor and Francis, London, UK. Klein, G. (1999). Sources of power, MIT Press, Cambridge, Massachusetts, 1999. Kletz, T. (1991). An engineer's view of human error, Institution of Chemical Engineers, Rugby, UK. Knoll, F. (1986). Checking techniques, Modeling Human Error in Structural Design and Construction, Nowak. A. S. (Ed.) American Society of Civil Engineers, Herndon, Virginia. Kontogiannis, T. and Lucas, D. (1990). Operator performance under high stress: an evaluation of cognitive modes, case studies and counter measures, Report No. R90103, Nuclear Power Engineering Test Center, Tokyo, Japan, Human Reliability Associates, Dalton, Wigan. Lancashire, UK. Libuser, C. (1994). Managing organisations to achieve risk mitigation, PhD Dissertation, Andersen School of Business, University of California, Los Angeles. Marine Technology Directorate (1989). Underwater inspection of steel offshore installa- tions: implementation of a new approach, Report 891104, London, UK, 1989. Marine Technology Directorate (1992). Probability-based fatigue inspection planning, Report 92: 100, London. Marine Technology Directorate (1994). Review of repairs to offshore structures and pipelines, Report 94/102, London, UK. Marshall, P. W. and Bea, R. G. (1976). Failure modes of offshore platforms, Proceedings of Behaviour of Offshore Structures Conference, BOSS '76, Vol. 11, Trondheim, Norway. Matousek, M. (1990). Quality assurance, Engineering Safety, Blockley, D. (Ed.) McGraw-Hill Book Co., London, UK. Melchers, R. E. (1993). Society, tolerable risk and the alarp principle, Proceedings of the Conference on Probabilistic Risk and Hazard Assessment, Melchers, R. E. and Stewart, M. G. (Eds.), The University of Newcastle, N.S.W., Australia.
- 348. 998 Chapter zy 12 z Moan, T. (1997). Current trends in the safety of offshore structures, Proceedings International Offshore and Polar Engineering Conference, International Society of Offshore and Polar Engineers, Golden, Colorado. Molak, V. (Ed.) (1997). Fundamentals of risk analysis and risk management, CRC Lewis Publishers, New York, 1997. MSL Engineering Ltd. (1999). Effect of vessel impact on intact and damaged structures, Doc. Ref. C209R007-Rev 1, Report to JIP on Comparative Evaluation of Minimum Structures and Jackets, London, UK. RAMB0LL (1999). Conceptual design summary report, Job No. 978503, Doc. No. 340- 005, Rev. 1, 1999-01-05,Report to JIP on Comparative Evaluation of Minimum Structures and Jackets, Copenhagen, Denmark. Rasmussen, J. (1996). Risk management,adaptation, and design for safety, Future Risks and Risk Management, Sahlin, N. E. and Brehemer, B. (Eds.), Kluwer Publishers, Dordrecht. Rasmussen, J., Duncan, K., and Leplat, J. (Eds.) (1987). New Technology and human error, John Wiley and Sons, New York. Reason, J. (1990). Human error, Cambridge University Press, London, UK. Reason, J. (1997). Managing the risks of organisational accidents, Ashgate Publishers, Aldershot, UK. Roberts, K. H. (1989). New Challenges in Organisational Research: High Reliability Organisations, Industrial Crisis Quarterly, Vol. 3, Elsevier Science Publishers, Amsterdam, the Netherlands. Roberts, K. H. (Ed.) (1993). New challenges to understanding organisations, McMillan Publishing Co., New York. Rochlin, G. 1. (1997). Trapped in the net: the unanticipated consequences of computerisa- tion, Princeton University Press, Princeton, New Jersey. Shetty, N. (Ed.) (2000). Comparative evaluation of minimum structures and jackets, Synthesis Report, Report to Joint Industry Project Sponsors, W. S., Atkins Consultants Ltd., Report No. AM3681, Leatherhead, UK. Soares, C. G. (Ed.) (1998). Risk and reliability in marine technology, A. A. Balkema, Rotterdam, The Netherlands. Spouge, J. (1999). A guide to quantitative risk assessment for offshore installations, CMPT Publication 991100, ISBN I870553 365, London, UK. Stewart, M. G. and Melchers, R. E. (1988). Checking models in structural design, Journal of Structural Engineering, Vol. 115,No. 17,American Society of Civil Engineers, Herndon, Virginia. Swain, A. D. and Guttman, H. E. (1983). Handbook of Human Reliability Analysis with Emphasis on Nuclear Power Plant Applications, NUREG/CR-1278, U.S. Nuclear Regulatory Commission, Washington, DC.
- 349. Design zyxwvutsrqponm for Reliabilitj,: Human and Organisational Factors zyxwvu 999 Vinnem, J. E. (1998). Evaluation of methodology for QRA in offshore operations, Reliability Engineering and System Safety,Vol. 61, Elsevier ScienceLimited, London, UK. Weick, K. E. (1995). Sensemaking in organisations, Sage Publishers, Thousand Oaks, CA. Weick, K. E. (1999). Organizing for high reliability: processes of collective mindfulness, Research in Organisational Behaviour, Vol. 21, JAI Press Inc. Weick, K. E. (2000). The neglected context of risk assessment zyx - a mindset for method choice, Risk Management in the Marine Transportation System, Transportation Research Board, National Research Council, Washington, DC. Weick, K. E. and Quinn, R. E. (1999). Organisational change and development, Annual Review of Psychology, New York. Weick, K. E., Sutcliffe, K. M., and Obstfeld, D. (1999). Organizing for high reliability: processes of collective mindfulness, Research in Organisational Behaviour, Staw and Sutton (Eds.), Research in Organisational Behaviour, JAI Press, Vol. 21, Greenwich, CT. Weick, K. E. and Sutcliffe, K. M. (2001). Managing the unexpected, Jossey-Bass, San Francisco, CA. Wenk, E. Jr, (1986). Tradeoffs, imperatives of choice in a high-tech world, The Johns Hopkins University Press: Baltimore. MD. Winkworth, W. J. and Fisher, P. J. (1992). Inspection and repair of fixed platforms in the North Sea, Proceedings Offshore Technology Conference, OTC 6937, Society of Petroleum Engineers, Richardson, Texas. Woods, D. D. (1990). Risk and human performance: measuring the potential for disaster, Reliability Engineering and System Safety, Vol. 29, Elsevier Science Publishers Ltd., UK. Wum, J. zyxwvuts S., Apostolakis, G. E., and Okrent, D. (1989). On the inclusion of organisational and management influences in probabilistic safety assessments of nuclear power plants, Proceedings of the Society for Risk Analysis, New York, 1989. Xu, T., Bea, R. G., Ramos, R.,Valle, O., and Valdes, V. (1999). Uncertainties in the fatigue lives of tubularjoints, Proceedings Offshove Technology Conference, OTC 10849,Society of Petroleum Engineers, Richardson, Texas.
- 350. Handbook of Offshore Engineering zyxwvutsr S . Chakrabarti (Ed.) zyxwvutsrq C 2005 Elsevier Ltd. zyxwvutsrq All rights reserved zyxwvuts 1001 Chapter 13 Physical Modelling of Offshore Structures Subrata K. Chakrabarti Offshore Structure Analysis, Inc., Plainfield, IL, USA 13.1 Introduction This chapter will describe the need, the modelling background and the method of physical testing of offshore structures in a small-scale model. The physical modelling involves design and construction of scale model, generation of environment in an appropriate facility, measuring responses of the model subjected to the scaled environment and scaling up of the measured responses to the design values. The purpose of duplicating the environment experienced by an offshore structure in a small scale is to be able to reproduce the responses that the structure will experience when placed in operation in the offshore site. This enables the designers to verify their design methods and take any necessary corrective actions for the final design of the structure before it is released for construction. The physical model also allows the proof of the concept for a new and innovative design for a particular application as well as verifies the operational aspects of a designed structure. For a successful physical modelling, the following areas should be known and will form the basis for this chapter: zyxwvu 0 0 0 0 0 0 0 0 0 0 0 Needs for model tests Similarity laws for modelling Froude number and related scaling Reynolds number and its effect Towing resistance and drag effect Scaling of a hydroelastic model Offshore model testing facilities and their qualifications and limitations Important components of a wave basin Modelling of environment Instrumentation requirements and measurement accuracy Modelling difficulties and distortion in scaling
- 351. 1002 zyxwvutsrqpo Chupter z 13 Data analysis and reporting While a general discussion on the physical modelling problem will be made, emphasis has been placed on modelling the present-day offshore structures in a small scale. Particular attention has been given to the testing of deep-water offshore structures. The chapter is laid out in such a way that it may be used in developing a request for testing proposal for a forthcoming model test. Challenges in testing of deepwater and ultra-deepwater structures zyx 13.1.1 History of Model Testing Model testing has been an integral part of the development of offshore structures starting with the shallow water structures in the early fifties to the present day. The operational elements for an offshore structure are routinely examined through model testing. Many of the design parameters are verified through model tests. As the water depth for offshore structures is getting deeper, the technique for small-scale testing is becoming increasingly more challenging. Model testing for today’s deep-water structures, however, is essential for a better understanding of the stability behaviour and survival characteristics of deep- water structures. A description of the role of model testing has been given in Dyer and Ahilan (2000). Many detailed aspects of physical modelling and testing may be found in Chakrabarti (1994). Experimental testing of physical scale models in a wave basin, in which the critical response parameters are determined by direct measurement, has been the traditional way of investi- gating the behaviour of offshore vessels [ITTC, 19991. It is recognized as the most reliable tool for reproducing realistic and extreme situations an offshore structure is expected to experience in its lifetime. In particular, it may be important for complex systems, where various kinds of static and dynamic coupling effects among the system components may occur. Additionally, the physical models have the advantageover the numerical models that unknown phenomena and effects, not described by theoretical models, can be discovered. Typically, model scales in the range 1 zyxwv :50-1 :70 are used for such testing. The complete floater system with moorings and risers are modelled. Often a simplified modelling is employed including a truncated mooring system and a reduced number of “equivalent” risers. These will be discussed further in the subsequent sections. Because of the limitation in the available basin depths, such truncation is a common occurrence. As the water depth goes deeper, question may be raised as to the suitability of such testing of distorted models. It is probably safe to state that deep-water development will proceed with or without model testing. The designers will use whatever design tools are considered appropriate for their design without direct verification through model testing. However, in practice, such will seldom be the case. Decisions for placing deep-water struc- tures will be made at a level different from designers and experimenters. All such developments will include plans and budget for model testing. In fact, no new structurewill probably be built and installed without some scale model testing. Therefore, intelligent decisions should be made in planning such testing when the full model scaling is not possible. This, however, is not new. Offshore structure development has taken a similar course in the past. We should not overlook the fact that many of the theories were developed only after a phenomenon was
- 352. Physical Modelling zyxwvutsr o f Ofshore zyxwvutsr Structures zyxwvutsr 1003 z discovered during a model (and prototype) observation. One example is the high tendon loads experienced by the tension leg platform (TLP). Model tests revealed very high vertical loads in the TLP tendon even though heave natural period was very low. Theory has evolved since its discovery, which mathematically describes the source of such loads commonly known as “springing”. Since then, impact-type loading has also been discovered in full-scale TLP measurements and theory for this “ringing” load has been developed subsequently. Other areas of measurements of model response include slow drift oscillation of a soft moored floating structure, green water impact on decks, air gaps, stability of floating structures, etc. Many of these tests used truncated models. Some of these physical phenomena still do not have adequate analysis tools. Structures have been developed and installed successfully in spite of these deficiencies. Let us illustrate by a simple example. In 1965, Chicago Bridge & Iron Co. (CB&I) built its first fixed offshore structure in the Persian Gulf, named the Khazzan Dubai oil storage tank. This structure was an ideal candidate for the 3-D diffraction design tool for a general shape, which did not exist at the time. In fact, in 1970, the traditional 3-D linear diffraction theory programme was first developed and commercially used in offshore structure development. This programme was not used in the design of Khazzan tank, which used simply the Froude-Krylov theory along with a large safety factor. The first model test of the Khazzan tank was very crude with substandard measurements compared to today’s technology. Later, verification of model test with the diffraction theory showed that the design was conservative. Today after 30 years, Conoco is still successfully operating the three CB&I-built Khazzan storage tank complex in the Persian Gulf. A scale model test in a wave basin is carefully controlled, and has far superior accuracy and sophistication today. For ultra-deepwater structures, scale distortion and truncation are here to stay no matter where the testing is performed. On the other hand, truncation is nothing new in model testing. Coastal people have been running successful tests with distorted models for many many years. Most of the recent deep-water model tests for the new generation semi-submersibles and SPARS needed truncation in the area of, among others, the mooring lines and risers in the system. They have been considered successful and meaningful results were obtained for motions, sectional structure loads and mooring line loads, air gaps, slamming loads, green water effects, etc. Moreover, there has hardly been a model test where the experimenters not learnt something new, no matter how trivial the tests were. Admittedly, the ultra-deepwater testing poses additional challenges. This means that such testing should be planned more carefully. In fact, as much time should be spent in the planning of these tests as the actual testing time in the basin. The test goal should remain focused in these cases, rather than all-inclusive. It is possible that different types of testing should be designed for different goals for the same system. Some testing of the structure component may be included in the overall plan along with the complete (distorted) testing. Sometimes, the multiple model scales of the system may be warranted. New phenomena will continue to be discovered for these ultra-deepwater structures yet undefined by theory. In particular, the stability as well as non-linearity issue is an open question. Additionally, model testing is a simple and efficient technique in improving and optimising a system or a concept, which is more difficult to achieve alternatively.
- 353. 1004 zyxwvutsrqp Chapter z 13 Therefore, physical model test of the overall system that takes care of distortion in a systematic way should continue. What modelling technique will work, of course, will depend on the particular system in question. In fact, certain specific ground rules may be laid out for deep water testing on the basis of what is available to us today. This may be an analysis of dos and donts and pitfalls to avoid. Some of these areas will be covered in the latter sections. zyxwvut 13.1.2 Purpose of Physical Modelling One of the principal benefits of model testing is that valuable information is provided which can be used to predict the potential success of the prototype at relatively small investment. The physical model provides qualitative insight into a physical phenomenon, which may not be fully understood currently. The use of models is particularly advanta- geous when the analysis of the prototype structure is very complicated. In other situations, models are often used to verify simplified assumptions, which are inherent in most analy- tical solutions, including nonlinear effects. An example of this is the discovery of slow drift oscillation of a moored floating tanker through model testing before a theory describing the second-order oscillating drift force and the associated motion was derived. Model test results are also employed in deriving empirical coefficients that may be directly used in a design of the prototype. Therefore, the following list gives the principal benefits to be gained from a model test: z 0 Validate design values Obtain empirical coefficients Substantiate analytical technique Problem difficult to handle analytically Verify offshore operation, such as, a specific installation procedure Evaluate higher order effects normally ignored in the analysis Investigate unpredicted or unexpected phenomena 13.2 Modelling and Similarity Laws It is important to have a clear understanding of the scaling laws before the model measurements may become meaningful. Why do we need scaling laws? We can cite the following reasons for scaling laws: Testing is generally done in a small scale, The scaling laws allow scaling up of the measured data to full scale. Modelling laws relate the behaviour of a prototype to that of a scaled model in a prescribed manner. The problem in scaling is to derive an appropriate scaling law that accurately describes this similarity. In modelling a prototype structure in a small scale, there are, at least, three areas where attention must be given so that the model truly represents the prototype behaviour zyxwvu - structure geometry, fluid flow and the interaction of the two. Therefore, we shall seek similarity in the structure geometry, similitude in the fluid kinematics and the similitude in the dynamics of the structure subjected to the fluid flow around it.
- 354. Physical zyxwvutsrqp Modelling zyxwvutsr o f Ofshore Structirres zyxwvuts 1005 z 13.2.1 Geometric Similitude zyxwvu Geometrically similar structures have different dimensions, but have the same shape. In other words, a model built for testing in a small scale must resemble the prototype in shape, specially the submerged sections. At least, the important submerged elements must be modelled accurately. This can be easily achieved if we assume that a constant scale ratio exists between their linear dimensions - [ P - - a (13.1) ern where zyxwvutsr tpand e,, are any two corresponding homologous dimensions of the two structures namely, prototype and model, respectively and a is the scale ratio between them. In this case, we say that the two structures are geometrically similar. The ratio of the two similar dimensions (e.g. diameter and length of a particular member) will, therefore, remain constant and establishes the scale factor for a model. This factor will be defined as E, throughout this book. 13.2.2 Kinematic Similitude The kinematic similitude is achieved in the model if the ratio of the fluid velocity and fluid acceleration are preserved. Thus, the ratio of the prototype velocity to the corresponding model velocity will be a prescribed constant. This applies to all velocities including fluid particle, wind speed, towing speed, model velocity in a particular direction, etc. Similarly, the ratio of the acceleration will be a different constant. Their relationships will be deter- mined from the scaling laws. When these laws are satisfied for velocity and acceleration, the model is considered kinematically similar to the full-scale structure. 13.2.3 Hydrodynamic Similitude Consider the masses of two similar structures in similar motions. Noting that the induced force may be written using the Newton’s law as the product of mass and acceleration, all corresponding impressed forces must be in a constant ratio and similar direction. Therefore, geometrically similar structures in similar motions having similar mass systems are similarly forced. When the model is forced similar to the prototype, the model is considered dynamically similar to the prototype. The chosen scaling laws establish this scaling relationship for the model. In order for a model to truly represent the full-scale structure, all three conditions, namely the geometric, kinematic and dynamic similarities must be maintained. Then only the model test data may be scaled up to the full scale without any distortion. Hydrodynamic scaling laws are determined from the ratio of forces. Table 13.1 gives the most common scaling laws from the fluid structure interaction problem. Several ratios may be involved in a particular scaling. One of these may be more predominant than others. The dynamic similitude between the model and the prototype is achieved from the satisfaction of these scaling laws. In most cases. only one of these scaling laws is satisfied by the model structure. Therefore, it is important to understand the physical process experienced by the structure and to choose the most important scaling law which governs this process.
- 355. 1006 zyxwvutsrqpo Table 13.1 Common dimensionlessquantities in offshore engineering zyx Chapter z 13 1zyxwvutsrqponmlkjihgfedcbaZYXWVUTSR Symbol IDimensionless number IForce ratio IDefinition I 1Fr 1Froude Number /Inertia/Gravity ]zyxwvutsrq u2/gD 1 1Re ~ Reynolds Number IInertia/Viscous iEu 1Euler Number , IInertia/Pressure I P I P 2 I 1Ch Cauchy Number I I UTID ~ 1KC Keulegan-Carpenter Number Drag/Inertia ISt ~ Strouhal Number In table 13.1,D =member diameter, T = wave period, g = gravity, u = kinematic viscosity, p = pressure, E = modulus of elasticity and zyxw fe = vortex (eddy) shedding frequency. The Froude number applies to gravity waves. The Reynolds number is related to the drag force in the structure. The Euler number is not as important as these quantities, except for vertically loaded long, slender structures. The Cauchy number plays an important role for an elastic structure, such as, compliant towers, risers and tendons. The Keulegan- Carpenter number is very important for small structural members where forces are com- puted based on hydrodynamic inertia and drag coefficients (see Chapter 4). The Strouhal number is the non-dimensional vortex shedding frequency and should be considered for a moving structure when the flow past a structuralmember separates and produces vortices past the structure. The typical current or wave-structure interaction problem involves Froude number, Reynolds number and Keulegan-Carpenter number. For structures that are subject to deformation, Cauchy number should additionally be considered. For structures vibrating in fluid medium, the Strouhal number is also included. The frequency of vortex shedding, fe, from a stationary circular cylinder of diameter D in a fluid stream of velocity u has been shown to be a linear function of Reynolds number Re over a wide range. A relationship between the Strouhal number St and Reynolds number Re exists in steady flows. It is generally accepted that St =0.2 in the range 2.5 x lo2 < Re < 2.5 x 10’. Beyond this range, St increases up to about 0.3 and then, with further increase in Re, the regular periodic behaviour of u in the wake behind the cylinder disappears. Some variation in this trend has been observed in experiments by several investigators, particularly, outside the constant range of St. In the offshore structure problem, the most common among the dimensionless scaling laws presented in table 13.1is Froude’s law. The Reynolds number is also equally important in
- 356. Pliysical Modeliing zyxwvutsr of Ojjshove Structures zyxwvuts 1007 z many cases. However, Reynolds similarity is quite difficult, if not impossible, to achieve in a small-scale model. Simultaneous satisfaction of zyxw Fr and Re is even more difficult. The Froude law is the accepted method of modelling in hydrodynamics. zyx 13.2.4 Froude Model The Froude number has a dimension corresponding to the ratio of u2/(gD)as shown in table 13.1. Defining Fr as the Froude model must satisfy the relationship: (13.2) (13.3) Assuming a scale factor of h and geometric similarity, the relationship between the model and full-scale structure for various parameters may be established. Table 13.2 shows the scale factor of the common variables that the Froude model satisfies. The variables chosen are the most common ones that are encountered in the offshore structure testing. For a scale factor of 1 : 50 and 1 : 100 for a model, the multiplying factors for these variables are also shown in the table. For scale factors other than these, this table may be easily converted to yield the multiplying factor for the desired responses. Thus, for a Froude model, the scaling of the model response to the prototype values is straightforward. There are instances, however, where this scaling may not be achieved simply. A few examples will be cited later where this table is not directly applicable and possible remedies or corrections that may be adopted for the above method will be discussed. One should note that fluid density and viscosity are different between the model and prototype, even though the difference is generally small. This difference is often ignored due to small corrections. However, the scaled up values may be corrected by the ratio of these quantities if desired. Chapter 3 lists the values of these quantities for the temperature difference. Examples of a few prototype quantities and environmental parameters are given in table 13.3. A few structure responses are also included in the table. Their values at different scale factors are listed. This is an exercise to illustrate what scale factor for a particular test requirement may be appropriately chosen and what becomes of the magnitudes of quantities in the model scale. Such a table will guide the user to choose the most appropriate scale given the limitation of a chosen testing basin and measurements. 13.2.5 Reynolds Model If a Reynolds model is built, it will require that the Reynolds number between the prototype and the model be the same. Assuming that the same fluid is used in the model system (viscosity ratio = l), this means that: uPDP= u,D, (13.4)
- 357. 1008 Variable All linear dimensions zyxwvu Table 13.2 Scaling of variables using Froude law Symbol Scale factor h=50 h=100 z D zyxwv h 50 100 z Chapter 13 Structure mass Structure moment of inertia nz h3 1.25E3 1.OE6 I h5 3.125E8 1.OE10 z 1Fluid or structure velocity l u 1 h12 1 7.07 I 10 1 Force Moment Stress 1Fluid or structure acceleration 1zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIH U 1 1 1 1 1 1 1 F h3 1.25E3 1.OE6 A 4 h4 6.25E4 1.OE8z U h 50 100 1Time or period I t 1 h12 1 7.07 1 10 1 Spring constant Wave period K h2 2500 1.OE4 T h'l2 7.07 10 1Section moment of inertia I I Izyxwvutsrqpo h4 1 6.25E4 1 1.OE8 1 Gravity Fluid density 1Structure displacement volume 1 V I h3 1 1.25E3 1 1.OE6 I g 1 1 1 P 1 1 1 1Structure restoring moment I C I h4 1 6.25E4 1 1.OE8 1 Fluid kinematic viscosity Reynolds number I , 1 1 1 Re h3 353.6 1000 IWave length I L l h I 50 I 100 I IPressure l v l h I 50 I 100 I IKeulegan-Carpenter number 1 KC I 1 1 1 I l l if a scale factor of h is used in the model, then this equality is satisfied if u , = All, (13.5) In other words, the model fluid velocity must be h times the prototype fluid velocity. In general, this is difficult to achieve, especially if a small-scaleexperiment is planned. It also points out the difficulty of satisfying both the Reynolds and the Froude number simultaneously. On the other hand, if the Froude's law is used in modelling, the distortion in the Reynolds number is large. As noted earlier, for a Froude model, the Reynolds number scales as: Re, = h3I2Re,, (13.6)
- 358. Physical Modelling of Offshore Structures zyxwvuts Length m 500 20 IO 5 zy 1 2.5 I 1 I 1009 IDraft ~m zyxwvu Table 13.3 Scaling of typical prototype parameters for various scale factors 100 4 2 I 1 1 0.5 I 1Parameter 1 Unit iPrototype i I : 25 1 1:50 1 I : loo i 1: 200 j 1 Min wave height Max wave period m 1 2 1 0.08 ~ 0.04 0.02 1 0.01 I sec 1 20 I 4 1 2.8 2 1 1.4 IColumn diameter 1 m 1 50 1 2 1 1 ' 0.5 I 0.25 1 1Structure mass i kg 1 1E6 1 64 I 8 ~ 1 1 0.125 1 IMax wave height 1 m 1 30 1 1.2 1 0.6 1 0.3 1 0.15 1 /Min wave period 1 sec 1 5 ~ 1 1 0.7 0.5 1 0.35 1 1Current speed ' m/s I 1 1 0.2 1 0.14 I 0.1 1 0.07 1 1Load ~ N 1 1E6 1 64 1 8 ~ 1 ~ 0.125 ~ 1Displacement l m I 2 1 0.08 ~ 0.04 1 0.02 I 0.01 Therefore, the larger the scale factor, the larger is the distortion in the Reynolds scaling. In fact, it is possible that the model flow will be laminar. while the prototype flow falls in the turbulent region. Experiments have shown that the flow characteristics in the boundary layer are most likely to be laminar at Re < lo5, whereas the boundary layer is turbulent for Re > lo6. In this case, two different scaling laws apply, (namely, both Froude and Reynolds), which cannot be satisfied simultaneously. (Use of different fluids to match the Reynolds number may not be practical.) In this case, it is most convenient to employ Froude scaling and to account for the Reynolds disparity by other means. There are several methods that may be used to account for the distortion in the Reynolds scaling for a Froude model: Maximise scale of the model to simulate the prototype effect closer Correct Reynolds effect in scaling up of data to full scale Trip the incoming flow by roughing the model surface in the forward area Induce turbulence in the flow by external means ahead of the model The larger the model, the closer is the flow simulation. This is, however, difficult to achieve for offshore structures. Sometimes, fluid of lower viscosity than water is used to increase the value of Reynolds number in the model. For equality of both Froude and Reynolds number, a fluid whose kinematic viscosity is about l/h3'of that of water should be used. When h is large, such as for offshore structures, this is impossible to achieve. One method of achieving a proper Reynolds number effect at the boundary layer is to deliberately trip the laminar flow in the model by introducing roughness on the surface of the forward part of the model. Then the model in most part will see turbulent flow. This works for a long model, because once the flow regime is turbulent, the drag effect is only weakly dependent on the Reynolds number. In testing tanker models, external means, such as studs, pins or sand-strips attached near the bow, are often used to induce turbulence.
- 359. 1010 zyxwvutsrqpon Chapter I3 z Flow can also be tripped ahead of the model by introducing a mesh barrier submerged from the surface. In towing tests, with horizontally long structures, such as, ships or barges, the skin friction resistance is comparable to the wave-making resistance and is dependent on Reynolds number. Thus, towing resistance depends on both Froude and Reynolds law. Corrections are made in the friction factor (which is known as a function of Reynolds number) based on the respective Reynolds number before the data on model towing resistance is scaled up to the prototype value. If this difference is ignored in scaling, the (scaled up) prototype data will generally be non-conservative. zyxw 13.2.5.1 Towing Resistance of a Ship Model in Wave Basin For a ship/barge model, the scaling is done with Froude scale and corrective mea- sures are taken to scale up the measured values in model scale. The following steps are adopted routinely as a standard procedure: Measure model resistance R, at model speed zyxw u, by towing the model in water with carriage Compute total resistance coefficient C,, by dividing R, by the factor zyx (O.SpA,u;,) where A,,, is the model submerged surface area Compute model friction coefficient C , , by the Schoenherr formula based on model Re, number 0.075 (1ogloRe- 212 zyxwvu Cf= Compute residual coefficient C,, = C,, - Cfm which corresponds resistance and is Froude-scaled Add to C,, a ship appendage correlation allowance Ca of 0.0004 recommendation) (13.7) to wave making (based on ITTC Compute prototype ship frictional resistance coefficient Cf, using equation (13.7) for the prototype Reynolds number Re, Add friction coefficient to the residual Cr, to obtain total prototype ship resistance coefficient C , Multiply by the normalisation factor O.SpA,u; (where A, is the prototype submerged surface area) to obtain the full-scale ship resistance Make correction in the density between the sea water and the fresh water used in the model test by multiplying by the ratio of the two Compute the prototype horsepower requirement. The above procedure is illustrated by an example calculation of total resistance of a barge from the model test results in table 13.4.The model represents a 1:55 scalemodel of a barge. The model was towed at the scaled speed (column 2) and towing loads (column 3) were measured at these speeds. The subsequent columns follow the steps outlined above until the total resistance of barge in full scale (in kips) is obtained on the last column. Since no appendages were present, no allowance for the appendages was considered in this example. This table illustrates how the correction for the Reynolds number distortion is accounted
- 360. Physical IModelling zyxwvutsrq ofzyxwvutsr Offshore Strucrures zyxwvuts Model wetted area zyxwvut = 29.61 ft2 Model length = 10.39 ft 1011 Prototype water mass density = 1.98 Model water mass density = 1.94 ~ Model kinematic viscosity of water = 1.17E-05 ft2Is 1zyxwvu Table 13.4 Scaling of measured model resistance to prototype resistance Proto- type knot Model Resis- 1 Ctm Izyxwvutsrqp Re Cfm ~ Crm I Re zyx C f tance =Crp I I I ft’s (lb) 1 xlO-’ 1 x105 xlOP4 1 x ~ O - ~ 1 x108 xlOP4 xlOP3 I (kips) ~ Proto I 4.83 1 1.1 0.218 0.627 9.74 ~ 3.49 5.92 3.97 ~ 5.18 1 6.44 ~ 38.0 ~~~~~ ,6.15 ~ 1.4 13.95 ~ 0.9 i 0.147 1 0.632 i 7.97 1 3.65 1 5.95 i 3.25 i 5.22 1 6.48 i 25.6 1 0.289 0.513 12.4 1 3.32 4.80 7.03 7.91 1.6 0.419 0.570 14.2 3.22 ~ 5.38 5.78 5.12 1 5.89 73.5 1.8 0.5 ~ 0.537 15.9 3.14 ~ 5.06 6.50 5.10 ~ 5.57 88.0 1 18.79 110.11 I l 10.99 i 2.5 1 0.92 I 0.513 I 22.1 1 2.94 i 4.83 1 9.03 1 5.04 1 5.34 1 162.6 ’ 2 0.6 ~ 0.522 17.7 3.07 I 4.92 7.22 5.08 5.42 105.8 1 2.3 0.791 1 0.521 20.4 2.99 1 4.91 8.31 5.06 5.41 139.7 1 111.86 1 2.7 1 1.06 1 0.506 1 23.9 1 2.89 I 4.77 1 9.75 1 5.03 1 5.28 1 187.6 I 13.18 1 3.0 14.06 I 3.2 14.94 ~ 3.4 15.82 1 3.6 I 1.334 0.516 26.6 2.83 4.88 1 10.8 5.02 5.38 236.1 1 1.492 0.507 28.3 2.79 4.79 11.6 5.01 5.29 264.4 I 1.68 0.506 30.1 2.76 4.78 12.3 5.00 5.28 297.9 I 1.832 0.492 31.9 2.73 4.65 13.0 4.99 5.15 325.4 17.14 118.02 for in the resistance calculations. The scaled-up prototype resistance is shown in fig. 13.1. For comparison, the model values are directly scaled up by Froude scale (see table 13.2 for the factor) without regard to Reynolds number distortion and are also plotted in the figure. In this case, the friction force was small compared to the inertia force so that the difference in magnitude between the two is quite small. 3.9 2.312 I 0.529 34.5 2.68 1 5.02 14.1 4.98 5.52 409.6 4.1 2.872 ~ 0.595 36.3 2.66 ~ 5.68 14.8 4.97 ~ 6.18 506.7 13.2.5.2 Drag Resistance of an Offshore Structure Unlike the ship-shape, many offshore structures are not elongated and have high forward speed and skin friction force is not a concern. However. many members of an offshore structure are subject to drag forces, which experience a similar problem of Reynolds distortion from the Froude scaling. This drag force is called form drag and has been
- 361. 1012 zyxwvutsrqpo Chapter zy 13 600 zyxwvut 500 n zyxwv 5 400 300 .- 200 a, zyxwvutsrqp t p? 100 zyxwvut 111 0 0 5 10 15 20 Speed, knots zyxw Figure 13.1 Scaled prototype towing load zyxw described in Chapter 4. The towing resistance for these offshore structures is expected to include both inertia and drag forces. For an offshore structure, where form drag is important, a similar procedure as in Section 13.2.5.1may be adopted: Measure towing resistance of an offshore structure in a towing tank Obtain C, for drag members from available chart on published model test data Compute drag force on drag members at model speed Correct C, due to any shielding effect from members using literature data Subtract model drag force from the measured total load Scale up the difference (residual force) by Froude scale Compute prototype drag force on drag members using design guide Account for any shielding from design guide (e.g. API guidelines) Add prototype drag force to scaled up data Correct for surface roughness, fluid density, etc., as needed. It is possible to derive the drag coefficients for the model and prototype members from Hoerner, (1965) and Sarpkaya (1976) as well as the certifying agency guidelines [e.g. API, 19791for offshore structures. The above correction is needed because of the difference in the flow regime between the model and the prototype. The laminar flow in the model may be equivalently compensated by artificial stimulation. In this case, no corrections are necessary in the process of scaling the model data. This is illustrated in figs. 13.2 and 13.3 for a semi-submersible production model. In fig.13.2 a semi-submersible rig model is seen being towed in a wave basin with the help of an overhead carriage. The load cells to measure the towing resistance may be seen between the model and the carriage. A submerged grid screen may be seen in the foreground mounted on the same carriage about 3 m (10 ft) ahead of the model (to simulate turbulence). The grid consists of taut strings about 6.4 mm in diameter spaced about 30 mm apart covering the frontal area of the semi-submersible. The test was done with and without the grid to illustrate its effect. The results are provided in fig. 13.3.It is clear that the presence of the grid disturb the flow (as was observed during
- 362. 1013 z Figure 13.2 Towing of a semi-submersibleon a carriage (courtesy Offshore Model Basin, Escondido, CA) z 1.o E 0.8 zyxwvu a m cn cn U .-zyxwvu 8 0.6 = 0.4 rr 8 m Ezyxwvu z 0.2 0.0 0 0.4 0.8 1.2 1.6 2 Model Towing Speed, ft/s Figure 13.3 Towing resistance of the semi-submersihle with and without turbulence screen zy the model test) seen by the model and the net effect is a reduction in the measured resistance. Since the prototype drag coefficient (in turbulence flow) is expected to be lower than the corresponding model drag coefficient (near laminar flow), the net load is expected to be lower. Thus, while the degree of turbulence compared to the prototype is unknown,
- 363. 1014 zyxwvutsrqpon Chapter z 13 1zyxwvutsrqponmlkjihgfedcbaZ Formula Item Bending moment ~ JW,,~ = h5M, the presence of the screen appears to duplicate the prototype flow. Therefore, the measured resistance will be close to the scaled prototype resistance and no additional corrections due to Reynolds number distortion are necessary in this case. Prototype/Model h5 zyx 13.2.6 Cauchy Model Let us now consider another type of structure where the flexibility of its members becomes important. In this case, the member is expected to undergo deformation due to the interaction with waves and this effect should be accounted for in the model for proper simulation. It is often desired to test structures to determine stresses generated in its members due to external forces, for example, from waves. It is well known that for long slender structures, the stiffness of the structure is important in measuring the response of the structure model in waves. In this case, the elasticity of the prototype should be maintained in the model. Hydroelasticity deals with the problems of fluid flow past a submerged structure in which the fluid dynamic forces depend on both the inertial and elastic forces on the structure. Therefore, in addition to the Froude similitude, the Cauchy similitude is desired. The Cauchy similitude requires that stiffness, such as in bending, of a model must be related to that of the prototype by the relation: Bending stiffness 1(EI),= h5(EI), Axial stiffness EA),= h3(EA), Section modulus irI,= h41, (13.8) h5 zyx h3 h4 1 where E = modulus of elasticity and I = moment of inertia. This provides the deflection in the model which is l/h times the deflection in the prototype (Froude’s law); also, stress must be similarly related, such that, zyxwv a , = ha, (table 13.5). Let us consider the example of a cantilever beam. The maximum deflection at the end of the beam is given by zyxwv S , , , = F13 /(3EI)where F is the load at the end of the cantilever Table 13.5 Scaling of structure stiffness parameters for combined Froude/Caucby scale ILung’s modulus iEp =LE,,, ~ ! 5 1 Torsional modulus Stress a , = ha, Torsional rigidity (GI),= P ( G I ) ~ 1 G, = hG,
- 364. Phqsicul Modelling zyxwvuts of Offshore Structures zyxwvuts 1015 z of length I, Equation (13.8) satisfies Froude’s law for this relationship. Since the section moment of inertia satisfies = PI,, (13.9) we have: Ep = hEm (13.10) Thus, the Young’s modulus of the model material should be l/A times that of the prototype. Scaling parameters for different important variables are given as the ratio of full scale to model values in terms of scale factor A in table 13.5. Assuming steel for the prototype material (Ep= 2.07 x 10’ kPa or 30x lo6 psi) and h = 36, the model E, should be 5.7 x IO6 kPa (83,300 psi). Therefore, a suitable material should be chosen with this value to build the model that will be elastically similar. zyx 13.3 Model Test Facilities Since scale models of an offshore structure are used to measure their responses accurately so that they may be applied in a design with confidence, it is important to choose a model testing facility that will fulfill the requirements of a testing programme. The primary purpose of wave tank study is to obtain reliable data by minimising scale effects and measurement error. The model testing facility for offshore structures should consist of the following capabili- ties - model building, instrumentation, simulation of environment and the software to record and analyse data. The physical facility should consist of a basin with the capability of generating waves, wind and current. An efficient wave absorption system is also essential in a basin. The simultaneous generation of waves and current allows the study of their combined interaction with the model. The wind effect is simulated on the superstructure of the model (the portion above the water) and is often accomplished using a series of blowers located just above the water surface near the model. A testing facility should have the following optimum requirements so that a variety of structures may be tested in the facility: Tests at a reasonable scale (1 :50-1 : 100 preferred) Capability of generating regular and random waves over a wide frequency range Wave spreading for certain structures may be preferred Period of waves from 0.5 s to 4 s Height of waves from a few centimetres (inches) to about 0.6 m (24 in.) Towing carriage with a steady speed range of 0.15-3 m/s (0.5-10 ft,s) and capable of carrying a large displacement structure without appreciable structure deformation Wind generating capability with a movable bank of fans Current generating capability with a return flow system Non-contacting motion measurement Underwater video documentation capability The preferred requirements for instrumentation and measurements are described later.
- 365. 1016 zyxwvutsrqpon Chapter 13 z 13.3.1 Physical Dimensions zyxwvu In today’s development of the deepwater fields, it is desirable to have a deep model basin. In fact, for ultra-deepwater, an extremely deepwater facility is required, which is not available. Even the deepest basin in the world is not adequate for the practical simulation of the deepwater depth of today and the full mooring system simulation. Therefore, it is recognised that some distortion in scaling all the important parameters in a model test is inevitable. The choice of scale for a model test is often limited by the experimental facilities available. However, within this constraint, optimum scale should be determined by comparing the economics of the scale model with that of the experiment. It should be kept in mind that too small a scale may result in scale effects and error. Too large a model is often very expensive and may introduce problems from physical handling of the model. When Reynolds effect (such as, presence of drag force) are important, a large scale is recommended to minimise the problem of scale effects. However, the adverse effects of the tank walls must also be considered and avoided in this case. As a rule of thumb, for circular cylindrical structures, the overall transverse dimension should not exceed 1/5th the width of the tank. When larger three-dimensional structures are tested in a wave tank, undesirable transverse reflections may generate in the tank from its sidewalls. This effect may be minimised by introducing lateral wave absorbers along the basin wall. For offshore structure modelling, a two-dimensional wave basin with a mechanical wave- maker is often utilised. There are two main classes of mechanical type wavemakers. One of them moves horizontally in the direction of wave propagation and has the shape of a flat plate driven as a flapper or a piston. The other type moves vertically at the water surface and has the shape of a wedge. In deeper water, a double flapper is often used. A double flapper wavemaker consists of two pivoted flappers, an actuation system driven hydrauli- cally and a control system. For a flapper type wavemaker, the backside (outside the basin) may be wet or dry. Both have advantages and disadvantages, which are taken into account in the design of such a system. The dry-back system appears to be more popular. The wave basin sometimes has a false bottom, which is adjusted to obtain the scaled water depth. In this way a facility may be made suitable for both deep and shallow water testing. Several facilities also have a deeper pit near their middle suitable for slender deep-water structures. A list of representative larger facilities of the world that perform commercially on a contract basis is given in table 13.6. Some overall particulars of these facilities including gross dimensions and capabilities are included in the table. These should be useful to a user for initial screenings. The website addresses of these facilities are also included, which should be consulted for further details about the capabilities of the facilities, They may help a user to obtain additional information to check against their needs for a suitable match. The earlier wave tanks built prior to 1980 only produce waves that travel in one direction. These are suitable for reproducing long-period ocean waves that are unidirectional. Wind- generated multi-directional waves require facilities that can generate multidirectional waves. These facilities have widths comparable to their lengths. Many modern facilities have this capability. These facilities are identified in table 13.6.
- 366. Table 13.6 Selected database on available test basin9 suitable for deepwater testing zyxwv Tow speed 3rrent Wind No. Facility Depth Size zyxwvuts - (m) 545 x 15 height 0.3-5.0 1.o 1 Bassin d'Essais des Carenes, France zyxwvuts w zyxwvutsrqponm ww.iahr.org/hydralah 12 5.0 2. 152 x 30 long and short 1.7m-I5 m 0.9 length 0.5 4.0 ~- CEHIPAR, Madrid, Spain www zyxwvutsrqpo teh/par.es/Engh.h/ Danish Hydraulic Institute, Denmark www.dh/ dk _ _ ~- 3. 3 12 long and short ionc fans 30 x 20 240x 12 79.3 zyxwvutsrqp x 73.2 4. Danish Maritime Institute, Lyngby, Dcnmark www.dunmar.dk 5.5 0.5-7.0 5. DTMB (MASK), CDNSWC, MD www.dt.navy.mil 6. I ionc long and short 0.5-3.0 1.0-3.0 0.5-10 6. DTMB, CDNSWC, MD (Dcep Basin) www.dt.navy.mil 846 x 15.5 6.7 lone long 7. IMD, NRC, Newfoundland www.nrr.ca/imd/ 200 x 12 7 lone long and short fans 8. KRISO, Korca www.kriso.re.kr 56 x 30 long and short long and 'I short IO.x 0.5-5.0 1.5 5 0.45 zyxw 16.0 9. MARTN, The Netherlands (Seakeeping and Manoeuvering) www.murin.nl 170 x 40 (Continued)
- 367. Current zyx (mis) 0.14.4 0.2 0.2 Wind (mh) Fans 10.0 zy ~ fans zy Table 13.6 zyxwvut Continued size i-Depth Centre lholc Pcriods Wave Tow height speed 0.4 3.2 Facility zyxwvutsrq y . , , . . short 45 x 36 0.3-3.0 MARIN, The Netherlands [Offshore) www.murin.nl ~ 0.9 Inone /long 0.8-10 5 80 x 50 50 x 30 90 x 14.6 50x 30 47.5 x 30.5 MARINTEK, Norway www.murintek.sintej.no Shanghai JT Univ., China www..sjtu.eclu.cn I- /long 0.5-3.5 0.74.0 OMB, Escondido, CA www.modelhasin. zyxwvutsrq corn COPPE, Rio, Brazil www.luhoceuno.coppe.u/~j. zyxwvutsrqp hr . . . . ~ 1.0 T Io I long and short OTRC, Tcxas A&M otrc. zyxwvutsrqpo tumu.edu 5.8 116.7 IFmtand 0.54.0 1 I
- 368. Ph~siral Modelling zyxwvuts o f Offshore zyxwvutsrq Strcrcrures zyxwvutsr 1019 z 13.3.2 Generation of Waves, Wind and Current In model testing, the environment experienced by the structure should be properly simu- lated in the laboratory. Two of the major environmental parameters required in offshore testing are waves and wind. The following capabilities are generally requested in model tests: Regular unidirectional wave Wave group Multi-directional wave Wind generation Current generation The following sections will describe the generation of these environments in a testing basin. The important frequency band applicable to offshore structures lies in the range of 5-25 s. The maximum energy of the ocean waves of design importance falls in the area of 10-16 s depending on the severity of the storm. Therefore, the model basin should have the capability of generating these waves with maximum heights (based on the prevalence of wind in the area) at a suitable scale. The magnitudes of these wave parameters are very important in the selection of a suitable scale for the model test. Generation of high frequency wave components at a small scale is difficult in a wave tank. For example, at a scale of 1 zyxwvu :200, a 0.5 s model wave represents a 7.0 s prototype wave. The wave generators seldom have quality wave generation capability much below 1 Hz. On the other hand, ocean waves at 5-7 s may have significant effect on the dynamic response of floating structures. Random unidirectional wave (including white noise) 13.4 Modelling of Environment The model testing facility should have the capability of simulating the wave, wind and currents commonly found at the offshore sites. The generation of model waves is essential for offshore structure testing. Many deep-water structures experience large current, which may also be an important environment, needed in a model scale. Often, towing of the entire structure and its components with the help of a carriage is used to simulate the uniform current speed on the structure. Wind is simulated by various means as well. 13.4.1 Modelling of Waves Modelling of regular waves is straightforward. The regular waves are given in terms of a wave height and a wave period. These quantities are appropriately reduced to the model scale by the selected scale factor. The waves of the given height are generated by the harmonic oscillation of the wavemaker at the required amplitude. For an acceptable regular wave, the waves should be of near permanent form and the height of the waves from one cycle to the next within the test duration should have minimum prescribed fluctuations.
- 369. 1020 zyxwvutsrqpon Chapter zy 13 z Random waves are generated in the model basin to simulate one of the many energy spectrum models proposed to represent sea waves (see Chapter 3). For the generation of random waves, a digital input signal is computed from the target spectrum taking into account the transfer function for the wavemaker. The transfer function generally accounts for the relationship between the mechanical displacement of the wavemaker to the water displacement, and the hydraulic servo control system. zyxw 13.4.2 Unidirectional Random Waves Two of the most common methods of wave generation [Chakrabarti, 19941in the basin are the random phase method and the random coefficient method. The former is spectrally deterministic while the latter is non-deterministic. The former is straightforward and approaches non-deterministic form for a large number of wave components. This is the most common method of wave generation in a basin and is described here. The random sea surface is simulated with the summation of a finite number of Fourier components as a function of time. Thus, the generated surface profile q(t)having the energy density of a specified (or chosen) spectral model has the form: (13.11) zy N q(t) = zyxwv a, cos(2Jrft +E,) n=l There are three quantities on the right hand-side that should be calculated from the speci- fied spectral model. The quantities a,, zyxwv f ,and E, are the amplitude, frequency and phase of the wave components and are obtained as discussed in Chapter 3 (Section 3.6.3). The number of wave components (N=200 minimum, preferably, 1000) is chosen. The spectral model is subdivided into N equal frequency increments as shown in fig. 3.17 having width Af over the range of frequencies between the lower and upper ends of the frequency spectrum, fi and f 2 (cut-off frequencies based on the basin limitations). For each of these frequency bands, the Fourier amplitude a(nAf) is obtained from the spectrum density value S(nAf)as a, = a(nAf) = J2S(nAf)AJ n = 1, 2, ... N (13.12) The frequency f n is chosen as the centre frequency of the nth bandwidth in the spectral model. The corresponding phase E, is created from a random number generator with a uniform probability distribution between -nand +n.The quantityf, is, sometimes, chosen arbitrarily within the nth bandwidth to provide further randomness. 13.4.3 Multi-directional Random Waves In the case of a directional sea, the directional spectrum is obtained as m zyx Q ) = s ( f ) m Q ) (13.13) where the spectral energy density S ( f ) is the same as used for the unidirectional sea. The directional spreading function usually has the same form for all the frequencies in the spectrum. The directional spectrum has been discussed in Chapter 3 (Section 3.6.4).
- 370. Physicul Modelling of Offshore Structures zyxwvuts 1021 z For directional seas, the simulation of the surface profile is shown in Chapter 3 (Section 3.6.5). It is important to calibrate the wave a priori so that the appropriate waveform, which satisfactorily matches the spectral shape, may be duplicated. Waves in the basin should be calibrated without the presence of the model. The random waves should be generated from the digital time history signal computed from the desired spectralmodel and the generated spectra should be matched with the theoretical (e.g. P-M or JONSWAP) spectral models. Once an acceptable match is found (e.g. fig. 3.19 or 3.21), the setting will be saved for later use. This will assure repeatability of the wave spectrum from one test run to the next. This repeatability in the wave generation is very important for the success of the test programme and should be satisfactorily demonstrated by the model basin. Sufficient duration for the random waves and white noise will be given during the test runs so that reliable estimates of the spectra and transfer functions as well as the short-term statistics may be made. Random sea state records should allow test duration of 30 min (prototype) for test data. For slow drift tests, a duration of 120 minutes (prototype) is recommended. The acceptable tolerances for the wave parameters are as follows: zyx Regular Waves Average wave height H, of a wave train consisting of at least 10 cycles: Tolerances: z f5% Average zero-up-crossing wave period T Tolerances: f0.2 s full scale. Irregular Waves Significant wave height H,: Tolerances: 1 5 % Spectral peak period Tp:Tolerances: f0.5 s full scale Significant part of the measured spectrum shape: maximum offset &lo% 13.4.4 White Noise Seas White noise spectra denote wave spectra with near uniform energy over the full range of wave frequency of interest. Sufficient energy content is required from say 5-20 s prototype to obtain significant motion response of the structure. The white noise is difficult to experience in a physical system at sea. However: the advantage of white noise spectrum is that it allows to spectrally analyse the response signal and develop response transfer functions. phases and coherence over the given wave period range in one single run. The method of data reduction is straightforward with the help of cross-spectral technique [Bendat and Piersol, 19801. The generation of white noise with significant amount of energy over a wide band of frequencies is a difficult task at any model basin. The energy level, particularly at the two ends, tapers down and hence, it is, sometimes, referred to as pink noise. The overall energy level is necessarily low and the response will be expected to be small as well. What may make the data reduction difficult is that most floating structures will respond to the white noise with a slow drift oscillation of significant amount. This will require possible digital filtering of data at the low frequencies. The cross-spectral analysis does not need filtering and the reliable range of transfer function is determined from the high value of coherence. The areas of low coherence are eliminated from the RAO.
- 371. 1022 zyxwvutsrqpo Chapter zy I3 Generally, this method provides reasonable accuracy and has the advantage of obtaining the transfer function from one single test run. However, it is recommended that at least limited number of regular waves may be tested to verify these values of transfer func- tion from the white (pink) noise run as well as any anomaly observed during the random wave tests. zyxwvuts 13.4.5 Wave Grouping The motions of floating structures are found to be sensitive to wave groups. The wave group is defined by the envelope wave of the square of the wave elevation. The slow drift oscillation, which depends on the difference frequency, is shown to differ significantly by the groupiness present in the irregular wave. In other words, two spectral realisations, having same energy contents, but different group spectra, will yield two different low- frequency responses. Therefore, it is important to model the groupiness function in the generated wave in addition to the spectral shape. The groupiness function is computed from the squared integral of the spectral density and represents the relationship among the difference frequencies within the energy density spectrum. This function should be computed for both the spectral model and the simulated wave in the basin during its calibration. The two should be matched as closely as possible to insure proper grouping of waves for the slow drift oscillation tests. An example of the comparison of the groupiness function for a JONSWAP wave is shown in fig. 13.4. z 4 x E-3 zyxw 3 zyxwvu < m A 4 / H r ) zyxwv 2 pa a -8) zyxwvutsr I D -5 1 -8 I -5 Frequency, Hz Figure 13.4 Comparison of wave groupinessfunction for a JONSWAP spectrum
- 372. Physical Modelling of OSfshore Structures zyxwvuts 1023 z 13.4.6 Modelling of Wind The wind loads on the structure may be particularly important in the design of such structures as a floating moored structure. However, it should be emphasised that these loads in the model system are limited by the associated scaling problems. Wind loads are functions of Reynolds number and Re is (an order of magnitude) smaller in the model compared to the prototype Re. Therefore, it is possible that the prototype wind effect falls in the turbulent region while the corresponding model wind effect is in a laminar region. In this case, the model test results may be considered conservative. This is why oftentimes, it is the properly scaled mean wind load that is simulated rather than the wind speed. In other words, wind speed in the model is adjusted so that the mean model wind load is matched. Wind may be generated with blowers positioned strategically in front facing the model. In this case, the model superstructure must be accurately modelled. While wind velocity is often taken as a steady value, the wind spectrum may be important in some applications. The frequency range of a wind spectrum is quite broad-banded, often covering a range from 0.005 to 1 Hz (see Chapter 3 for the description of wind spectrum model). While a bank of fans is most commonly used in generating wind over a model super- structure, there are several other methods of simulating wind effect on offshore structures. The earliest method of simulating a steady wind load on an offshore structure is still in use. It is achieved with a weight hanging in the direction of wind from the model with the help of pulleys and strings. A fan mounted on the deck of a floating model has also been applied to simulate steady and oscillating wind load. The pros and cons of these three wind simulation methods in a model scale are compared in table 13.7. Each method has its place in model testing. To demonstrate the capabilities of the model mounted fan, we present an example of a wind condition that was considered in a model test of a floating platform. Table 13.8 summarises the parameters that were used in wind spectral simulation. There are two separate data files needed for the test. The first file is the time history of the signal that is used to control the pitch angle of the generator blades. The second one is a time history of the load computed from the wind drag formula [equation (4.13)]. This load is considered appropriate for the generation of wind load on the model. The measured load during the test is compared with the computed one. Adjustments are made in the control signal to obtain a satisfactory match. For the above example, the desired and computed wind load spectra are shown in fig. 13.5. The comparison between the two in the load spectra is considered acceptable in this case. 13.4.7 Modelling of Current In many deepwater locations considerable current prevails. In the design of structures in these locations, the effect of current may be extremely important. For such structures, simulation of current in the wave basin is essential. While current may be simulated well with towing the model (already discussed), this method may not be adequate for many offshore components. There are several options in generating local current in the general area of the model in the test basin [Chakrabarti, 19941. The current generator should, in general, have the
- 373. 1024 Variable (see Chapter 3) Reference elevation (m) zyxw 1 Mean velocity (mk) zyxwv Chapter zy 13 Wind condition 10 29 1zyxwv Table 13.7 Comparison of wind generation methods Peak frequency coeff. RMS velocity Different Simulation Methods Weights hung from the model superstructure 0.025 0.15*V (lh, zy ZR) Fixed bank of fan Fan mounted on the model Pros Simple to implement No detailed superstructure model needed Generates the steady load as well as the spectrum Generates the steady load as well as the spectrum More accurate generation of the wind possible No modelling of the superstructure Easy to change wind heading Cons Simulates only the mean wind speed Can add some inertia to the model accurately over a large area Need a precise model of the superstructure zy 0 Difficult to generate Fan becomes part of the model and its inertia should be included as part of model heading with model yaw the scaled wind load Causes change of wind Large pitch angle changes Table 13.8 Wind parameters used in the example 1Area (m2) 13884 1 1Air density (kg/m3) 11.21 j 1Starting frequency (HZ) IO I 1 1Ending frequency (Hz) 10.1 i IFrequency components i4096 i 1Sampling period (ms) 154 I following capabilities: The hardware to generate current is reasonably transparent and has minimum influence on the waves generated simultaneously in the basin. The current in the region of the model is reasonably steady.
- 374. Physicul Modelling of Offshore Structures zyxwvuts 1025 z 4 zyxwvu 3 zyxwv f zyxwvutsrqp 0 0.00 0.02 0.04 0.Q6 0.08 0. z z FmFtlCy, zyx rr;r Figure 13.5 Comparison of target and measured wind load spectra In any physical generation of current in a basin, some turbulence is present. Small turbulence may be acceptable, as it will better simulate the prototype situation and minimise the effect of distortion in the model Reynolds number. The current profile is extended to a desired depth over the width of the model region. 0 Some vertical shear in the current profile is possible by selectively throttling the flow. The modelling of current in a laboratory test with or without waves is an important consideration in offshore structure modelling. This capability in a wave basin allows studying the wave-current interaction on the model. In current modelling in a facility, the uniformity and distribution of current should be carefully investigated. The generation of current is simplified if a closed loop is placed in the facility. It is one of the most desirable methods and is achieved by pumping water into and out of the two ends of the tank by a piping system. If a false bottom exists in the facility, underwater pumps can circulate the water in a loop above and below the false floor. Counter-current is generated by reversing the flow. If an installed current generation is not available. local currents are often generated by placing portable current generators in the basin. These may take the form of series of hoses with outside water source or portable electric outboard motors. Uniformity of flow is achieved by proper control of the velocity. For a local current generation, a manifold may be created over the area covering the width and depth of the model in the basin. The manifold may consist of small-diameter PVC pipes of adequate size and number through which flow can be generated. The manifold is
- 375. 1026 zyxwvutsrqp BUTTERFLY VALVE ps z Chupter 13 I WAVE TANK FLOOR zyxwvu SLAB Figure 13.6 Technique for generation of shear current in test basin zyx supported on a structure and hung from the carriage above and placed, say, about 3 m ahead of the model. The flow is created and controlled by a controllable pump. The water is circulated from an intake pipe from the wave basin. Flow straighteners, such as a tube bundle, may be accommodated in the basin to stabilise the flow as long as they do not interfere with the waves. Individual controls are provided in manifold at each elevation with valves so that the flow through them may be individually controlled. It is possible to generate some vertical shear in the current profile by selectively throttling the flow. This is illustrated in fig. 13.6 with the butterfly valves on flow strengtheners on a false bottom of a testing basin. If current is inadequate or unavailable, it is sometimes simulated by attaching the model to a towing carriage and towing the model at steady speeds down the tank with or without waves. While towing does not duplicate the current effect exactly, it is generally considered acceptable for steady currents. A thorough calibration of the current generation should be performed before the model is placed in the basin. The uniformity and distribution with depth of the current profile is established at the test site by a series of current probes placed vertically. The temporal variation of current should be limited to 10% or less for a steady current test. The current velocity required for the test will depend on the scale factor and is generally scaled with Froude scaling. If current is an important consideration in the testing, the scale factor should be chosen such that the available current can simulate the desired environment. z 13.5 Model Calibration While the calibration of the environment is being carried out in the basin, the following calibration procedures may be simultaneously undertaken with the model itself for a
- 376. Phjsical Modelling zyxwvutsrq of Offshore Structures zyxwvuts 1027 floating structure. The completed model is weighed without ballast. The centre of gravity as well as the natural period in pitchlroll of the model in air are determined. Calculations are performed to determine the amount and location of the ballast to achieve the necessary properties of the model. These ballast weights are placed in the model and the location of CG and the natural period of the model in air is verified on the calibration table. For a non-rigid model, the actual stiffness of the model should be carefully determined and compared to the computed stiffness. For verification of numerical modelling software, it may not be necessary to match the computed stiffness very closely as long as the model stiffness is established well. zyxwvu 13.5.1 Measurement of Mass Properties The mass properties of the structure model are measured using a specially built calibration table (fig. 13.7). The table is designed to accommodate the largest structure expected to be tested in the basin. We describe here one such table being used at the offshore model basin (OMB). The tilt table has a large bed to hold the structure. The table is set on a pair of knife-edge fulcrums at its central axis such that it is free to swing in a vertical plane. The position of the fulcrum is adjustable in the vertical direction. The use of counter weights allows the tilting platform to be balanced at any fulcrum adjustment. At each fulcrum adjustment, counter weights are moved to align the CG of the table with the tilt axis. The table is attached to coil springs at its edges (see fig. 13.7) with known spring constant. The weight, centres of gravity, and pitch and roll radii of gyration of the model structure are measured with the help of the calibration table. Before the model is placed on the tilting Figure 13.7 Setup of the model on calibration table (Courtesy of Offshore Model Basin) Next Page
- 377. 1028 zyxwvutsrqpon Chapter z 13 z table, the proper model displacement is achieved by ballasting the model to the desired draft with specified weights. Next, the table height is adjusted such that the desired KG of the model is measured between the knife edge fulcrum and the top of the table. Then, the ballasted model is centred on the platform table. Ballast is arranged vertically to arrive at the model CG on the tilt axis of the table so that the model and table have the same KG. The radius of gyration (kJJ) of a mass zyxwv (m)is defined by k,? = zyxwv Jl/m (13.14) where I is the moment of inertia of the mass about an axis of interest. The moment of inertia of the table is defined as (13.15) where K,. is the rotational spring constant and TJt) is the period of oscillation of table for small angles. The pitch radius of gyration is set using the tilting table restrained by the spring system. The natural period of oscillation of the tilt table alone is measured first about the tilt axis. The moment of inertia of the table without the model is then obtained using the relation in equation (13.15). The moment of inertia of the platform with the model is measured by observing the natural period of oscillation of the system with the model. The moment of inertia, and consequently the radius of gyration of the model structure are computed by subtracting the moment of inertia of the table from the combined inertia of the table - structure system and then using equation (13.14). The inertia of the model alone is then defined as I(m)= I(t +m) - I(t) (13.16) where the local variables zyxwv rn and t stand for model and table respectively. For a long model, a compound pendulum called a bipolar system, may be used to define the roll radius of gyration of the model. The compound pendulum is made out of two single pendulums, one supporting the bow and the other supporting the stern of the model. The moment of inertia of the model about the pin axis is measured by observing the natural period of oscillation. The moment of inertia of the model about the pin axis is defined as: TimgL I . -- 4x2 pin - (13.17 ) where g is the gravitation constant, and L is the distance from the pin to the CG. The moment of inertia of the model about its centre of gravity is obtained from the parallel axis theorem defined as ICG= IPin-mL2 (13.18) The roll radius of gyration zyxwv (kXX) is obtained from the relation: (13.19) Similar expressions may be obtained for the pitch direction. Previous Page
- 378. Physical Modelling zyxwvutsrq of Offshore Structures zyxwvuts 40 1029 1.21 0.0133 56.67 0.0041 13821.95 z Table 13.9 Model ballasting in pitch and roll on tilt table 60 zyxwvuts 1 1.38 zyxwvu I?? (model) = 1446 lb = 44.99 slugs KG (model) = 1.70 ft d = centre of table to centre of placed weight = 2.83 ft zy x = centre of table to deflection measurement point = 3.25 ft 0.0133 56.67 0.0041 13821.95' I (a) Calibration of spring Computed moment of Inertia Desired moment of Inertia 1Load (Ib) 1 Reading (in.) 1 Defl. (ft) 1 Mom. = Load*d 1 8 = Defl., x 1 Mom.iQ 1 Slug.ft2 233.8 250.5 350.2 Slug.ft2 221.7 353.2 IO 1 0.89 1 I20 I 1.05 1 0.0133 1 56.67 1 0.0041 113821.951 1 1.54 1 0.0133 1 56.67 1 0.0041 113821.951 ~ 100 , 1.70 1 0.0133 1 56.67 1 0.0041 113821.951 I I (b) Calibration of model I 1Item 1 Units 1 Table I Roll 1 Pitch 1 1Measured natural period* I s 10.825 1 1.19 1 1.303 1 I 1Measured radius of gyration I ft 1 2.36 1 2.79 1 ~ Desired radius of gyration I ft I 1 2.22 12.8021 * Average over 10 cycles; measured roll and pitch periods include the table An illustrative example of a model calibration on a table is given here showing the details of the calibration of the table and model inertia. The properties of the model and the table springs are given on the top section (a) of the table. The pitch and roll properties (as found from the calibration table) are included in the bottom portion (b) of table 13.9. After the dry properties are known and verified for accuracy, the model is placed in the water and ballasted to the proper draft with ballast weights. The static tests are carried out by adjusting the location of the ballast adjusted to achieve the scaled G M values of the model. The positions of the weights are chosen such that the moment of inertia of the model is relatively unchanged. The natural period in heave, pitch and roll of the model are determined by displacing the model from its equilibrium position and recording its motion with the help of a rotational transducer or an accelerometer. The mooring lines may be calibrated by choosing a short section of each type of material making up the line and determining its elastic properties by a tension test. Care should be
- 379. 1030 zyxwvutsrqpon Chapter zy 13 z taken in choosing the springs for the non-linear mooring system. These springs are calibrated to establish the scaled stiffness of all the individual mooring lines. zy 13.6 Field and Laboratory Instrumentation In model testing, the simulated environment and the model response to that environment are measured. Usually, the test environment is intended to scale a specified ocean environ- ment. In order to verify that the sea state has been properly modelled in the laboratory test, measurements are made with the wave height gauge (e.g. resistance or capacitance wave probe) and current meters. These instruments are commercially available. The instruments are placed near the model to measure the wave profiles experienced by the model. One probe is often placed across from the model in line with its centre to determine the phase relationship between the model response and the corresponding environment. The instruments that are necessary for the successful measurement of the model environ- ment and the responses of the model are described here. 13.6.1 Type of Measurements Structure responses of interest might include environmental loads on a fixed structure, motions of a floating or moored structure and stresses on individual members or components of a structure. The interaction effect of waves with a structure may also be of importance in a design. For example. wave reflection or the run-up of waves on the face of a structure can be an important consideration in the design of an offshore platform. The instruments in these measurements are often specially designed to meet the requirements of the model. For example, load cells are designed to fit between the model and its mounting system in the wave tank in the range of expected loads. Strain gauges are mounted directly on the model surface to measure stresses. The standard instruments that are required during a typical fixed/floating structure test and their applications are listed in table 13.10. Typically acceptable in-place accuracy of these instruments on the model is noted in column 3. 13.6.2 Calibration of Instruments Transducers receive a physical input from the model such as displacement, acceleration, force, etc. subjected to a model environment and produces an equivalent electrical output. The transducer is designed so that this transformation from the measured response to volts is in the linear range for the level of response expected. This allows a single-scale factor for conversion of the output to the required engineering unit. A few common means of measuring an input signal include a bonded strain gauge, a linear variable differential transformer (LVDT) and a capacitance gauge. These components are placed in a trans- ducer stock, which is designed to measure an expected response in a model test. For example, the strain gauge is glued strategically on a tension/compression member of a load cell designed for the desired load range. The load cell is attached between the model and the mounting system. As the model is subjected to waves, the load imposed by the wave on the model is recorded by the load cell. Before this placement, these instruments are
- 380. Table 13.10Typical instrumentsfor offshore model testing zyxwvu Load cells zyxwvutsrqpo _ _ - Ring gauges Strain gauges Instrument IApplication Measures loads between the model parts where attached, c.g. towing loads, and wavc loads on member Measures line tensions of the mooring lines at the fairlead Measures the stresses on the mounting point on the model _ _ _ _ _ _ _ ~ Wave probes Measures incident and Air gap probes Measures air gap between thc frce surfacc and the deck of the model tracking system Motion sensing _____ Light specially built mechanical system to measure six DOF Measures (XYZ) accelerations at the point of attachment, e.g. CG of the model towing carriage Accelerometers Towing speed indicator rowing dynamometer Two component (XU) load cell capable of measuring the towing Accuracy 1/16 in. 1/16 in. 1/8 in. and 1/2 dcg. after application of software l/S in. linear and 1/2 deg. angular 0.1 lb or less depending on load range; zyxwvuts 5% cross talk 0.1 lb 0.01 / l 0.01g ftjs lb Comments Mounted from the air above the water surface Splashing of water introduces inaccuracy in measurement __ Cameras are mounted on the side wall or the carriage Inertia and damping effect of the mechanical system on the structure should be known Attaches between model and fixtures or two members of a model Mounts on model at the mooring h e fairlead May be mounted directly on elastic members Mounted on the model where accelcration/displacement is desired Part of the towing carriage A hinge provided between the staff and dynamometer to allow freedom in pitch
- 381. 1032 zyxwvutsrqpo Chapter 13 z - __* Response TRANSDUCER placed on a specially designed calibration stand and calibrated over the range of expected values. For example, the load cell is fixed on the calibration stand and known standard weights are hung in the direction of measurement from the load cell in increments and the associated voltages are recorded. In case of a capacitance wave probe, the calibration is achieved by placing it submerged at the water surface and moving it up and down in water. The linearity of the instrument is verified and a scalefactor in terms of the response unit per volt is generated. This factor is used to multiply the output voltage during the testing. Each instrument should be checked for waterproofing and calibration prior to setting up the test in the wave tank. The wave probes are calibrated by lowering and raising the probe in still water over the range of water level covering the height of the generated waves for the test programme. The ring load cells used in the mooring lines are calibrated in tension by hanging the load cells vertically and using standard weights over the range of mooring line loads expected. For sectional loads on a structure component, the zyx XYZ load cells are calibrated in each direction on a calibration stand. Cross talks between two orthogonal axes (it. reading on one due to loading on another) are recorded. If the cross talk is high, the instrument should be rejected or re-assembled. In each case, the linearity of the gauges is assured by least square technique and checking the correlation coefficient and standard deviation. The MST (six degrees of freedom motion sensing transducer) is calibrated in each direction, and a calibration curve is developed for each transducer. For pitch and roll angles, the angular potentiometer is turned in steps. For heave, the model mounting plate in MST is raised in steps. For surge and sway, the MST mounting plate is moved forward or sideways in steps. For yaw, the mounting plate is rotated about its vertical axis in steps. Additional calibration checks are performed to demonstrate that the calibrations, polarities and uncoupling software result in measured data, which corresponds to actual displacements by displacing the MST in several directions at the same time. If a non-contacting position tracking system is used, then a complete dry calibration is required for the system on a calibration stand before mounting it on the basin. In-place calibration should also be performed to verify the set-up and the software used for the data reduction for the camera system. Instruments are electrically connected to an automatic data acquisition system (DAS) so that the transducer signal may be automatically recorded. A simple schematic of a data acquisition system is shown in fig. 13.8.The typical transducer signal is such that its output is given in microvolts. It is first amplified by a gain factor to yield a voltage in the limit AMPLIFIER & AID zyxw Y SIGNAL I Figure 13.8 Schematic of data acquisition system COMPUTER DATA BUS
- 382. Physical zyxwvutsrqpo ,ModeNing of zyxwvuts Offshore Structuves zyxwvuts 1033 of 0-5 or 0-10 V. The signal is conditioned for recording, which may include analogue filtering of noise and other unwanted signals and then converted from the analogue to digital form through an AID converter. Unlike analogue signal, digital signals are non- continuous and stored into a computer memory at the specified sampling rate. Today these operations are accomplished efficiently on a desktop personal computer. Instruments for the measurement of responses at a small scale may pose a problem due to its size compared to the model. What creates the inaccuracy in the system is the introduction of superfluous physical phenomenon not present in a larger scale model or prototype, for example, effect of the instrumentation cables, and physical size of instruments. However, many small precise and reliable instruments are available today. The measurement accuracy or instrument sensitivity at a small scale, say 1 zyx : 100, is not a serious problem. The generally accepted overall in-place measurement accuracy is about z 5%. At a much smaller scale, this accuracy may drop down to as much as 20%. For a small- scale testing (smaller than 1: loo), this measurement error must be recognized and considered in the correlation and extrapolation of data. Regular checks of the instrumentation should be performed during testing to confirm that the instrumentation has not undergone any significant changes during the test programme. Checks should be performed each day prior to commencement of data acquisition and whenever the test set-up is changed. Typically, this will include cleaning of the wave probes, re-adjustment of load cell zeroes to correct for drift errors, and simple static tests. zyxwvuts 13.7 Pre-Tests with Model Before the test set-up begins in the wave tank, it should be assured that the proper- ties of the model and associated parts are properly modelled. The following tests, at a minimum, should be performed on a floating structure model. 13.7.1 Static Draft, Trim and Heel Purpose of Test: Record draft, trim and heel. Test Procedure: The floating model is placed in water and the draft, trim and heel are recorded and compared with the scaled values. If there is a discrepancy on the draft. then it is rectified before proceeding. 13.7.2 Inclining Test Purpose of Test: To determine the metacentric height (GM) of the model. Test Procedure: Weights in increments are set at accurately measured distances from the floating model centreline, and the inclinations measured. From these measurements, the metacentric height is evaluated, and compared with that calculated for the model with the specified KG, corrected for the inclining mass. Inclining tests are performed in the transverse and longitudinal directions at increments of the heel and trim angles and the righting moments are determined and verified. Any adjustments are made in the model properties to match the calculated value within 5%.
- 383. 1034 zyxwvutsrqpo Figure 13.9 Mooring line offset test in model zyxw Chapter 13 13.7.3 Mooring Stiffness Test Purpose of Test: zyxwvuts The aim is to measure the restoring force characteristic of the moored model, and to demonstrate that this characteristic is representative of the full-scalemooring system. Test Procedure: With the model moored at the specifiedpre-tensions, the draft is measured, and compared with the expected value. A set-up should be provided (e.g. with a line and pulley system) for applying a known steady horizontal force to the model above the water line. Forces in equal increments should be applied, and the resulting offsets, vessel trim and mooring line tensions are to be measured. The offsets are to include horizontal and vertical components and the inclination. The force should be applied in two directions, the first along the longitudinal direction, and the second in the transverse direction. The expected load range in the positive and negative direction should be covered. An example of the measured restoring force of mooring lines compared to the computed model line forces is shown in fig. 13.9.The data represents scaled-up prototype values. The offset shown is the expected range of offset during the model test. Regarding the measurement system, particularly the motion measurement system, the carriage with the instrumentation system is positioned after the pretension displacement of the model has taken place. This will allow the measuring system to stay within the limits of motion of the model from the wave and slow drift oscillations. 13.7.4 Free Oscillation Test Purpose of Test: To determine the natural periods and damping coefficients of the moored model in free oscillatory modes in six DOF including surge, sway, heave, roll, yaw and pitch. Test Procedure: The model is located reasonably well away from the edges of the tank, to avoid reflections of radiated waves. The model is given an initial displacement one at a time in the selected mode of motion and is released. Time histories of the resulting motions in all six DOF are recorded by MST or motion sensors. The tests are conducted in calm sea conditions. Care should be taken to achieve a nearly pure single degree of oscillation. If the oscillation in any other direction is significant. indicating coupling effect, then the test
- 384. Physical ,Modelling of Offshore Structures zyxwvuts 1035 z should be repeated. This motion time history will provide the natural period of oscillation and damping of the system in the degree of oscillation in question. Note that it is not necessary to measure the displacement of the model in particular. An adequately measur- able response in the oscillation mode from any model-mounted instrument (such as wave gauge, accelerometer, etc.) will provide the desired results. The idea is to record a decaying oscillation from the instrument mounted on the model as the model moves. zy 13.7.5 Towing Resistance Test Purpose of Test: To evaluate the towing resistancelcurrent drag of the complete vessel. Test Procedure: The towing carriage tows the model from one end of the basin to the other. The steady part of the towing speed is used to record the test run. Towing may be performed in waves, as well. Note that the encounter frequency of wave by the model will be different from the generated frequency by the Doppler shift (see Chapter 3), the magnitude of which will depend on the towing speed. For towing in an irregular wave, multiple test runs may be necessary so that the total run length is sufficient for the RAO and other statistical calculations. In this case, the subsequent waves should start where the last one was left off. During towing tests, the quantities measured are towing speed, towing load, and centre of resistance. Towing may also provide the values of drag coefficient for the model. 13.8 Moored Model Tests in Waves and Current In this series of tests the floating model is moored in its permanent in-place condition. The tests are performed in wind, waves and current. Tests are carried out with the vessel and mooring intact. For each test, the environmental conditions are generated, and the behav- iour of the model is recorded with the installed instruments. The following measurements are usually made during such tests: Wave elevation at several positions in the tank to measure wave profile and phasing Six DOF motion response of the floater about a fixed coordinate system Surge, sway and heave accelerations measured at the desired deck level Stresses at several locations on the model if it is flexible Tension in each mooring line Air gap at several model location under deck measured by capacitance probe or similar 13.8.1 Regular Wave Tests Purpose zyxwvuts o f Tests: To establish transfer functions for all measured responses in regular monochromatic wave conditions. To observe any non-linearity in the response transfer function by varying wave height at a few selected wave periods. To define steady state drift force for each regular wave of given period and height. Test zyxwvutsr Procedure: The model is subjected to a series of regular waves. The data sampling rate and test duration are chosen such that the steady state values of the responses may be obtained accurately. Typically, about 10 cycles of steady state data are recorded. The offset
- 385. 1036 zyxwvutsrqpon Chapter 13 z from zero value gives the steady drift force. The second-order drift force on a floating vessel due to regular waves is proportional to the square of the wave height. zyx 13.8.2 White Noise Test Purpose of Tests: loads over the expected range of wave periods. Test Procedure: The model is subjected to a wide band spectrum having nearly equal spectral energy level. The data sampling rate and test duration are chosen such that the transfer function may be obtained spectrally, using a cross-spectral approach. Typically, about 10 min of model scale data will be required for reliable spectral results. To define the transfer functions for the model motion, and mooring 13.8.3 Irregular Wave Tests Purpose of Tests: To establish the behaviour of the complete moored vessel in an irregular sea state with and without the influence of current and wind. Generally, the sea states experienced at the offshore field are simulated in these tests to study the operational and survival characteristics of the system. Test Procedure: The irregular wave runs correspond to the random waves calibrated without the model in the basin. The model is moored with a specified mooring system and pre-tensioned the specifiedamount. In the absence of physical current or towing, the steady load due to current may be simulated with a line and a force transducer attached to the model. The transducer is pre-tensioned a specified amount representing the scaled steady load at the desired point of application. Alternately, the wind and current are physically generated in the basin where such a facility is available. The length of the test run should be sufficient such that a reliable spectrum may be estimated from the measured channels. Irregular wave tests are performed for a period of 120 min full scale to better define the spectral and second-order response characteristics. The transfer functions are computed for the responses from these test runs. They also allow the statistical analysis for the short-term extreme responses. 13.8.4 Second-Order Slow Drift Tests Purpose zyxwvut o f Tests: To establish the quadratic transfer functions for the second-order motions of the moored model. Test Procedure: In addition to the steady drift force, a slowly oscillating drift force is generated on a moored floating structure due to an irregular wave. This drift force is excited around the long natural period of the system from the difference frequencies in the irregular wave components. Thus, a low-frequency response of the vessel is expected covering the frequency band around the natural frequency of the system. This response spectrum due to a random wave is related to the wave spectrum through an integral in terms of a quadratic transfer function. It is often difficult to establish the values of this quadratic transfer function through the spectral approach from irregular waves. Therefore, an alternate technique is recommended to develop the quadratic transfer function for the slow drift motion. Since the slow drift motion appears as the difference in the frequencies in the irregular wave and has a
- 386. Physical Modelling zyxwvutsrq of Offshore Structures zyxwvuts 1037 bandwidth around the natural frequency of the system, this bandwidth can be established from the irregular wave runs. Then, frequency pairs can be chosen from the input wave spectra such that they produce a difference frequency in this band. This will give rise to a symmetric matrix based on the pairs of wave frequencies, which form a wave group. Wave groups from the frequency pair of equal amplitude are generated in the basin with these frequency pairs and the responses are measured. The low-frequency components are filtered through fast Fourier transform (FFT) and the response amplitudes are derived. These response amplitudes are normalised with respect to the square of the wave group amplitudes to give the quadratic transfer function as a matrix. The number of test runs will depend on the width of the response spectrum. This method is quite accurate, since it will directly measure the slow drift response for one difference frequency pair. zyxwvut 13.9 Distorted Model Testing Distorted models are often used in offshore structure testing. The distortion appears in several areas, one of which is model scale. In shallow water coastal engineering, it is quite common to use two different scale factors zyxw - one for the vertical direction and one for the horizontal direction. Because of the limited water depth in the testing facility, the vertical components of a deepwater offshore structure, e.g. mooring lines and tendons are truncated in the model. This distortion should be carefully designed so that the goal of the model test is achieved and the information for the full scale may be derived from the test. This section describes the common distortions found in the model of an offshore structure and the usual remedies taken to correct the problems. 13.9.1 Density Effects In a wave tank, almost invariably fresh water is used to represent the seawater found in a prototype application. This creates a small difference in the density, which is about 3%. This difference reflects a similar change in the measured forces, which need to be corrected. All model weights should be corrected for the difference in water density between that at the test facility and sea water (1025 kg/m3).This is achieved by the ratio of the two water densities. 13.9.2 Cable Modelling In modelling very long cables in a laboratory facility. experimentally realistic diameters should be maintained. This is achieved by combining the proper choice of the elastic material, the role of drag coefficient in conjunction with buoyant devices, and increased kinematic viscosity of the test fluid. Elasticity of a cable/wire (tensile stiffness) is an important property that should be scaled with a suitable material at a small scale. This involves the Cauchy similarity as well as the Froude similarity. The Reynolds number based on cable diameter is involved indirectly with the drag coefficient, zy CD. The requirements for scaling a large cable structure in a laboratory are governed by its length, L, = hL, Once the length scale is chosen, the flow velocity is determined from the
- 387. 1038 zyxwvutsrqponm Chapter zy 13 z Froude number. In addition, the density ratio is fixed, which determines the modulus of elasticity for the model cable, namely, Ep = LE,. Material, such as plasticised polyvinyl- chloride (PVC), can be used in the model cable to provide the required modulus of elasticity. The proper density for the material may be achieved by impregnation of powdered lead. The diameter of the model may be determined by making proper adjustment of the drag coefficient based on the Reynolds number. zyx 13.9.3 Modelling of Mooring Lines, Risers and Tendons The following are the properties for a mooring line or a steel catenary riser (SCR) listed in order of their importance. These should be modelled as accurately as possible. Vertical stiffness There are several alternates we can use to model mooring chains and catenary risers: Vertical and horizontal pretension components Line mass and drag characteristics Horizontal stiffness over the range of anticipated offsets. Model the stiffness curve with multiple springs that is pre-tensioned - drag damping on the mooring line or SCR is not considered here. Use model chain that has the correct submerged weight - the geometry is complex and calculation of drag on the model chain is difficult. Use an outer flexible tube e.g. a thin-walled tygon tubing of a diameter representative of the model diameter and a weighted cable inside - this method provides the scaled drag effect, but is time-consuming in modelling and may have large bending stiffness. Use a plastic rod of suitable material of correct submerged weight per foot. Segment the rod in about 1-2 ft length connected by eye hooks. This will provide a uniform diameter for the lines and risers (except for the small area of the hooks). It is easier to build and still provide a reasonable estimate for the drag coefficient. 0 The most important property of a chain is its weight (per unit length). The material and size of the model chain can be chosen such that the weight can be achieved at a small scale. The elasticity of the material should be verified to ensure the order of magnitude. The geometry of the chain is difficult to scale, which introduces inaccuracy in simulating hydrodynamic damping of the chain generated from its own motion as well as from the wave and current action in the upper part of the ocean. Hydrodynamic damping of the mooring line has been shown to have a significant effect on damping for the low-frequency response of the floating structure. Distortion due to truncation in length is provided by additional springs. Means of correcting for the damping effect from the truncated chain may be introduced in the model chain. This area of truncation is discussed further later. 13.9.3.1Truncated Mooring Line Simulation The (taut and catenary) mooring system needs careful attention since the dimensions, and especially, the depth of the tank, often do not allow a direct scaling of the geometry of a
- 388. Physical Modelling of Offshore Structures zyxwvuts 1039 z typical deepwater floating structure. This is particularly true today as the exploration and production of minerals are going into deeper water. zyxw A truncated mooring line is a common occurrence in testing models of structures placed in deep water. The stiffness of the missing line segment is modelled by additional springs at the bottom of the truncated line. The truncation appears at the basin floor. The tension and initial angle of the mooring line are matched at the fairlead to the prototype design condition. However, the tension and the bottom angle at the line truncation point rarely duplicate the prototype situation. Moreover, the line angle changes as the floater moves in waves and even the fairlead angle cannot be maintained at the prototype values. The difficulty of modelling and set-up of a floating moored system in a basin arises from the following considerations: zyxwv 0 0 0 The mooring stiffness is often non-linear The fairlead angle changes with time and loading for a given environment The initial line angle requires change as different environmental conditions and model orientations are simulated. The mooring lines are usually modelled such that the correct non-linear stiffness behaviour is achieved at the fairlead connection points. A truncated mooring spread is often considered acceptable as long as the stiffness properties at the vessel are correctly represented. It is often important to model all mooring lines individually. For the success of a small-scale testing, it is important that the mooring system simulation is kept simple and the mooring arrangement does not change with every environment. If the pre-tensioned line force is roughly linear with the line extension, the mooring line may be modelled with a set of linear springs arranged in a straight line. The spring set is attached to a cable to achieve the required length of the mooring line. When required, the mooring line model can be non-linear. The springs are chosen such that the stiffness may be easily adjusted to match the linear slope by adding or removing a set of springs. One end of the cable is attached to the fairlead at the model through a load cell. The other end of the cable is attached to an anchor plate at the bottom of the basin in order to maintain the initial fairlead angle for the particular environment. This bottom attachment point is sometimes brought to the carriage with a pulley system so that the initial angle can be adjusted from the surface. The initial fairlead angle is adjusted in order to match the calculated values. It is under- stood that this angle will change with loading from the environment. Since there is a large pre-tension in most cases, the error in the angle with load will be small. The initial tensions at the fairleads, which are monitored with the help of the load cell located at the end of the mooring lines at the fairleads are adjusted and maintained for different wave headings. The procedure during testing is as follows: the model is moored with the taut mooring system with the initial fairlead angles and the ballasted anchor plates set at pre-marked locations at the basin floor. For the test runs where steady loads are needed, the load is applied with the line and pulley arrangement. The model displaces in the aft direction under this load. The model is pulled back by the anchor lines to its initial position (marked
- 389. 1040 zyxwvutsrqpon Chapter zy 13 z on the carriage). This will maintain the initial position while properly pre-tensioning the mooring lines. In order to simplify set-up changes, the anchor plates remain at the same locations between test runs with different drafts. Under this arrangement with draft, the initial fairlead angles will be different, as expected in the real case. The anchor plates will be re-located for different wave heading. Another possible set-up used in several model basins is a Simple Mooring and Riser Truncation (SMART coined by the Offshore Model Basin). SMART consists of a combination of lines running from the model’s fairlead to ring gauges downward at the elevation angle to a specified weight fastened on the line and then continuing up to a fixed point on a vertical pole. This arrangement is represented in fig. 13.10.SMART is geometry dependent. The restoring force of the SMART system follows the desired stiffness characteristics of the non-linear mooring line. There are four variables. Adjusting the distances A, B and C, together with the magnitude of the suspended weight, it is possible to model the desired mooring line characteristics. The mooring line loads can be decoupled for drag estimates. The calibration and installation of a SMART system is simple and fast. The stiffness characteristics are set by the location of the fairlead and weight. Additionally, the weight contributes a realistic inertia load of the hanging chain, which is, generally, absent in the simulation with springs. The weight introduces some hydrodynamic damping as well. The pretensions are set automatically, which allows dynamic load readings. It is quickly adjustable for model draft, easily rotated to change model heading, and readily towed for the simultaneous simulation of current load by towing. This system has been used by several model basins in various projects, including truss Spars and semi-submersibles. A static offset test was conducted on the model of this system shown in fig. 13.11(a). The measured horizontal forces due to horizontal displace- ments are shown in fig. 13.11(b). The system static offset characteristics are checked in I VERTlC.4L zy Figure 13.10 Mooring line arrangement in model
- 390. Physical .Modelling of Offshore Srructures zyxwvuts 1041 z P zy b zy 2 zyx .C I c
- 391. 1042 zyxwvutsrqpon Chapter 13 z Curved Rail zyxw Figure 13.12 Simulation of the bottom end of a truncated mooring Line the basin by applying a series of horizontal loads (with weights over pulleys) to the model while measuring offset distances and line tensions. Similar to the mooring lines, steel catenary risers can be simulated to match the forces induced from the risers on the model. Risers can either be modelled individually or combined into one system for a group to match the non-linear (at large offsets) restoring force exerted at their attachment point on the model. Decoupling the loads on the individual SMART lines also provides useful information of the mooring/riser-induced moments on the vessel. An alternative arrangement to the above-mentioned truncated mooring line systems is to use a moving system at the bottom joint of the truncated line. The purpose is to allow the scaling of the fairlead angle during the model motion. This is accomplished by using a curved rail of a pre-selected curvature and the line is attached to a set of wheels travelling on the rail (fig. 13.12).The spring simulating the stiffness of the balance of the line length is added as described before. The curved wheel is designed such that the angle at the fairlead changes to the scaled value with the motion of the floater. The size of the mooring line may be set to incorporate the approximate load (and associated damping) experienced by the entire line. This will provide a closer scaling of the coupled motion of the floater and line. One difficulty of this arrangement is the possible additional damping introduced by the wheels moving on the rail. On the other hand, the mooring line on the ocean bottom produces frictional damping. The friction in the wheels may be designed to approximate this effect. 13.10 Ultra-deepwater Model Testing Traditionally, model testing verifies the hydrodynamic response of new systems for oil and gas production systems. It is preferred to perform tests in laboratory basins, which can accommodate the full depth of moorings and risers. For ultra-deep waters, however, the modelling of full-depth system becomes difficult, since no test facility is sufficiently large or deep to perform the testing of a complete floating system with compliant mooring in 1500- 3000 m depth, at a reasonable model scale. In this case, the validity of truncation described in the earlier section may be questioned. Various procedures have been proposed and developed to meet this challenge in ultra-deepwater testing. Some of these are: Physical model tests of complete system - Ultra small-scale testing zy (A > > 100) Passive Equivalent Mooring Systems [see Buchner 19991
- 392. Physical Modelling zyxwvutsrq of Offshore Structures zyxwvuts 1043 OMB 5 zyxwv 1 500 OTRC 5.8 580 COPPE 15 ~ 1500 zyxw 0 Active Equivalent Mooring Systems - e.g. Active Truncated Line Anchoring Simulator (ATLAS) [see Buchner. et a1 19991 Outdoor large-scale model tests at sea or in lakes Field tests in full scale Numerical wave tank, e.g. computational fluid dynamics Combination of model tests and computations The actual choice may depend on several factors, such as the type of structure to be modelled, most important parameters to be studied. and the environmental conditions (depending on the location, etc.). The last procedure in the list above combines model test at a reduced depth coupled with computer simulation. This is termed hybrid method. Some of these alternatives are briefly discussed here. 1000 1500 1160 1740 3000 4500 zyx 13.10.1 Ultra Small-scaleTesting As discussed earlier, the first alternative (Le. complete system modelling) is considered to be the most direct, independent method for determining model response. Considering the size of the existing model testing basins, a very small-scale model is needed for testing a complete system in deep water. This scenario is illustrated in table 13.11 where, as an example, the available depth of the basin versus equivalent model scale depth is shown for a few available ocean basins. As can be seen, a depth of 3000 m will require a scale factor of 1:200 for a complete model in the deepest available basin of the world. This scale should be considered limiting for a full model test. At a scale of 1:300, the modelled depth goes up to 4500 m. For use of ultra-small scales, one has to assure that the uncertainty of results is within specified acceptable levels, and there is a need for quantification of these uncertainties. Some of the practical restrictions are the reduced repeatability of waves, currents and wind modelled at very small scales (1 : 150 and smaller). This may be improved if small portable generators are used closer to the models. On the other hand, their presence may have a direct influence on the model response. Several additional areas of concern may be stated as follows: Table 13.11 Available prototype depth in different ocean basins Model Basin ~ Ava:rle ~ l l y : : ~ 1 : : ~ y:: ~ DTMB MASK depth (m) MARIN 10 I 1000 2000 3000 MARINTEK 1 10 1 1000 1 2000 1 3000 1
- 393. 1044 zyxwvutsrqp Chapter zy 13 Accuracy related to model construction Scaling of geometry and mass properties, and response levels Accuracy of instrumentation Possible influence of instrument probes and cables on model response Generation of environmental condition, and capillary effects Viscous scale effects Increasing importance of current loads Damping and inertia effect of the mooring and riser systems A few tests were performed in multiple scales in the same basin so that the scale effects may be studied. An FPSO was tested in scales of 1 zyxw : 170 and 1:55 and comparisons of results were made [Moxnes and Larsen 19981. A similar study was made with a semi-submersible [Stansberg, et a1 20001, where tests in scales of 1:55, 1 : 100 and 1: 150 were compared. Particular care was needed during the planning, preparation and execution of these model tests, since the required accuracy is at a level considerably higher than for conventional scales. The experience from these studies shows that model testing in ultra small-scales down to 1: 150-1 : 170 is, in fact, possible, at least for motions and mooring line forces of FPSOs and Semis in severe weather conditions. For floating systems, not requiring a large “footprint” area on the bottom, such as TLPs, tests in deep pit section of the wave basin may be an alternative [Buchner, et a1 19991.It is, however, difficult to generate a specified current over the entire depth in that case. zyxwv 13.10.2 Field Testing A field test of large models or prototypes, of course, is one method for the verification of design tools. Note, however, field experiments are very expensive and complex, are not guaranteed for success and are at the mercy of the environment. Brazilian Oil Company, Petrobras is a pioneer in deepwater development using many first-of-a-kind technologies. Their philosophy has been that the field experience will prove these technical firsts. Testing in fjords or lakes is another alternative to basin tests, and presently may be the only one, without having to compromise on scale and system simplifications. For research projects, and for use as reference check (benchmark test) of the numerical computations of specific details, testing in fjord is a very attractive alternative. Examples are reported in Huse et a1 (1998) and in Grant et a1 (1999). The main problem of Fjord-testing is, of course, that the environmental conditions are not controllable and, therefore, cannot be used on a routine basis as a design tool. In conjunction with an installed technical facility at sea (e.g. a floating dock, a wavemaker, a top-end actuator, etc.) it may be possible to bring in some control of the environment, even though they will be expensive to install. An at-sea test of the small deepwater semi-submersible, called Motion Measurement Experiment was performed by the US Navy at a site with 900 m (2910 ft) water depth off the coast of Port Hueneme, California. The submersible was proposed as an unmanned Navy facility to support offshore aircrew combat training programme. A three-point mooring system was used in which each line comprised of chain platform pendant, poly- ester line, anchor chain and anchor. The reason for the full-scale testing for this system is obvious because of the deep water and small structure size. It was expected that the dynamics of the mooring lines themselves would have a substantial coupling effect on
- 394. Phj zyxwvutsrqpon srcd ModeNing zyxwvutsrq of Offxshove Stt L K ~ U I es zyxwvuts - - z Subsurface Buoy zyxwvuts Positioning Acourt' Figure 13.13 At-sea test of a small semi-submersible [Shields, et al. (1987)j 1045 z the motions of the semi-submersible. The test set-up is shown as a schematic in fig. 13.13. The environment was measured by a directional discus buoy, wave staffs and electromag- netic current meters. The platform responses were measured by a motion sensor package including accelerometers and rate gyroscopes. Shackle load cells measured mooring line loads. The platform experienced a storm with significant height as high as 8 m, which was close to the design wave height for the system. The advantage of a full-scale testing is that generally minimal scaling effect is involved in the measurements. However, full-scale testing has its own drawbacks: Such testing can only provide feedback on the design after the structure has been built, but fails to provide information at the design stage. There is little control on the environment, so that the structure seldom can be tested in survival conditions such as a design storm. The environment on wind, waves and current are not well defined and are at the mercy of nature. The interference effect of the offshore structure with the instrument and the measure- ment accuracy is not known. From the point of view of cost and practical aspects, the testing is often limited and only a small number of measurements are possible. The reliability and accuracy of full-scale measurements are influenced by large loads, vessel-mounted instruments and vessel motions, the difficulty with the reference values (zero values, position reference) and external aspects (wave directionality, turbulence, temperature, etc.).
- 395. 1046 zyxwvutsrqpon Chapter zy 13 z 13.10.3Truncated Model Testing zyxwvu We have already discussed the truncated system in Section 13.9.3 in which mooring lines and risers are truncated. In designing truncated systems, one needs to apply an efficient methodology in choosing the right system. For example, one should apply an optimisation technique to establish a truncated system with the required properties. The method has to consider at least the following items: Uncertainties in model scale versus uncertainties introduced by the gap between full- depth system and truncated system The importance of interaction effects between the mooringlriser system and the floater motions More important loading effect, e.g. wind, waves, current, VIV etc. Room to explore unknown effects in the test setup. There is also a need for general guidelines to help set the criteria for the requirements for the properties of the truncated system. These requirements are dependent on the system (and site) in hand and have to be evaluated on a case by case basis. 13.10.4Hybrid Testing A realistic alternative is the use of a hybrid form of testing. In this case, the challenge for the design verification of a deepwater system is to apply model tests and numerical computations in such a manner that the reliability is ensured and the critical system parameters are verified at an “acceptable” level of accuracy. Reliability analysis will quantify the effect of the uncertainties. Ultimately, the accuracy of the design verifica- tion must be reflected in the selection of the level of the safety factors in the design of the deepwater system. Of course, for a cost-effective design, these safety factors should be optimised. Another important issue is the very long natural surge/sway periods of deepwater systems and their impact on the procedures used in statistical analysis for the verification. For hybrid verification, the complete modelling is replaced by a hybrid modelling, which introduces an uncertainty gap. The question is how to know that the final simulations give the same results as would have been obtained from a complete model test. Proper model scale and proper truncated set-up should be chosen to reduce these uncertainties. A schematic illustration of how the uncertainty of the verification process depends on the model scale and the degree of truncation is given in fig. 13.14. It qualitatively shows that the uncertainty increases in physical modelling as the scale factor increases, while the uncertainty in the hybrid system increases with smaller value of the scale factor. Therefore, an optimum scale factor shown by a range in the middle of the intersected curves should be arrived at for the model test. Possible hybrid approaches are discussed in more detail in the following sections. 13.10.4.1 Truncated Systems with Mechanical Corrections The simplest approach with a truncated system is the one without computer assistance at all. This has been discussed already. In this case, all connections to the full depth system is incorporated passively in the model test set-up itself, by means of springs, masses and
- 396. Physical Modelling zyxwvutsr ofzyxwvutsrq Offshove Structures zyxwvuts 1047 z Figure 13.14 The balance between uncertainties related to truncation and to small scales [MARINTEK, 19991 mechanisms connected to the floater. Although static characteristics can be modelled quite well by this method [Clauss and Vannahme, 19991, it has been found [Dercksen and Wichers, 1992; Oritsland, 1996; Chen, et a1 20001 that it is difficult to combine a proper line dynamics that reproduces floater damping. When such issues are of less significance, this procedure may be considered as an alternative. A passive system involves model tests with truncated system (equivalent mooring/riser system) and subsequent extrapolation to full depth by use of numerical simulations. The main motivation to perform model test with truncated system is to validate and/or calibrate the numerical tool for a system similar to the actual full-depth case. Various procedures have been described for combining a “passive” truncated test set-up with a subsequent off- line computer analysis. For examples, see Dercksen and Wichers (1992), Kim et al (1999), Chen et a1 (2000) and Stansberg et a1 (2000). 13.10.4.2 Hybrid Passive Systems In order to reduce the uncertainties related to an off-line extrapolation of test results from a truncated to the full-depth systems, one should strive at obtaining the same motion responses of the floater as would result from the full-depth mooring. The truncated mooring system should preferably have a similarity to the physical properties of the full- depth system. In practice, the design of the test set-up should follow the following rules, in order of their priority: Model the correct net, horizontal restoring force characteristic Model the correct quasi-static coupling between vessel responses (for example, between surge and pitch for a moored semi-submersible) Model a “representative” level of mooring and riser system damping, and current force Model “representative” single line (at least, quasi-static) tension characteristics. To the extent that these requirements may not be fully realised, the philosophy of the procedure is that the numerical simulations shall take care of the effect of the deviations between the full-depth and the truncated system.
- 397. 1048 zyxwvutsrqpon Chapter 13 z The purpose of the model test will dictate the actual procedure proposed. Thus, if the purpose of the experiment is to study only a specific effect, the main focus of the physical modelling is placed on that particular detail, while other details are simulated on the computer. For example, tests can be run with a single mooring line for a study on line dynamics, or with the vessel moored in a very simple spring system to study only the vessel hydrodynamics. On the other hand, if the aim is to observe the behaviour of the total system, one will try to model the physical model as much as possible, including, for example, individual mooring line models, albeit truncated. In the latter case, the purpose of the tests is to check and calibrate the numerical programme on the whole system, including the vessel and the lines and risers, on the reduced depth system. Subsequently, the full system is executed along with the numerical model with the relevant information in an extrapolated version. There may also be an “intermediate” case, where lines and riser systems are modelled in a realistic way, but where the main focus is still on the floater. The more advanced the available computer programmes, the more “new” information can be expected from the computations. But they will have to be extensively verified a priori against a range of experiments. zyxwv A particular two-step (passive) hybrid verification procedure was developed by Stansberg, et a1 (2000) for numerical reconstruction. Similar ideas have been suggested in Dercksen and Wichers (1992). The principle is illustrated in fig. 13.15, and can be summarised as follows: Design truncated set-up (according to above guidelines) Select and run a proper test programme with representative tests for the actual problem Reconstruct the truncated test (coupled analysis) numerically for calibration and check of the computer code Extrapolate to full depth numerically. For the computer simulations, coupled analysis is generally recommended. Figure zyxwvut Full Depth 13.15 Two-step hybrid verification procedure [MARINTEK, 19991
- 398. Physical Modelling zyxwvuts of Offshore Structures zyxwvuts 1049 z 13.10.4.3Hybrid Active Systems zyxwvu Active hybrid model testing systems make use of real-time computer-controlled actuators replacing the truncated parts of moorings and risers. The system must be capable of working in real time in model-scale, based on feedback input from the floater motions. Thus, the mooring line dynamics and damping effects are artificially simulated in real time, based on a computer-based model of the system. System identification from model tests of a single mooring line can be used as input to the computer model. A feasibility study with such a system used on a 1: 80 scaled FPSO model moored in a relatively shallow water basin has been described in Watts (1999, 2000). Buchner et a1 (1999) described another system, which might be used in a deep-water basin. In place of a passive system, an active system is installed at the truncated end of the lines. The main features for such a system may include a robot arm on the basin floor (e.g. the MARIN ATLAS system) which will be driven from the surface via an analogue control. The system is designed in such a way that it actively simulates the behaviour of the truncated portion of the mooring lines or risers. The virtual mooring lines (and risers) below the basin floor are coupled to the real mooring lines in the basin. It requires a rigorous computational effort on a real-time basis that simulates the behaviour of the complete mooring (and riser) system. The system can accommodate the soil mechanical aspects of the problem as well. The method can simulate the interaction effect of the mooring/riser system on the low-frequency vessel motions. However, such testing procedure is highly dependent on the accurate performance of sensitive electronic equipment at the basin floor controlled by the numerical simulations. Moreover, the robot arms can induce hydrodynamic effects themselves, which can interact with the mooring and riser system. A complete model test verification system based on these ideas is a challenging, but interesting task for future considerations. It requires powerful computers, as well as well advanced and accurate control systems. The motion range required in 6 degrees of freedom for actuators simulating very deep systems may be another limiting factor. One should also ask: how “intelligent” does the computer model have to be for hydrodynamic verification purposes? It is expected that significant developments will take place in this field in the future. Advantages and disadvantages of this system are: No numerical representation of the floater force model exists. Scaling is taken care of by real-time tests visually resembling “real” model tests. It is difficult to validatejverify correct performance of numerical simulations that control the actuators. Advanced (intelligent) software is needed, requiring rigorous computations. 13.10.4.4 Challenges in Numerical Simulation Whether a passive or an active system is applied, a numerical tool is essential. For an active system, the numerical tool has additional requirements. The computational tool should have the following attributes:
- 399. 1050 zyxwvutsrqpo Chapter z 13 0 Faster and more efficient computers. Real-time feedback requires ultra-fast data computation. Faster and more efficient algorithms in general. Efficient algorithms for time-domain wave kinematics (viscous drift forces and local wave loading on individual mooring lines and risers). Utilisation of multiprocessor hardware. Coupled vs. uncoupled analysis (uncoupled approach needs verification with coupled analysis). Improved mathematical formulation for the floater force model. Formulation of non-linear material properties. Hysteresis effectslenergy dissipation for taut mooring made of synthetic ropes. zy 13.11 Data Acquisition and Analysis So far, we have discussed the modelling technique, scaling methods and measurements. In this section, we briefly comment on the data collection in the test and the analysis procedure that is adopted. The purpose is to obtain technically meaningful results that can be used by the structural engineer in the design of the full-scale structure. 13.11.1 Data Acquisition System The data acquisition system should be automatic using an A/D system to convert the analogue signal from the instruments to digital form. The signal conditioners should consist of amplifiers, switchable filters and bridge sensors. There should be ample data channels available for accommodating all the instruments required for a test. The data throughput capability of all channels should be high of the order of 50-100 kHz. The data collection/reduction system should be such that the data after each test run may be examined within a short time after the run is completed. 13.11.2 Quality Assurance Several steps should be taken to assure that the data collected in the basin during the tests are accurate and that all instruments are working properly. The signal conditioners should be checked every morning to check any drifting of instrumentation. Suspected instruments should be check-calibrated and any problems should be fixed before testing continues. The wave probes should be cleaned periodically to avoid erroneous reading. In-place calibration is performed of all installed instruments to ensure proper measurement and their accuracy. Proper verification of the data acquisition and software routines required for the proces- sing of the recorded data should be made prior to the testing. Several verification problems on the software for the wave generation and data acquisition should be run and the programmes verified. The hardware including the amplifier system should be checked for accuracy using standard calibration technique.
- 400. Physical Modelling zyxwvutsrq of Offshore Structures zyxwvuts 1051 z A known calibration wave should be run daily and checked against specifications. If the resulting calibration signal is outside the specified region, a logical procedure should be instigated to verify the component parts of the wave making process Le. input control signal, wave maker motion, wave probe, logging system and pre-processing. The problem should be fixed before the testing resumes. The collected data should be compared with the standard run made at the commencement of the test in order to make sure all channels are giving similar results within acceptable tolerance. zyxw 13.11.3 Data Analysis Data analysis consists of several steps. All data are collected in the time domain using a suitable high pass filter that removes the high-frequency electrical noise inherently present in the system. All data are normally presented in prototype units using scale factors discussed earlier (table 13.2). Preliminary results of testing should be made available to the client after each test. This should include: 1. zyxwvutsr 2. 3. 4. zyxwvuts 5 . Force vs. offset results after the offset tests Natural period and damping estimates after pluck tests For regular wave tests and white noise tests, motion RAOs plotted for the structure For regular and irregular wave tests, statistics of each channel should be calculated (including mean, maximum, minimum and standard deviation of all responses) Selected time history plots of the data channels as necessary to examine the data quality and trend. The regular wave test data are reduced to obtain the transfer functions (RAO) and plots are presented showing the RAO results for the various motions, sectional loads, stresses on the hull and the mooring line loads. Any problem related to the natural period response of model should be discussed. The offset due to wave drift force is measured and accounted for in the determination of the transfer function at the wave frequencies. The magnitude of the wave drift of the vessel is reported. For irregular waves, spectral energy densities are calculated and compared with the theoretical values. The spectral calculation of the responses is given. The RAO of the low- and high-frequency responses is computed by a cross-spectral method. For channels, which are subject to statistical analysis, the following parameters, at a minimum, should be determined. Significant values Mean periods It is recommended that the design software be executed prior to testing once the model and test conditions are known. These results should be available during the test runs. This allows a direct comparison with the test data during the data reduction while the test is Mean, minimum, maximum, standard deviation
- 401. 1052 zyxwvutsrqpon Chapter z 13 z being executed. This permits uncovering and rectifying possible problems encountered in the test. It also allows redesigning the test to investigate and understand a particular discrepancy between the model tests and design tool results. In addition to the analysis listed above, the following data analysis should be included at a minimum in the final report: 1. 2. 3. 4. 5. 6. 7. The final report should include: Estimate of damping factor vs. response amplitude for 6-DOF motions from pluck tests Comparison of RAOs from white noise and irregular wave tests RAO for response and airgap for each gauge location for regular wave and white noise tests Tension RAOs for mooring lines and risers for regular and irregular waves and white noise tests Plots showing coherence and phase along with all RAOs Time history plots and spectral density plots of all channels for irregular wave tests Extrema1 analysis of responses for the irregular wave tests Model test set-up, coordinate system and sign conventions Detailed drawings and pictures for the model as used in the test List of instruments and their functions Measured mass properties and distributions compared with the computed Wave, current, wind and instrument calibration Test matrix Test results including transfer functions as noted above Any significant visual observation during the test of significance. zyx References American Petroleum Institute (1979). “Recommended Practice for Planning, Designing and Constructing Fixed Offshore Platforms”, API-RPZA, March, Washington, DC. Bendat, J. S. and Piersol, A. G. (1980). Engineering Applications of Correlation and Spectral Analysis, John Wiley and Sons, New York. Buchner, B. (1999). “Numerical simulation and model test requirements for deep water developments,” Deep and Ultra Deep Water Offshore Technology Conference, March, Newcastle. Buchner, B., Wichers, J. E. W., and De Wilde, J. J. (May 1999). “Features of the state- of-the-art Deepwater Offshore Basin”, Proceedings on Offshore Technology Conference, OTC 10841. Chakrabarti, S. K. (1994). Offshore Structure Modelling, World Scientific Publishing, Singapore.
- 402. Physicul Modelling of zyxwvutsr Offshore Structures zyxwvuts 1053 Chen, X., Zhang, J., Johnson, P., and Irani, M. (2000). “Studies on the dynamics of truncated mooring line”, Proceedings on the 10th ISOPE Conference, Vol. 11, Seattle, WA, USA, pp. 94-101. Clauss, G. F. and Vannahme, M. (1999).“An experimental study of the nonlinear dynamics of floating cranes”, Proceedings on the 9th ISOPE Conference, Brest, France. Dercksen, A. and Wichers, J. E. W. (1992). “A discrete element method on a chain turret tanker exposed to survival conditions”, Proceedings on the BOSS’92 Conference, Vol. 1, London, UK, pp. 238-250. Dyer, R. C. and Ahilan, R. V. (2000). “The place of physical and hydrodynamic models in concept design, analysis and system validation of moored floating structures”,Proceedings of Offshore Mechanics and Arctic Engineering Conference, Paper No. OMAE2000/ OFT- 4192, New Orleans, LA, USA. Grant, R. G., Litton, R. W., and Mamidipudi, P. (1999). “Highly compliant (HCR) riser model tests and analysis”, Proceedings on Offshore Technology Conference, OTC Paper No. 10973, Houston, TX, USA. Hoerner, S. F. (1965). Fluid Dynamic Drag, Published by the author, Midland Park, New Jersey. Huse, E.: Kleiven, G., and Nielsen, F. G. (1998). “Large scale model testing of deep sea risers”, Proceedings on Offshore Technology Conference, OTC Paper No. 8701, Houston, TX. ITTC (1999). Environmental Modelling, Final Report and Recommendations to the 22ndITTC. Proceedings of22nd ITTC Conference, Seoul, Korea. Kim, M. H., Ran, Z., Zheng, W., Bhat, S., and Beynet, P. (1999). “Hulljmooring coupled dynamic analysis of a truss spar in time-domain”, Proceedings on the 9th ISOPE Conference, Vol. I, Brest, France, pp. 301-308. MARINTEK (1999). “Deep Water Model Test Methods: Recommendations and Guidelines on Hybrid Model Testing”, Report No. 513137.15.01, Trondheim, Norway. (Restricted). Moxnes, S. and Larsen, K. (1998). “Ultra small scale model testing of a FPSO ship,” Proceedings of Offshore Mechanics and Arctic Engineering Conference, OMAE-98-381, June, Lisbon. Oritsland, 0. (1996). “VERIDEEP. Act. 2.4, Simplified Testing Techniques zy - Type I”, MARINTEK Report No. 513090.45.01,Trondheim, Norway (Restricted). Sarpkaya, T. (1976). “In-line and transverse forces on cylinder in oscillating flow at high reynolds number”, Proceedings on Offshore Technology Conference, OTC 2533, TX, USA, Houston, pp. 95-108. Shields, D. R., Zueck, R. F., and Nordell, W. J. (1987). “Ocean model testing of a small semisubmersible”, Proceedings on Offshore Technology Conference, Houston, pp. 285-296.
- 403. 1054 zyxwvutsrqpon Chapter zy 13 z Stansberg, C. T. (2001). ”Data interpretation and system identification in hydrodynamic model testing”, Proceedings on the 11th ISOPE Conference, Stavanger, Norway. Stansberg, zyxwvu C.T., Yttervik, R., Oritsland, O., and Kleiven, G. (2000). “Hydrodynamic model test verification of a floating platform system in 3000 m water depth”, Proceedings of Offshore Mechanics and Arctic Engineering Conference, Paper No. OMAEOO-4145, New Orleans, LA. Stansberg, C. T., Ormberg, H., and Oritsland, 0. (2001). “Challenges in deep water experiments - hybrid approach”, Proceedings of Offshore Mechanics and Arctic Engineering Conference, OFT-1352, June, Rio de Janeiro, R. J., Brazil. Watts, zyxwvuts S. (1999). “Hybrid hydrodynamic modelling”, Journal of Offshore Technology, The Institute of Marine Engineers, London, UK, pp. 13-17. Watts, S. (2000). “Simulation of metocean dynamics: extension of the hybrid modelling technique to include additional environmental factors”, SUT Workshop: Deepwater and Open Oceans, The Design Basisfor Floaters, February, Houston, TX,USA. Wichers, J. E. W. and Dercksen, A. (1994). “Investigation into scale effects on motions and mooring forces of a turret moored tanker”, Proceedings on Offshore Technology Conference, OTC paper 7444, Houston.
- 404. Handbook of Offshore Engineering zyxwvutsr S . Chakrabarti (Ed.) zyxwvuts 02005 Elsevier Ltd. zyxwvutsr All rights reserved zyxwvuts 105s Chapter 14 Offshore Installation Bader Diab and Naji Tahan zyxwvu Noble Denton Consultants, Inc., Houston, Texas 14.1 Introduction While civil engineering structures are normally built at their installation site, offshore structures are built onshore and transported to the offshore installation site. The process of moving a structure to the installation site involves three distinct operations referred to as the loadout, transportation and installation operations. Collectively, these operations are also known as the “temporary phases” and the engineering work associated with them as “Installation Engineering”. During the temporary phases, the structure is subjected to loads that are different in magnitude and direction from the in-place loads. The shape, the weight and the cost of offshore structures are, therefore, influenced by these temporary phases. The temporary phases also affect the choice of the fabrication yard and the cost and schedule of the overall project. Given a large number of the temporary phase concepts and the numerous types of offshore structures, it would be difficult to present a comprehensive study of installation in a single chapter. The objective of this chapter is to provide the reader with a basic understanding of the most common concepts together with their advantages, disadvantages and limitations. While some design guidance is offered within, the chapter is not meant to provide a comprehensive design guidance on all the installation concepts. For such guidance, the reader is referred to the volumes of technical literature such as research papers, codes or recommended practice, regulatory authority publications and the rules of the classification societies and the marine warranty surveyors. Different types of structures require different methods of transportation and installation. Different installation methods can also be used for the same type of offshore structure. The work presented in this chapter is arranged along the types of structure and the installation concepts.
- 405. 1056 zyxwvutsrqpon Chapter z 14 z 14.2 Fixed Platform Substructures 14.2.1 Types of Fixed Platform Substructures A zyxwvutsr fixed substructure is that part of an offshore platform which sits on the seabed and is rigidly connected to it by means of foundation piles (e.g.jackets) or under the effect of its weight (e.g. gravity base structure). The installation methods of the following substructures are covered in this section: Jackets Compliant towers Gravity base structures. 14.2.2 Jackets The jacket is a space frame structure made of tubular steel members. The jacket legs and braces transmit environmental and topsides loads into the piles and subsequently into the seabed. Jackets typically have three, four, six or eight legs. Jackets with three legs are known as tripods. Jackets with a single caisson type leg also exist. These are also known as monopods. Piles made of tubular steel are installed through the legs of the jacket or through the pile sleeves connected to the jacket legs at its base. The piles installed inside the jacket legs normally extend to the top of the legs. Through leg piles are connected to the jacket legs at the top using shim plates, known as “crown shims”, that are installed in the annulus between the leg and the pile and are welded to both. In some structures, the annulus between the jacket and the pile is grouted, although this is no longer a common practice. Piles installed through sleeves on the outside of the leg structure are connected to the sleeve by grouting the pile-sleeve annulus. Regardless of the size or the type of jacket installation, once the jacket is on the seabed, its weight is temporarily supported by mudmats. Mudmats are added to the bottom of the jacket legs to provide the required bearing area to support the jacket weight and resist environmental loading during installation and until the strength of the piles has sufficiently developed. This phase is known as the “unpiled stability” phase. They are flat panels that are made of stiffened steel plate or, to reduce weight, from glass reinforced plastics. Mudmats are sized so as to support the combined loads of the jacket weight and buoyancy, weight of piles that have to be supported on the jacket and environmental loads associated with the installation window. Section 14.9.4 lists the typical unpiled stability requirements. The method of installation depends on the weight and the physical dimensions of the jacket and on the capacity of the installation equipment. The following methods are the most common for a jacket installation. 14.2.2.1 Lift and Lower in Water This method is used for small jackets, in very shallow water, which are transported on barges in the upright position already pre-rigged for offshore lift and installation by a crane vessel. Once offshore, the jacket is lifted off the deck of the barge and lowered down to the seabed. Jackets installed in such a configuration are typically less than 50 m tall.
- 406. Offshore zyxwvutsrqponm Installation zyxwvutsrq 1057 z The foundation piles for this size of jacket structure are typically transported together with the jacket on the same cargo barge. Once the jacket is set on the seabed, the piles are installed using the same crane vessel and a pile hammer of an adequate size. zy 14.2.2.2. Lift and Upend As the size of a jacket structure increases, it is built and transported on a cargo barge in the horizontal position. The jacket is lifted off the cargo barge using one or two cranes. Following pick-up, the cargo barge is withdrawn and the jacket is upended. Single cranes with two blocks can be used for upending smaller jackets with the jacket length aligned with the plane of the crane boom. There are several methods of upending jackets: Two-block upending - upending in air or partially in water using two crane blocks. In this method, the jacket does not have sufficient buoyancy to float without crane assistance. Instead, the upending is achieved by hoisting down the block of one of the two cranes while the other is hoisted up. Figure 14.1 shows a two-block upend operation. The size of the jackets that can be upended with a single crane is limited. Single-block upending. A jacket installed using this method needs to have sufficient buoyancy to float in the horizontal position by itself. In this method, the jacket is pre- rigged with two sets of four slings. The first set of slings - the lifting slings - are attached somewhere along the top jacket frame, while in the horizontal position. The second set of slings - the upending slings - are attached to padeyes at the top of the legs when the jacket is in the upright position. The jacket is lifted off the cargo barge with the lifting slings and lowered into the water until its buoyancy balances its weight. The lifting slings are then disconnected from the crane hook and the upending slings are connected Figure 14.1 Two block upending (Marathon East Brae jacket)
- 407. 1058 zyxwvutsrqpon Figure 14.2 Single block upend zyxw Chapter z 14 z to the hook. The jacket is then ballasted in a controlled manner until it is upended a few meters above the seabed. Further ballasting is then carried out until the jacket is positioned on the seabed. This method only requires one crane albeit it has to be capable of lifting the full jacket weight without assistance. The jacket legs need to be made buoyant by installing rubber diaphragms at the bottom of the legs and steel caps at the tops. Additional equipment such as flooding valves, umbilicals and pumps are also needed. The jacket buoyancy has to be designed so as to allow easy access for rigging the upending slings, while the jacket floats horizontally. Sufficient buoyancy and subdivision is also required to ensure that flooding of one compartment does not lead to the jacket sinking or making the installation operation impossible to complete. Some consideration should be given to provide remotely operated valves with manual back-up. Figure 14.2 shows a single block upend operation. 14.2.2.3 Launching Jacket structures that are too heavy to be lifted can be launched into the sea off a launch barge. A launch barge is a flat top cargo barge equipped with skid beams, a rocker arm, launch winches and a suitable ballasting system. Jackets are designed to be either self-upending or upended with the assistance of a crane vessel. Launched jackets need to have sufficient reserve buoyancy in order to ensure they float at the end of the launch sequence. The jacket legs are made buoyant by the use of rubber diaphragms at their bottom ends and steel caps at the top. Additional buoyancy located appropriately is sometimes required to achieve the required level of reserve buoyancy or to ensure the jacket will upend itself at the end of the launch sequence. Launching operations require the jacket to be fitted with a launch truss. The launch truss is an integral part of the jacket structure and serves to transfer the weight of the jacket into the skid beams and the rocker arm during the launching operation. The weight of the launch truss normally constitutes a significant part of the jacket weight.
- 408. 0 zyxwvutsrq ffsshore Insraiia rion zyxwvutsrq 1059 z Figure 14.3 Launch simulation of a self-upending jacket zyx The rocker arms are two beams that are installed at the stern of the barge in line with the skid beams. They are connected to the stern through hinges. The rocker arms serve to support the jacket weight as it rotates over the barge stern and dives into the water. As such, the rocker arms and the supporting hinges can be substantial structures. Figure 14.3 shows a typical launching sequence of a jacket that was designed to be self-upending. Sections 14.8.2 and 14.9.3 include more information on launching. 14.2.3 Compliant Towers Compliant towers are made of several rigid steel sections joined together by hinges such that the tower can sway under environmental loads. A compliant tower’s mass and stiffness characteristics are tuned such that its natural period would be much greater than the period of waves in the extreme design environment. This reduces their dynamic response to such environment and extends the applicability of fixed platform to deeper water such as 1000m. A compliant tower structure can be divided into four basic structural components: The foundation piles, The base section, The tower section(s). Depending zyxwv on the water depth and the means of transport, the tower can be made in one or more sections, The deck. The base and the tower sections are lattice space structures fabricated from tubular steel members and thus termed the jacket base and the jacket tower sections. Normally the tower section is much larger than the base section. A typical installation sequence of a single tower section is described next. The jacket base section is transported on and launched off the deck of a launch cargo barge at site. The top of the jacket would be connected to a derrick barge and the bottom to its assisting tugs. Once in water, the jacket base section would be upended by the derrick barge assisted by
- 409. 1060 zyxwvutsrqpo Figure 14.4 Installation of the Baldpate piles [De Koeijer, et al 19991 zyx Chapter z 14 z the jacket buoyancy. Once vertical, the jacket will be lowered and manoeuvred into position often with the guidance of a pre-installed docking pile. Piles are transported to the site on cargo barges, lifted off and upended, using the cranes of the derrick barge, lowered, stabbed through the jacket base pile sleeves and driven to target penetration as shown in Fig. 14.4.Pile driving is addressed in Section 14.4.2 where a more detailed description is provided. After, the verticality and orientation of the jacket base are achieved, piles are grouted to the pile sleeves. The base structure would now be safely secured to the seabed and ready to receive the next tower section. Then, the tower section is transported on the deck of a launch barge and launched into water. Due to the large weight and height of the tower section, it is designed such that it is self-upending after separating from the launch barge and going into the water. Once vertical, the tower section, is ballasted to the required float-over draft. The tower section is then towed and positioned over the pre-installed jacket base section as shown in Fig. 14.5. With assistance from the attending derrick barge, and position-holding by tugs, ballasting continues until the pins at the base of the tower section engage one by one in their respective receiving buckets at the top of the pre-installed base section. Grout is then injected into the gap between the pin and bucket, which provides the structural continuity and the integrity of the entire subsurface structure (base and tower sections). The tower is now ready to receive the topsides deck. The topsides deck can then be lifted by the derrick barge and set onto the tower structure.
- 410. Offshore Installation zyxwvutsrq 1061 z Figure 14.5 Upend and transfer of Baldpate tower section [De Koeijer, et zyx a1zyx 19991 14.2.4 Gravity Base Structures zyxwvu Gravity base structures (GBS) are very large structures that sit on the seabed and resist sliding and overturning loads by friction and soil bearing capacity. The hull of a GBS is made of several tanks that are used to store oil and ballast. The lightship displacement of the gravity based structures can be of the order of several hundred thousand tonnes. GBS have been installed in water-depths of up to 300 m. Most gravity based structures are made from concrete although one steel gravity base platform, Maureen, was installed in the North Sea. Concrete platforms are built and installed in a different way from steel jackets. The construction commences in a dry dock adjacent to the sea. The structure is built vertically. from the bottom up, in a similar manner to onshore buildings. When the structure is complete, the dock is flooded and the structure floats under its own buoyancy. The topside structures are normally installed at an inshore location by deck mating or any other suitable method. Multiple tugs are used to tow the structure to its offshore location. Once on location, the structure’s tanks are filled with sea water to a predetermined ballasting plan and the structure is sunk down to its final position on the seabed. The GBSs are typically trial-ballasted prior to tow to site. GBS are towed at a large draft and their towing requires very detailed analyses and marine procedures including the following aspects: Available water depth, underkeel and horizontal clearances in the tow route. Stability and freeboard. Required number of tugs, bollard pull and design of the towing attachments. Given its size, several tugs tow the GBS at a very slow speed of zyxw 2 knots or less. Table 14.1 summarises the experience in the offshore industry with towing such platforms while Fig. 14.6 shows the tug towing arrangement of one of the early GBS.
- 411. 1062 Beryl zyxwvuts A Brent B zyxwvutsr Chapter z 14 1975 120 m 170 nm 6 days 1975 140 m 170 nm 6 days Table 14.1 Previous GBS towing distances and duration Gullfaks C zyxwvut 1Platform 1Installed IWater depth 1Towed distance and duration 1 1989 216 m 160 nm 6 days IEkofisk 11973 /70m 1216 nm 7 days I Snorre A Frigg CDP-1 I1975 1104 m I120 nm 5 days 1990 I309 m 1180 nm 6.5 days IBrent D 11976 1140 m 1 160 nm 6 days 1 1Frigg TCP-2 I1977 1104 m 180 nm 4 days I 1Statfjord A I1977 1146 m 1220 nm 7 days I 1Statfjord B 11981 I146 m 1220 nm 7 days I 1 Statfiord C 11984 1146 m 1230 nm 7.5 days I IGullfaks A 11986 1135 m I160 nm 6 days I 1Gullfaks B 11987 I142 m I160 nm 6 days I 1Oseberg A 11988 I109 m 1 130 nm 5 days I IDraugen I1993 1251 m 1333 nm 8.9 days I 1Sleipner A 11993 182 rn I156 nm 7 days I 1Troll 11995 1303 zyxw r n 180 nm 6.5 days I 1Hibernia i 1997 180 m I260 nm 9 days i Figure 14.6 Tow of Beryl A GBS
- 412. Ojfshore zyxwvutsrqpon Installation zyxwvutsrq 1063 z 14.3 Floating Structures 14.3.1 Types of Floating Structures zyxwv The most common floating production storage and offload (FPSO) vessels are converted tankers. While most of the new-build FPSO retain the aspect ratios of tankers, their bow and stern hull shapes tend to be more square than the ship-shaped tankers. An FPSO, as the name suggests, supports production and storage operations with some of the largest ones being built today capable of storing 2 million barrels of oil. Given their length-to- width aspect ratio, environmental loading on the beam of the vessel is much higher than that on the bow or stern. Turret mooring systems that allow the FPSO to weather-vane so as to minimise the environmental load are a common choice for station keeping particularly where there is very little directionality in the design environment such as in the areas exposed to hurricanes or typhoons. In environments where the weather is directional, such as in West Africa, or semi-directional as in Brazil, there is scope for using the spread- moored systems for station keeping. Semi-submersible vessels are also referred to as the column-stabilised units. Their most common hull form consists of four columns supported by two pontoons. The pontoons are submerged under normal operations and the only water-piercing part of the hull are the columns. These vessels are commonly used for drilling operations in water depths in excess of 100 m. Several semi-submersibles have also been used as production rigs in deep water. Some of these vessels support the combined drilling and production operations but have no storage capabilities. While mooring remains the most common type of station keeping, deepwater semis are equipped with thrusters that maintain station with a dynamic positioning (DP) system. Semi-submersibles are sensitive to additional weight and increases in water depth as their operating water plane area is small. With a length-to-width aspect ratio close to 1 zyxwvutsr .O, a spread mooring pattern is usually adopted for the semi-submersibles. Conventional tension leg platforms (TLP) have similar hull forms to semi-submersible vessels with water-piercing columns and pontoons. TLPs are anchored to the seabed via vertical tendons that are made of high strength steel pipes commonly joined by mechanical connectors or. less frequently, by welding. Tendon tensions at the operating draft are balanced by hull buoyancy. This system is self-restoring since any offsets from the mean position caused by environmental loads results in a gain in hull buoyancy and tensions and generating a restoring force that pulls the TLP back to its mean position. TLPs have been installed in water depths ranging between 148 m and 1432 m. A conven- tional TLP can support drilling and production operations while the smaller mini TLPs can only support production operations. Because of the high stiffness of the tendons, the TLP motions are much smaller than the semi-submersibles and FPSO. Figure 14.7 shows different TLP configurations. The TLP foundations are typically driven piles although other pile types are feasible. Sometimes, foundation templates are used. Deep draft caisson vessels (DDCV), also known as spars, are an alternative to the TLP in deep water. A conventional spar hull form consists of a vertical cylinder made of a combination of voids and ballast tanks. Truss spars are a variation on the theme with the lower part of the length of the cylinder substituted by a truss structure. Spars are inherently stable as their centre of gravity is located below their centre of buoyancy. They
- 413. 1064 zyxwvutsrqpon Chapter zy 14 z Figure 14.7 Schematic of Seastar, ETLP and Moses TLP designs zyx are normally moored by a semi-taut spread mooring system although at least one spar is currently being designed with a taut leg polyester mooring system. Figure 14.8 shows a schematic of a spar. The spar dimensions vary with the largest built to date being of the order of 150 ft in diameter and 750 ft long. The immersed part of the spar hull consists of a hard tank (usually the mid-section) which provides the buoyancy, and a soft tank at the bottom where the fixed ballast is stored. Figure 14.8 Schematic of a truss spar
- 414. Offshore Insfallation zyxwvutsrq 1065 z The mooring line fairleads are positioned close to the pitch centre which is well below the water line. This minimises the fairlead excursions in rotational movements allowing the mooring system to be reasonably taut which, in turn, minimises lateral excursions. The heave motions are also low because of the spar’s low water plane area compared with its hydrodynamic mass giving low motion characteristics overall. zyx 14.3.2 Installation of FPSOs Although the installation of the topsides onto the hull of the FPSO is considered to be part of the construction phase, the topside integration lifts are often carried out by floating crane vessels making the operation akin to an installation operation. Since the integration lifts are carried out along the quay, or in sheltered conditions, the criteria that are applied to the lift are those for an inshore lifting. The availability and the size of the lifting equipment in the vicinity of the yard is normally a significant consideration when selecting a construction yard for the integration of the FPSO topsides. If the capacity of the available lifting equipment is low, the topsides would have to be split into a greater number of modules of a manageable size. The installation of the topside is followed by a period of a few months during which the installed modules are hooked up to the ship systems or to each other. During this phase, the FPSO has to remain moored along the quay. Whether FPSOs are converted tankers or purpose-built vessels, they are unlikely to have any propulsion, since it is not required during the service life. They are therefore towed to site using at least one tug and, more likely two or three tugs. The mooring system of the FPSO is installed prior to the arrival of the FPSO and laid on the seabed or, in the case of polyester mooring lines, suspended at mid-depth using buoyancy cans. The FPSO is towed over the mooring pattern and the tow switches from the towing configuration to the station-keeping configuration. While the tugs hold the FPSO in position, other tugs pick up the ends of the pre-laid mooring lines and bring them towards the FPSO fairleads where they are connected to winches or chain jacks that are installed on the FPSO. The tugs are released when a sufficient number of lines are connected. The mooring line hook-up operation continues until all the lines are connected and tensioned. 14.3.3 Installation of Semi-Submersibles Drilling semis normally carry their mooring legs and anchors on board. The mooring system consists of a chain, wire or a combination of both. When the vessel arrives on location, its anchors are handed over to anchor-handling (A/H) vessels. The A/H vessel then moves towards the designated anchor installation position while the mooring line is paid out from the semi’s on the winches. The anchor is lowered to the seabed at the designated location. Preloading the anchors and tensioning of the mooring lines is carried out using the anchor handling vessel and the on-board winches. With the introduction of the taut leg polyester mooring systems for semis, particularly in the case of the production vessel, the mooring system can be pre-laid ahead of the semi arriving on site. The semi is then “hooked-up’’ to the mooring legs one by one, using temporary or permanent winches or chain jacks installed on board.
- 415. 1066 zyxwvutsrqpon Chapter z 14 z 14.3.4Installation of Tension Leg Platforms zyxwv The main components of the TLP are the hull, the deck, the piles and the tendons. Pile installation is discussed in Section 14.4.The tendons can be installed ahead of the hull or installed at the same time as the hull. Similarly the deck can be integrated with the hull at the fabrication yard or installed after the hull. This section describes the various stages involved in the installation process. 14.3.4.1 Wet Tow of Hull and Deck Once installed, TLPs derive their stability from the tendons. Free floating stability is deemed to be an issue only during the temporary phases including wet tow and installation. This issue determines at what stage the deck is installed. There are two installation philosophies of TLPs: Installation of Complete Platform The deck is installed inshore at or near the integration site and the completed platform is transported to site. This saves the cost of expensive derrick barges and hook-up and commissioning work offshore. The transportation operation involves a wet tow for at least a part of the voyage. The platform is therefore required to have adequate free floating stability. The hull is designed to provide sufficient buoyancy and the water plane area to meet the stability requirements during wet tow and installation. More recently TLP designers added temporary stability tanks to the hull in order to meet the stability requirements. These tanks are removed once the TLP installation is complete, thus leaving the hull with only the necessary structure to meet the in-place conditions. ABB’s Extend Pontoon TLP (ETLP) is an example of such a concept. This is thought to reduce the cost of the platform as the temporary stability tanks can be re-used and their cost can be spread over several projects. Installation of the Hull and Deck separately This installation philosophy was adopted on several mini-TLPs such as the Seastar and the Moses. With weights of less than 6000 tomes, the topsides can be installed in a single lift and offshore integration time is not perceived to be a significant handicap. A crane vessel is required on-site during the installation of the hull, hook-up to the tendons and deck installation. During installation of the hull, an additional hull stability is often required during hull ballasting for installation. This is achieved by applying an upward load on the hull by the crane hook. 14.3.4.2 Tendon Assembly This section describes the means of delivering the tendons to site and assembly. a. Dry Tow of Tendon Sections and Assembly Offshore A typical tendon string is made up of a bottom section, several main body sections and a top section made of a length adjustment joint (LAJ). The bottom connects through a mechanical connector to the pile or foundation template. The individual sections are joined together with a mechanical connector such as the Merlin connector. The main body sections are typically fabricated in sections of 240-270 ft lengths, shipped on a
- 416. Offshore Installation zyxwvutsrq 1067 z cargo barge to the installation site, where they are lifted and upended by a crane barge. During the tendon assembly process, the weight of the tendon string that has already been assembled is supported on a tendon assembly frame (TAF) which is a purpose built structure that is installed over the side of the derrick barge. The maximum length of individual sections is determined by the available hook height of the derrick barge. Tendon strings with longer sections require fewer mechanical connectors but a larger installation crane boom. Wet Tow of Complete Tendon As an alternative to using tendon connectors, tendon strings are assembled by welding individual sections together. The tendons are subsequently launched and wet towed to the site in the same way as the pipe bundles. Buoyancy modules may be strapped onto the tendons to provide additional buoyancy and control stresses during the wet tow as shown in Fig. 14.9. Once at site, the tendons can be upended with the help of winches or cranes and controlled removal of the buoyancy modules. This method saves the cost of mechanical connectors. The tow operation has to be designed carefully to ensure that failure of any component during the wet tow does not lead to the total loss of the tendon string. b. zyxwvutsr 14.3.4.3 Tendon Hook-up a. Pre-installed Tendons Tendons can be installed prior to the hull arrival to site. To ensure that the tendon and its components remain taut, upright and to keep the stresses within design allowables, temporary buoyancy modules are provided in the form of steel cans which connect to Figure 14.9 Buoyancy modules fitted to the Heidrun TLP tendons
- 417. 1068 zyxwvutsrqpon Chapter z 14 z the tendon at its top as described in Section 14.3.4.4.This temporary buoyancy module (TBM) is clamped to the tendon after the tendon assembly is complete. The tendon and the TBM assembly is lifted from the side of the derrick barge and manoeuvred until the tendon bottom connector is stabbed into the foundation and latched in position. The TBM is then deballasted such that it applies sufficient tension to the tendon until it is hooked up to the TLP. Figure 14.10 shows a schematic of a pre-installed tendon. When the platform arrives on site, it is ballasted until the tendon connector engages the z LAJ teeth at which point the connector is locked off. Once the connector is locked off the connector allows the downward movement of the platform under wave action but prevents any upward movement. This is known as “ratcheting”. The ballasting operation continues in parallel with the ratcheting motions, until the desired draft is reached. At that point the ballast water is pumped out causing the tension in the tendons to increase while the hull draft only reduces marginally by the amount of tendon-stretch. The de-ballasting operations are considered complete when the desired pre-tension is reached in the tendons. Figure 14.11 shows a typical time history of the tendon loads during the ratcheting operation. PONTOON EXTENSION TENDON PORCH PONTOON EXTENSION TENDON PORCH TENDON MIDDLE SECTION TENDON BOTTOM SECTION FOUNDATION RECEPTA Figure 14.10 A pre-installed tendon with a TBM before hook-up to ETLP
- 418. Offshore zyxwvutsrqp lnstuilution zyxwvutsrq 1069 z Tendon Tension I zy I I I 1500 1875 2250 2625 3000 Time (s) Figure 14.11 Time history of tendon ratcheting loads b. Hull and Tendons Installed Concurrently Once the tendon is assembled on site, the derrick barge hands it over to the platform where it is hung from the tendon porches. Once all the tendons are hung from their respective porches, their bottom connectors are stabbed into their piled foundat- ions. Tendon pre-tensioning is achieved using mechanical tensioners similar to chain jacks. The pre-tensioning operation can proceed in several stages with only one group of tendons being tensioned during each stage in order to limit the number of the tensioning devices required. The tendon porches in this type of installation have to be open on one side to allow the tendon to be inserted laterally. This restriction does not apply to the pre-installed tendons. Figure 14.12 shows a schematic of the tendon stabbing operation. 14.3.4.4 Temporary Buoyancy Tanks Where the tendons are pre-installed, temporary buoyancy modules (TBMs) are used to maintain tension in the tendon until the hull arrives. Each TBM is subdivided into several chambers or a cluster of tanks such that the tension in the tendon is not lost with any one compartment getting accidentally flooded. The TBM is normally clamped to the tendon after the tendon is assembled, while it is still hanging over the side of the derrick barge. The TBMs are located below the LAJ such that they do not interfere with the operation of hook-up to the platform. The TBMs are flooded when the tendon is stabbed into the
- 419. 1070 zyxwvutsrqpon Chapter zy 14 Chain jack tensioner Chain paid out from tensioner Hydroulic connector -Tendon porches 1i k . z Y TENDON BOTTOM SECTION ,,-Tendon bottom indexed into latch - z Figure 14.12 Stabbing of tendons hung-off from the Auger TLP (Offshore engineer) zy foundation piles. While the top of the tendon is still supported by the crane hook, the TBM is de-watered by pumping compressed air into it, until the desired buoyancy is achieved. The top of the tendon is subsequently released from the crane hook. TBMs can have a closed bottom or an open bottom. One of the critical areas of design is to ensure that the TBMs have sufficiently large openings in their top to allow air to escape while they are lowered through the wave zone. Once the TLP is hooked up to its tendons and sufficient tension is achieved in the tendons, the TBMs can be flooded so that they become neutrally buoyant and removed by the attending installation vessel. 14.3.5 Spar Installation 14.3.5.1 Wet Tow and Upending Spar hulls and decks are normally installed separately. The hull is normally wet-towed to site and upended by ballasting.
- 420. Ofshore Installation zyxwvutsrq 1071 z Figure 14.13 Upending of Nansen spar [Beattie, et zyxw a1 20021zyx In conventional spars, fixed ballast is added to the soft tank at the bottom end of the hull followed by variable ballast added to the hard tank at the top. A fixed ballast in the soft tank could either be water or hematite. The truss spars are made of a hard tank at the top, a truss section which substitutes the soft tank in a conventional spar, and a fixed ballast tank at the bottom. The ballasting operations can be done by free-flooding the tanks. In this case, large “rip- out” plugs are removed from the tanks to facilitate free flooding. The vent size has to be carefully designed so as to allow the escape of large volumes of air in a very short period of time. Where tanks are ballasted by pumping water into the tank, the pumping rates need to be maximised to ensure a rapid operation that can be completed within a reasonable window. Figure 14.13 shows a typical spar upending operation. 14.3.5.2 Mooring Line Hook-up Mooring lines are installed prior to the arrival of the spar on site and laid on the seabed until the spar arrives or, in the case of the polyester mooring lines, kept above the seabed using buoyancy devices. The spar end of the mooring line is normally made of a chain. This segment could either be pre-installed with the rest of the mooring line or, alternatively, installed during the hook-up operation. Recovery of the mooring lines is normally performed using a crane vessel. The weight of the mooring line dictates the size of the crane vessel required. The connection of the mooring line to the spar is performed using the pull-in winches installed on the spar. A messenger line is deployed from the spar through the fairleads and connected to the end of the mooring line which is supported by the crane vessel as shown in Fig. 14.14. Once the messenger line is connected to the mooring line, the winch pulls the messenger/mooring line assembly back. Once the correct pretension level is achieved, the chain stoppers are locked off. The hook-up operation described above often requires a substantial pull-in system. The size of the pull-in system can be reduced by supporting the weight of the mooring line at an intermediate point close to its top end on a clamp installed on the crane vessel [Dijkhuizen,
- 421. 1072 zyxwvutsrqpo Chapter z 14 z Figure 14.14 Handover of mooring line to Nansen spar messenger line [Beattie,et a1 20021 20031. An equaliser system is rigged up so as to allow the crane vessel and the spar to be winched closer together. The short length of chain beyond the clamp is handed over to the spar for connection. This system is shown in Fig. 14.15. During the hook-up operation, the spar is held in position using tugs connected to the spar hull. These tugs serve to keep the spar on location during the hook-up operation. The tugs have to be sized to resist environmental loads and loads from the mooring lines already connected to the spar hull. zyxwv 14.4 Foundations 14.4.1 Types There are four main types of foundations: Driven piles Drilled and grouted piles Suction embedded anchors Drag embedded anchors Each type requires a different method of installation. Gravity structures may be regarded as a type of foundation, but are considered in this chapter as a “fixed platform”. Refer to Section 14.2.4.
- 422. Offshore zyxwvutsrqponm Installation zyxwvutsrqp 1073 z Figure 14.15 Pulling Horn Mountainspar and crane vessel with equaliser system [Dijkhuizen, zy 20031 14.4.2 Driven Piles Driven offshore piles are steel tubular members which consist of a driving head, the main body of the pile and a driving shoe. The pile length, diameter and the wall thickness depend on the soil characteristics and the magnitude of design loads. Pile lengths to over 500 ft and diameters greater than 96 in. have been installed. 14.4.2.1 Transportation and Installation Piles are normally transported on cargo barges to the offshore location. They may be lifted off the deck of the cargo barge and transferred onto the deck of the installation vessel before the commencement of the installation activities. Alternatively, they can be upended immediately after lifting from the cargo barge. Piles are lifted and upended using two crane blocks or a single block with an internal lifting tool (ILT). The ILT is a specially designed tool, which consists of a mechanical device inserted into the inside of the tubular pile head, with hydraulic pistons which push a set of grippers against the inner walls of the pile driving head and support the weight of the pile through the friction generated between the ILT grippers and the inside wall of the pile head. Other lifting options have been used such as the padeyes welded to the exterior of the pile, at some distance below the top of the pile so as to avoid any interferences with the pile driving hammer. Once in the vertical position, each pile is lowered through the water and stabbed into the seabed or the template structure. For steel jackets piles can be driven through the jacket legs or through the pile sleeves connected to the jacket legs at its base. Thejacket leg and the pile sleeve both act as a guide for the positioning and the directionality of the pile.
- 423. 1074 zyxwvutsrqpon Chapter z 14 z 14.4.2.2 Hammer Types and Sizes zyxwv The most common types of offshore pile driving hammers are steam and hydraulic hammers. The steam driven hammers can be used when driving piles through jacket legs or in shallow water where pile followers may be used which ensure that the hammer remains out of the water. However, with offshore developments moving into deeper waters, hydraulic hammers have been used to drive piles both below and above water. Hammers vary in size, weight and capacity depending on the characteristics of the pile to be driven and the soil properties to be driven into. They can be classified in terms of the maximum energy they can deliver. Existing hammers can drive piles up to 120 in. in diameter. Hydraulic hammers are more efficient than the steam hammers in terms of the energy delivered to the pile and, as such, their energy output needs to be carefully controlled and monitored. The hammer has to be sufficiently large to drive the pile to design penetration in the given soil conditions. Typically the soil conditions considered correspond to both the lower and upper bounds. Also, the pile is assumed to be either “plugged” or “unplugged”. The plugged condition refers to the case where the soil plug inside the pile is assumed to have become an integral part of the pile and moves with it as the pile closes at the bottom. The unplugged refers to the case where the soil plug is assumed to remain in level with the soil outside the pile and that resistance from skin friction continues to develop both inside and outside the pile. The combinations of soil upper and lower bounds and plugged and unplugged behaviour give rise to four cases of analysis, which need to be considered in the pile design. A pile “driveability” analysis is normally carried out to establish the following: Whether the pile can be driven to the required depth with the proposed hammer size/ energy in the four analysis cases. zyxwv 0 Whether the dynamic stresses in the pile exceed allowable stresses. The driveability analysis is based on the wave equation method, first proposed by Smith (1960). In the absence of specific driveability analyses being carried out, guidance is available in the industry on required pile wall thickness and diameter combination for a given hammer size [API RP2Al. Almost all offshore pile installation projects, however, are now-a-days based on the pile driveability analyses. Pile driving criteria are summarised in Section 14.9.5. Figure 14.16 shows a typical pile driveability analysis plot of blows per foot for the expected penetration depth. 14.4.3 Drilled and Grouted Piles The drilled and grouted steel pile concept has been used successfully in offshore applica- tions. Typically, a hole is drilled to a given depth into the sea floor through the leg of a jacket structure. A pile is then fed through the jacket leg and lowered into the drilled hole. Cement is then pumped down from the top through and around the pile to fill the gap between the pile and the sides of the hole in the seabed. Pumping is continued until the annulus between the jacket leg and the pile is filled with grout cement. In this way, structural continuity and load transfer is achieved from the jacket to the pile through the grout annulus between the pile and the inside wall of the jacket leg.
- 424. Offshovr lnstuiiutiori zyxwvutsrq Figure 14.16 Pile blow count/penetration analysis plot zyx 1075
- 425. 1076 zyxwvutsrqpon Chapter z 14 z Drilling operations should be done carefully to minimise the possibility of hole collapse. Steel pipe casings are used when a hole instability is expected. It is worth noting that the drilled foundations have a distinct advantage over the other types where holes can be drilled through the rock while pile driving may not be considered as an option. zyxwvut 14.4.4 Suction Embedded Anchors Recently, suction embedded anchors have been used to anchor floating exploration and production platforms particularly in a soft cohesive seabed soil. They have been introduced in deepwater applications where alternative foundation concept may prove more costly and most probably require the use of a large derrick barge. The suction piles are made of an open bottommed cylinder with a hole somewhere near the top through which water is pumped out to “suck” the pile into the seabed as shown in Fig. 14.17. The suction anchors have been installed in water depths from as shallow as 40 m to as deep as 2500 m. Diameters ranging from 3.5 to 7 m have been used, with a penetration up to 20 m. Unlike the drag embedment anchors, the location of the suction piles can be determined with great accuracy. This provides a distinct advantage in fields with congested subsea facilities. An added benefit of the use of suction anchors is that they do not need to be Figure 14.17 Suction pile schematic
- 426. Offshore zyxwvutsrqponm Installation zyxwvutsrqp 1077 dragged in order to be proof loaded. The choice of the installation vessel depends on the size of anchor and other operations that are taking place during the same installation campaign. For deep water mooring installations, the suction anchor is often installed at the same time as the mooring line, thus avoiding the need to connect those two components under water. There are also connectors which can be used to connect mooring lines to a pre-installed suction anchor. The suction anchors can be lifted or skidded onto the deck of an anchor handling tug z (AHT) which transports it directly to its offshore location ready for installation. The installation process consists of the following stages: Over-boarding Lowering to the seabed Penetration into the ground. Deploying the pile over-board the installation vessel can be carried out using a crane or an A-frame depending on the size of the pile. Other low cost installation options are also available. Once in water, the pile is lowered to the seabed using the vessel crane or deck-mounted winches. The most critical phase of the lowering process is the “hovering” stage where the suction pile is suspended several meters above the seabed. During this phase, successive heave cycles can cause the pile to partially penetrate and then retract from the seabed. As the pile approaches the seabed, the entrained water escapes below the lower rim and through the hole on top, thus creating a damping force on the pile motions. It is important to ensure at this stage that the damping loads and the seabed resistance to penetration do not cause slackening of the slings leading to subsequent snatch loads. Heave compen- sators fitted to the crane or the winch help make this stage much more controlled. Once the soil resistance to penetration exceeds the self-weight of the pile, the crane wires are slackened. A survey package is normally attached to the suction pile at its top to give verticality and orientation information, but more usually this function is provided by an attending ROV which attaches itself to the anchor. The horizontal positioning of the anchor may be assisted by using pre-measured and installed guide-ropes which are tied back to an existing structure e.g. wellhead frame. Alternatively, a set of small buoys can be pre-installed to mark the target position of the pile. Once the self penetration ceases, the attending ROV, which is equipped with suction pumps connects to the suction valves and pumping of water from the inside of the anchor can commence. The anchor penetrates the soil as a result of the water being pumped out of the hole at the top thus creating an under-pressure that drives the pile into the ground. Water is pumped out at a pre-determined and controlled rate so as not to implode the anchor. The total soil resistance to penetration, RToT is the sum of the resistance from the side friction, Rsideand the resistance from the tip including any stiffeners that may be present, Rtip: zyxwvuts (14.1)
- 427. 1078 zyxwvutsrqpo Chapter z 14 The amount of under-pressure, zyxwvu Au, needed to penetrate the pile into the soil is: (14.2) where, W is the submerged weight of the pile, A is the projected horizontal area inside the pile. The required under-pressure is inversely proportional to the pile projected area, A . Since A is proportional to the square of the pile diameter while is proportional to the diameter, the required suction pressure reduces as the pile diameter increases. If the suction pressure exceeds the soil capacity, the soil fails by upheaval in the soil plug inside the pile. The suction pressure is also limited by the structuralintegrity of the pile. It is important to verify that the soil capacity is not exceeded either during the lowering stage when the pile is accelerating while suspended from the crane hook, or during the penetration stage. Since the required suction pressure is inversely proportional to the pile diameter, piles with larger diameter can achieve higher penetrations before reaching refusal. Refusal is defined by the suction pressure being equal to the limiting soil-failure loads. Care has to be taken into designing of the suction anchor to ensure that the anchor does not rotate during penetration. The provision of the vertical cross walls inside the anchor, in the lower part, can stabilise the anchor during the penetration phase. zyx 14.4.5 Drag Embedded Anchors Drag anchors have been in use by ships for a very long time and have been used in the offshore industry, since its early days for mooring semi-submersible drilling vessels, single point moorings (SPMs) and installation vessels. The drag anchors generate their holding capacity by self-embedding in the seafloor when pulled horizontally mobilising the shear strength of the seabed soil to resist the pulling force. The ultimate holding capacity of the drag embedment anchors is several multiples of its weight, depending on its type and on the soil conditions. Some anchors embed themselves in the soil, irrespective of their orientation on contact with the seabed, for example the Bruce anchor. Other types of anchors, for example the Delta anchor, will only embed if it arrives at the seabed in the correct orientation. The installation of such anchors will involve the use of a second line, in addition to the anchor line, for correct orientation. Anchors are normally installed by anchor handling tugs (AHT). When mooring a vessel the AHT approaches the vessel stern until it is in close proximity to the fairlead. The anchor is handed over to the AHT winch and the AHT heads towards the designated anchor location while the mooring vessel’swinch pays out the mooring line. The anchor is then lowered to the seabed by either a wire attached to a ring chaser or a pendant wire. In the case of a ring chaser, the AHT pulls the ring back along the mooring line by the line attached to the ring
- 428. Offshore zyxwvutsrqponm Installation zyxwvutsrq 1079 z and offers it to the offshore vessel. In the case of a pendant wire, the free end of the wire is attached to a buoy and left in position. Once all anchors are in place, the mooring lines are subjected to tension test loads by pulling diagonally the opposite mooring lines against each other (cross-tensioning). The anchors can also be dragged by the AHT some distance (up to some 200 ft) along the seabed to achieve the required holding capacity. Piggyback anchors may also be added, if additional holding capacity is required. zyxwv 14.5 Subsea Templates Subsea templates are fabricated from steel members; they vary in size and weight depending on their functional requirements. Template weights typically range from a one hundred- tonne skid frame to several hundred tonnes. 14.5.1 Template Installation Subsea templates can be installed using the same mobile offshore rig (MODU) used for well drilling or a heavy lift crane vessel. Figure 14.18 shows a procedure for installing a subsea template by keelhauling it below the MODU. This installation method involves pre-installing the piles through a temporary pile guide frame, keelhauling the template below the rig and lowering the template to the sea floor using the drill pipe. In this type of installation, the weight of the template is restricted z Piisewe Fmmm 1. Install Pile Guide Frame Keel Haul Rigging 3. Keel Haul Manifold 4 4 4. Lower Manifold Figure 14.18 Installation of a subsea template by a MODU [Homer, 19931
- 429. 1080 zyxwvutsrqpon Chapter z 14 z 1. Mob Manlfold and Crane Vessel zyxwvutsrq B zyxwvutsrqp Llfi Manlfold zyxwvut 3 .Add Rlgglng Extensions id 2 Lower Manilold 4. Sei Manifoldon Piles zyxwv Figure 14.19 Installation of a subsea template by a crane vessel [Homer, 19933 to the lift capacity of the rig’s draw-works. The keelhauling phase can be simplified by the use of templates which are buoyant. Buoyancy may be obtained by using steel tubulars in the construction of the template structure. However, as the water depth increases, wall thickness to diameter ratio increases rapidly negating any perceived benefit from using the MODU for installation. Figure 14.19 shows a procedure for installing a subsea template using a heavy lift crane vessel. The heavy lift vessel directly lifts the template off the deck of a transport barge and lowers it into water. The template is then further lowered to the sea floor on the crane hook using the crane’s underwater block. If an underwater block is not available or that its capacity is insufficient, the template can be transferred to a deep-water-lowering winch system. The transfer from the crane hook to the lowering system typically occurs after the template is in water. Particular attention has to be paid to the method of transfer as it imparts additional risk to the overall installation operation. 14.5.2 Positioning and Monitoring Templates are installed within tight tolerances in terms of position, direction and level. The position and the orientation of the template are achieved through the use of the pre-drilled wellheads as guides. The docking piles. installed either before or after the drilling activities, are also used as guides. The template is lowered to some height, typically a few feet, above the seabed, at which point the position and orientation of the template are verified and corrected, if necessary.
- 430. Offshore zyxwvutsrqponm Installatiori zyxwvutsrq 1081 z Inclinometers are mounted on and used to monitor the levelling of the template on the seabed. The inclinometers can be linked to a control room on the installation vessel through an umbilical line. Levelling is achieved by using hydraulic jacks, which act by pushing against the pre-installed piles and are remotely operated. zyx 14.5.3 Rigging Requirements Whether a MODU or a heavy lift is used, as the water depth increases, the lift capacity is diminished by dynamic load margins and by the weight of the lowering string. Subsea templates are typically lifted using a single point lift with four wire slings connected to a single hook at the top and attached to four lift padeyes at the corners of the template structure. Figure 14.20 shows a subsea template ready for installation. The rigging of the template is designed for the three phases of the installation operation: The lifting in air of the dry weight of the template from the transport barge. Dynamic factors apply, which account for the lifting by a floating structure from another floating structure in an offshore environment. The dynamic amplification factors for this phase are typically less than 1.25. The lowering of the template into water through the wave zone. The drag and the inertia due to the direct wave load impart additional loads on the template and the supporting riggings. Slam and slap loads can also be significant. The lowering into water of the template to the seabed. The drag and inertia loads in this phase result from the template’s dynamic motions that are caused by the motions of the installation vessel. The combination of the template mass and hoisting wire stiffness can give rise to natural heave period for the template that are in the same range as the installation wave periods. The resulting resonant response of the template induces Figure 14.20 Subsea template lifting
- 431. 1082 zyxwvutsrqpo Chapter z 14 dynamic tensions that equal, or exceed, static tension in the rigging. zyx A dynamic analysis can be carried out to calculate the motions and line tensions in either time or frequency domains. Cranes or winches with heave compensation are often used in deepwater installation to avoid such resonant response. 14.5.4 Existing Subsea Facilities The design and the installation method of a subsea template should take into account any existing installations on the seabed. Wellheads are generally pre-installed; flow lines may also be installed before the template. Other cluster well systems,jumpers, may exist in close proximity of the intended location of the subsea template. All these factors need to be taken into consideration in the design and installation methodology of the subsea template. 14.5.5 Seabed Preparation The surface of the seabed is rarely horizontal or even. The horizontality of the template is crucial to successful drilling operations. For these and other reasons, piles are installed prior to the installation of subsea templates. The piles can be driven, jetted or drilled and grouted. For foundation types and installation methods, refer to Section 13.4. 14.6 Loadout The phase of transferring the completed structure from the quay onto the deck of a cargo vessel is referred to as the loadout operation. Most loadout operations take one of four forms: a. Trailer loadout where mutliwheel hydraulic trailers are brought underneath the structure, in order to lift it and wheel it onto the deck of a barge which is placed right up against the quay; Skidded loadout where the structure rests on steel rails and winches are used to push or pull the structure onto the deck of a barge which would have to be equipped with skid beams to take the structure onto its final location on the barge; Lifted loadout where the modules are lifted onto the deck of the barge using shore- based cranes or floating crane barges; Float-away loadout where a structure is built in a dry dock facility, such as semi- submersible hulls, TLP hulls, FPSO hulls, etc. Upon completion, the dry dock is flooded, or ballasted down in the case of floating dry docks and the structure which floats under its own buoyancy is towed away by tug boats. b. c. d. The decision on the type of loadout should be made as early as possible in the design process as, it has direct consequence not only on the configuration and size of structural members but also on the economy of the project. 14.6.1 Loadout Methods The choice of a loadout method depends on a multitude of factors such as the geometry of the structure, its weight and the availability of trailers close to the fabrication site.
- 432. Offshore zyxwvutsrqpon Installation zyxwvutsrq 1083 z The experience of the designer and the fabricator also influences the choice of the loadout method. Like any other project phase, pure commercial factors are quite often the reason behind a certain loadout method to be adopted. zyxw 14.6.1.1 zyxwvutsr Trailer Loadout For a trailer loadout, the module is supported on multi-wheel trailers for the movement onto the cargo vessel. The trailers may be self-propelled or may be pushed or pulled onto the vessel. Trailers accommodate uneven ground surfaces and small movements between the barge and the quay. The support configuration over the trailers is likely to be different from the in-place configuration leading to different load path and set of stresses being imposed on the structure. A separate analysis is normally carried out to verify the structural integrity in this mode. Trailers are normally arranged in three hydraulic groups such that the load on each group can be calculated by simple statics. The reactions from the axles in each group is applied as a uniformly distributed load acting upwards against the weight of the structure. They can be regarded as a series of linear springs, if necessary. In this type of loadout, it is important that the loadout barge maintains elevation against the quay within a specified tolerance, which is typically a few inches. An adequate ballasting system with sufficient redundancy is essential for the success of the loadout operation. The ballast system compensates not only for the transfer of load but also for the effects of tide. The global and the local strengths of the loadout barge, in addition to the stability of the system, are important considerations in determining a ballast plan for the loadout operation. A typical trailer loadout is shown in Fig. 14.21. Figure 14.21 Topside module loadout on trailer
- 433. 1084 zyxwvutsrqpo Steel/Teflon zyxwvu Chapter z 14 0.12 zyxw 0.05 zyx 14.6.1.2 Skidded Loadout Stainless Steel/Teflon 0.10 Teflon/Wood 10.14 In a skidded loadout, the structure is pushed (by jacks) or pulled (by winches) onto the cargo vessel. Skidding may also be achieved by utilising skid units which travel together with the skidded structure. The moving is effected by a combination of push/pull hydraulic jacks and clamps. The structure is usually supported on skid shoes that are guided over the skid beams. The force required to move the loadout object along the skid rails depends on the friction between the skid shoes and the skid rails. The initial load required to move the structure from the static, typically the erection location, is referred to as the breakout or the static load. As the structure moves forward, the force required to keep it moving is less than that at breakout. Table 14.2 shows typical values for the static and the dynamic coefficients of friction. Teflon pads are sometimes mounted on top of the skid rails, with grease applied to them in advance of the moving structure to minimise frictional resistance. The speed of the loadout operation is dictated by several factors such as the stroke of the jacks, the number of parts on the pulling winch or the speed of the ballasting system. For planning and design purposes, the capacity of the skidding equipment such as jacks, wires and anchor points should exceed the breakout force described. 0.05 0.06 1 14.6.1.3 Lifted Loadout When considering a lifted loadout, the designer should take into account the lift capacities of available cranes. These may consist of land cranes with lift capacities measured in hundreds of tonnes, or of floating sheer-Ieg cranes with capacities reaching thousands of tonnes. The same rigging arrangement as for the offshore lift can sometimes be used. A visual inspection of the lift points is required upon completion of the lifted loadout. If the rigging arrangement is different from the installation one, a separate loadout lift analysis is required. The stability of the land cranes or the floating sheer leg also requires checking. Table 14.2 Coefficient of friction used in skidded loadout operations 1Contact Surfaces iStatic IDynamic I 1Steel/Steel [ 0.15 10.12 I 1Steel/Waxed Wood (0.10 10.06 I i
- 434. Offshore Insrullntion zyxwvutsrq 1085 z 14.6.1.4 Float-away Loadout zyxwvu The weight of some offshore structures increases to levels where it is not feasible or economical to load them out from the quay directly onto the transport vessel. In this case, these structures are designed to be launched from a dry dock and wet-towed either to their offshore location or to an awaiting transport vessel. Examples of floated-away structures include the hulls of semi-submersibles, Tension Leg Platforms, FPSOs and FSOs. 14.6.2 Constraints The type of constraints that need to be considered depends on the loadout method chosen. For the skidded or trailer loadouts. the following parameters need to be considered: 0 Pump capacity and redundancy. Pressures imposed by trailer wheels on the ground or transport barges are often less than 10 ton,&. Typically, the local quay strength and the barge deck strength are adequate for this level of loading. Skid beams, on the other hand, impose concentrated loads on the quay and the barge deck. The skid beams are often supported on piled foundations on the quay. Where possible they are also aligned with the barge’s longitudinal bulkheads to minimise stress on the barge transverse frames. Transfer of the ballast water, either between the barge tanks, or between the barge and the sea is carried out to compensate for the weight of the structure or the effects of tide. The pumps should have adequate capacity to keep the barge level with the quay within a specified tolerance that depends on the loadout equipment. Pumps should have some redundancy and spare capacity, typically 50%, to allow for individual pump failures. Timing is also another factor where, in order to take advantage of the tidal conditions, a loadout operation may commence at low tide so that unnecessary de-ballasting of the barge is avoided. Water depth and quay height above the water line at the loadout quay represent additional constraints. A minimum underkeel clearance of around one metre shall be maintained at all stages of the operation. In some instances where the water depth is not sufficient, a grounded loadout is considered where the barge sits on a well leveled and prepared seabed. If a barge needs to be grounded for a certain loadout operation, due to limitations on the water depth and quay height, the condition, levelness and bearing capacity of the seabed at the quay are some important considerations. In this case, the ballast plan has to be developed so as to ensure that only a proportion of the barge and cargo weight is resisted by the seabed. This proportion is limited to the bearing capacity of the soil. The constraints associated with the lifted loadouts are similar to those considered in any lifting operation. Mooring large crane vessels in the vicinity of loadout facilities is Quay and barge local and global strength. Barge freeboard in relation to quay height. Tidal range and rate of variation with time.
- 435. Chapter z 14 z 1086 zyxwvutsrqpo sometimes difficult. When land-based cranes are used, the strength of the quay and the load-sharing between the cranes, if more than one is used, are important considerations. Other constraints relate to the weather and include the swell, the current, the wind speed, visibility and the general weather conditions. zyxw 14.6.3 Structural Analysis The loadout procedure provides a detailed description of all the stages of the loadout operation. A representative structural model is normally set up to incorporate the support configuration during the various phases of the operation. For the trailer and lifted loadouts, the support configuration does not change significantly during the loadout and a single analysis would be adequate to model all the loadout stages. Care should be taken to ensure that the hydraulic connectivity of the trailers and the potential for variations in the load sharing between the different trailer groups is understood and accounted for in the analysis. In the case of a structure supported on four skid shoes or more, the structure needs to be analysed and checked for settlement or loss of support due to the barge movement or ballasting inaccuracies. The amount of mis-alignment that needs to be considered depends on the loadout procedure. It is usually difficult, and too restrictive, to keep the mis- alignment below 25 mm. For analysis purposes, a loadout is regarded to be a static condition. If the loadout analysis is carried out on the basis of the working stress design code (e.g.AISC or API RP2A-WSD) no increase in the allowable stresses can be taken into account. 14.7 Transportation 14.7.1 Configuration Structures can either be wet or dry-transported. In a wet transport the structure floats on its own hull and is towed by one, or more tugs to the offshore site. In the case of dry transport, the structure is loaded onto a flat top cargo barge, on a general purpose cargo carrier or on a purpose built submersible ship often referred to as a heavy lift vessel (HLV). Topsides, jackets, piles and subsea units have no or little buoyancy and are normally transported “dry”. Structures such as semi-submersible vessels, gravity base platforms, tension leg platforms, spars and jack-up rigs can be either wet or dry transported. The decision to transport these structures dry or wet depends on: Dimensions, weight and centre of gravity height of the structure: The current cargo weight record on submersible ships is 60,000 ton. Transport route design environment: If the direct environmental loads or motions associated with a wet tow are too onerous on the structure, it needs to be dry transported. Distance and schedule constraints: Heavy lift vessels and general purpose ships are the fastest mode of transportation and are therefore the most common modes for long
- 436. Offshore zyxwvutsrqpon Installation zyxwvutsrq 1087 distance transportations. Typically, heavy lift vessels can achieve calm weather speeds of 12-16 knots while the wet tow speeds are in the range of 4 8 knots. Large structures as the gravity base platforms, spars and large TLPs are towed at speeds of less than 4 knots. Cost: Heavy lift vessels are competitive for long and medium distance transports, while towing is more cost effective for short tows. Ability to avoid bad weather: In areas with tropical revolving storms or generally harsh environment, tows can only be undertaken at certain times of the year. Tows are generally too slow to change course to avoid forecast bad weather or seek shelter. 14.7.2 Barges and Heavy Lift Ships Transport barges vary in size and capacity. Their availability also depends upon the geographical location. There are mainly two types of transport vessels: Towed barges and Self-propelled ships including submersible heavy lift vessels. Cargo Barges These are barges which are towed or pushed by tug boats to transport from one location to another. These, in the majority, are flat top and bottom and are simply equipped with navigational lights, fairleads and towing points. A small proportion of these barges are designed to be submerged so as to pick up floating cargoes. These are equipped with a forecastle and a deck structure at the bow and have their own ballast system. Large steel boxes, stability casings, are added at the stern to provide additional water plane area necessary for the stability of the barge and its cargo as the deck goes through the water line. These stability casings are removable and can be stowed away on the deck of the barge or stored onshore when not required. Towed barges are classified not only by their length and width and also by their mode of utilisation (e.g. Launch barges, submersible barges). The typical barge sizes and their uses are: Barges less than 200 ft in length and 50 ft wide. These are small pontoons used for carrying small structures in sheltered inshore waters. 250 ft zyxwvuts x 70 ft barges. These are relatively small pontoon barges with no ballast systems of their own. They are used to transport small offshore modules, small jacket and piles, tendon sections for TLPs, containers for pile driving hammers, modules of drill rigs, etc. 300 ft barges. These can be 90 or 100 ft wide barges. They represent standard cargo barges used quite extensively in the offshore industry. Most of these barges are not equipped with a ballast system of their own. Medium size structures, in the region of 3000 Te have been transported on barges of this type. 400 ft x 100 ft barges. These barges are often equipped with a ballast system of their own. Due to the deck space available on the barge, more than one structure can be transported onboard these barges. These barges are ideal for transporting piles and
- 437. 1088 zyxwvutsrqponm Chapter z 14 z bridges as they avoid the risk of immersion in water and wave slamming. Some of these barges are used for launching shallow water jackets. Barges of 450 ft and longer have been used for jacket launching. These barges are equipped with ballasting systems in addition to skid beams and rocker arms at the stern to enable the launching of jackets. Heerema’s H851 barge, which is nominally 850 ft long by 200 ft wide, is the largest barge available in the industry. Submersible barges. These are towed barges equipped with stability casings aft and a ship-like bow structure and a bridge, sufficient to enable the submerging of the barge above its main deck. The Boa barges (nominal dimensions 400 ft x 100 ft), the AMT barges (nominal dimensions 470 ft x 120 ft) and the recently built Hyundai barges (nominal dimensions 460 ft x 120 ft) are examples of these submersible barges. These barges can submerge up to 6-8 m above their decks. Vessel owners and operators publish data of their vessels in terms of deadweight which provides a broad indication of their carrying capacity. Additional requirements need to be met in terms of their global strength, local deck and frame strengths and height of the cargo’s centre of gravity. While a cargo barge may be able to transport a 10,000 ton structure with low vertical centre of gravity and supported on a large number of points on the deck, it may only be able to transport a 6000 ton topsides module which has a relatively high centre of gravity and supported on fewer support points. A typical tow arrangement with a towing bridle is shown in Figs. 14.22 and 14.23. Two lines run from tow brackets through fairleads on the barge and connect to a triplate through towing shackles. These two lines are referred to as the towing bridle. A third line connects the triplate to the winch of towing tug. An emergency wire is installed along the length of the barge. The line is attached to a synthetic rope that terminates with a buoy which trails behind the barge during tow and forms part of the towing arrangement. The size of a tug is determined on the force required to hold the tow in a given environment. The Noble Denton Guidelines require the tow to hold station in a Beaufort Force 8 with a corresponding significant wave height H, of 5.0 m, a wind speed of 20 mls (at 10 m above mean sea level) together with a current speed of 1 knot with the barge heading into the wind. The resulting load is multiplied by an efficiency factor, which accounts for the difference in the tug-pulling capacity between calm weather and storm conditions. Further reduction in the efficiency applies when multiple tugs are used. The towline pull required (TPR) is usually calculated by adding the wind, the current and the wave forces. The wind force (in tonnes) is calculated as: F,+= 0.0625 zyxwvuts (ACh C,)V: where, A,, is the projected wind area (in m2) Ch is the height factor (from MODU) C , is the shape factor (from MODU) V,, is the wind speed in m/s. (14.3)
- 438. Offshore Instullatioii zyxwvutsrq 1089 CLOSED b 3R:O.Ezyxwvutsrq APEX-. . . (OELTA zyxwvutsrqp PLATE) . , . WiHt OR CWY RRI3LE - . , , - - zyx CHAFE CHIN .-. ---.RECOVERY WINCH 1 . . Azy Figure 14.22 Typical barge towing equipment [Noble Denton Guidelines00301 Figure 14.23 Typical arrangement of tow line and bridle
- 439. 1090 zyxwvutsrqpo Chapter z 14 z The current force (in tonnes) is calculated as: (14.4) where, A, is the projected current area (in m2) Cd is the drag coefficient which varies between 1.0 for a barge with a flat bow and 0.3 for the spoon bows zyxwvuts V, is the current speed. The wave force (in tonnes) is normally calculated from a diffraction analysis. In the absence of any specific data, wave forces can be conservatively calculated using: F, = pBH:/16 (14.5) where, B is the beam of the barge (in m) H, is the significant wave height (in m) Submersible Heavy Lift Ships These ships often have two propulsion systems that are independent of each other and provide an adequate margin of safety against the ship being completely incapacitated. Some heavy lift ships also have a dynamic positioning system. The ship’s ballast system enables it to submerge its deck, allowing the floating cargo to be floated on or off. The speed of these ships makes them attractive for long haul transportation operations. Their speed also gives a greater ability to avoid forecast storms. This is considered to be a distinct advantage in places and seasons that are prone to severe weather conditions such as tropical revolving storms. Table 14.3 lists the largest self-propelled transport vessels with some of their characteristics. Topside decks as well as semi-submersible and TLP hulls have been carried on the decks of heavy lift ships. For structures that float on their own hulls, the heavy lift ship submerges such that its deck and cribbing clear the keel of the floating cargo by a safe margin of about 3 ft. The cargo is then floated over the ship’s deck and positioned against pre- installed guides by means of wires and winches. The ship is then de-ballasted back to transportation draft. Figure 14.24 shows the dry transport of a semi-submersible vessel with a displacement in excess of 40,000 ton. The topside structures are loaded on the heavy lift ship using one of the conventional methods such as skidding, using trailers or using cranes. zyxwvuts 14.7.3Design Criteria and Meteorological Data Stability and strength are the main aspects of a transportation operation that need to be verified. The following engineering studies are normally undertaken when planning a Next Page
- 440. Offshore zyxwvutsrqponm InsIR//Rtioll zyxwvutsrqpo (m) Swan Class (Tern, Swift and Teal) 1180.5 zyxwv 1091 ( 4 ( 4 (m) (ton) 32.26 13.3 7.3 32,650 ,zyx Table 14.3 List of largest self-propelleddry transport vessels /Mighty Servant 3 1Black Marlin 1Vessel Name Vessel IVessel IDeck Submergence Dead 1 length beam depth depth above ~ 1 maindeck iweight ~ 181.2 40.0 12.0 10.0 27.700 217.8 42.0 13.3 10.1 57,000 lTai An Kou and Kang Sheng Kou 1156.0 132.2 110.0 1 9.0 118,000 1 1Transshelf 1173.0 140.0 112.0 1 9.0 134,000 1 1Mighty Servant 1 1190.0 150.0 112.0 1 14.0 141,000 1 1Blue Marlin i217.8 163.0 113.3 ~ 10.1 178,000 1I Figure 14.24 Dry transport of P-40 on self-propelledvessel mighty servant 1 transportation operation: A route study to evaluate the design environmental criteria. This is normally carried out when a voyage-specific motion analysis has to be carried out. A stability study to demonstrate that the carrier vessel, in the case of a dry transport, or the hull of the transported structure, in the case of a wet tow, meet the requirements of the IMO or the classification society. The analyses are normally carried out using the generic wind speeds of 100 knots for intact stability assessment and 70 knots for z 0 Previous Page
- 441. 1092 zyxwvutsrqp Chapter z 14 damaged stability. Lower wind speeds are sometimes considered on a case-by-case basis for restricted tows in sheltered waters. The stability requirements are covered later in this section. Motions and accelerations study. Typically, motion analyses are carried out with the voyage specific environmental criteria using diffraction or strip theories. In the absence of such meteorological data, deterministic motions are often used. zyx A structural assessment taking into account the loads associated with the motions and accelerations. Seafastening design. A local and global strength assessment of the carrier vessel in the case of a dry transport. The most widely used deterministic motions criteria are those introduced by Noble Denton for flat bottom cargo barges and other types of carrier vessels. The criteria are: 20" roll angle in 10 s period zyxwvu & 0.2 g heave acceleration. 12.5" pitch angle in 10 s period & 0.2 g heave acceleration. When deriving the voyage-specific environmental data for the transportation route, the 10-yr return environment is normally considered. Given the temporary nature of the transportation phase, the data is normally derived specifically for the departure month so as to take advantage of seasonal variations. The transportation route is normally split into several sectors within which the environment is assumed to be uniform as shown in fig. 14.25. The duration of exposure within each of those sectors is calculated based on the Figure 14.25 Typical transport route sectors between Korea and the North Sea
- 442. Offshore zyxwvutsrqpo Installation zyxwvutsrq 1093 z vessel speed. Given that the exposure periods are normally less than 1 month, the environmental data may be reduced to allow for the shorter exposure periods. The monthly wave-height scatter for each sector are normally used to define the Weibull distributions using the method of moments. Additionally, the Fisher-Tippet Type 1 distributions can also be fitted to the wave data. An average month in a IO-yr period will have approximately 2435 periods of 3 hour storms. The probability of non-exceedence associated with the 10-yr return monthly storm is therefore 0.9996. Meteorological data sources include the satellite databases and the voluntary observatory fleet (VOF) data sets. The most comprehensive satellite data set available is a satellite radar altimeter data for which 15 years of data is now available. Each altimeter measures the significant wave height over a 5-10 km footprint every second (corresponding to 7 km steps) giving an accuracy comparable with estimates of wave height from a 20 min buoy record. Synthetic Aperture Radar (SAR) data allows the computation of the directional wave spectra from the satellite-measured data so that all the wave parameters are available for analysis. This type of information has only become available recently and may not be as accurate as the satellite altimeter data for wave height, but nevertheless it provides very useful descriptions of the sea surface. The most comprehensive databases are the CLIOsat database and the ARGOSS internet-based wave climate database. VOF data sets include visual observations of wind speed, wind direction, wave height and direction. wave period and swell height, period and direction, among other parameters, provided voluntarily by ships officers of many nations. zyxw 14.7.4 Transport Route Transportation routes are selected based on the economic. environmental and safety considerations. The following factors are considered: The environmental conditions along the transport route affect the motions of the vessel and the voyage speed. The weather conditions after the commencement of the transport operation often dictate local deviations from the planned route. The existence of safe havens. As part of a contingency planning, particularly for long transports, safe havens have to be identified in case the conditions require the vessel to seek refuge in a port. Vessel or cargo dimensions and hull draft which restrict passage below certain obstructions, such as bridges, or in shallow water or through locks and waterways. Costs of the passage through canals, such as the Suez Canal. 14.7.5 Motions and Stability Motion analyses are carried out to estimate the motions and accelerations of a vessel during transport normally using the frequency domain analysis techniques. These analyses are often carried out in two phases. In the first phase, the motion response to regular waves for a range of wave periods is derived in the form of vessel response amplitude operators (RAOs) for all six degrees of freedom. In the second phase, the response to irregular waves is derived using the significant wave height and wave period.
- 443. 1094 zyxwvutsrqpon Chapter z 14 z The following parameters are needed for the motion analysis: Significant wave height representative of the tow route A range of peak wave periods Wind speed and Vessel heading relative to the waves. The design wave height, H,, can be based on the 10-yr return adjusted for the periods of exposure. The range of peak wave periods, Tp,is used to account for the different wave steepnesses and can be obtained from the following expression: where H, is expressed in metres and T, in seconds. If the peak roll period of the barge falls outside the T, range for the design wave, smaller waves with periods similar to that of the barge roll period are also analysed. In the absence of a motions analysis, the loads can be combined deterministically as follows: The vertical force is given by: The rotational moment of inertia is given by: where, (14.7) (14.8) (14.9) Roll, or pitch, period (in seconds) Roll, or pitch, angle (in radians) Height above the centre of rotation (assumed to be at the waterline) Horizontal distance from the centreline of the barge Gravitational acceleration (mls2) Moment of inertia of the cargo about its longitudinal axis Inertia force parallel to the vessel’s deck Inertia force normal to the vessel’s deck. Intact ‘StaticStability Stability requirements are stipulated by classification societies or, in some cases, marine warranty surveyors. Stability requirements such as the range of positive stability. the required area ratio and the damaged stability scenarios to be considered depend on the
- 444. Ofisshore Insrullution zyxwvutsrq 1095 z shape of the hull. The following are extracts from Noble Denton's stability requirements for ships and barges: "The range of intact stability about any axis shall not be less than 34" for large barges and 40" for small cargo barges [less than 23 m in beam or 74 m in length). Alternati- vely, zyxwvuts i f model tests or motion analyses are carried out, the minimum range of static stability shall not be less than 120+0.8*0)", vvhere zyxw 0 is the maximum amplitude of nzotion plus the static angle of inclination from the design wind. The buoyancy of a watertight cargo may be considered in the computation of the stability characteristics. Any opening giving an angle of dobvn flooding less than zyxw !0+5)' shall be closed and watertight when at sea, )$*here 0=20'for large barges and 25"for small barges. A cargo overhang shall not immerse as a result ofheeling in a 15m/swind in still water conditions. The area under the righting moment curve to second intercept of the righting and wind overturning moment curves or the down flooding angle shall not be less than 40% in excess ofthe area under the wind overturning moment curve to the same limiting angle". 14.7.6 Seafastenings/Tie downs Where possible, the strong points on the cargo, such as the legs of topside decks, are located over the strong points of the carrier vessel, such as the bulkheads. Where this is not possible, the weight of the structure is supported on steel grillages that distribute the static and the dynamic loads into the carrier vessel's strong points. When the dry transport cargo is a plated structure, such as the hulls of mobile offshore drilling units or tension leg platforms, the weight of the cargo is distributed into the deck of the carrier vessel through wood cribbing. Cribbing could be aligned with the cargo's frames to avoid overstressing it or arranged in a random fashion. Where random cribbing is used, the dynamic stresses in the cribbing are normally limited to 1 N/mm2. Otherwise stresses are limited to 4 N/mm2. Other limiting stresses are considered depending on the type of wood used. Seafastenings are structural members that are made of steel members or steel wire that are used to restrain the structures on board a vessel against movement due to the vessel motions. Steel wires lashings are normally used for smaller cargoes that are transported on board cargo ships. For large offshore structures, seafastening can consist of a system of steel tubular members which are welded to the cargo and to the deck of the transport vessel. The design of the grillage and seafastening is usually carried out to the requirements of the AISC and the API RP2A. For the design of the seafastening members, the allowable stress may be increased by a third to reflect the transient and extreme nature of the transport load. The third increase in the allowable stresses does not however apply to the local strength of the deck of the carrier vessel. 14.7.7 Structural Analysis When seafastening members are modelled, the structural analysis is carried out in two phases: Transport dynamic condition. Still water static condition and
- 445. 1096 zyxwvutsrqpon Chapter z 14 z The first phase consists of a static analysis in still water condition,where the full structure is modelled and the seafastening members omitted from the analysis. Only the gravity loads are considered at this stage. In the second phase, the seafastening members are added, or the boundary conditions are modified to reflect their addition. The analysis is carried out against the dynamic loads such as: Vessel motions and accelerations listed above. Standard motions are combined as follows: +Roll +heave &Pitch & heave Deflections of carrier vessel, if they are significant Direct wave loads due to the inundation of cargo overhang, if they are present. Where the voyage-specific motion analyses are carried out, direct wind loads on the cargo are also considered. However, to allow for the coinciding of the low probability of extreme motions and extreme wind loads, the combined loads are reduced by 10%.The design load is then the highest of Motion-induced loads alone Wind-induced load alone Static and dynamic loads are combined to calculate the highest stresses in the cargo and the highest and lowest reactions to the carrier vessel. 0.9 x combined motion and wind loads zyxw 14.7.8 Inundation/Slamming Parts of large cargoes that overhang the transport vessel may be subject to direct wave loading such as slamming, drag and buoyancy. Slamming loads result when the structure makes the first contact with the water surface. These are impulsive loads with durations in the order of several milliseconds. As such, their effect is generally localised to the area of the slamming with little or no loads transmitted globally to the rest of the structure. The drag and the buoyancy loads result from the subsequent submergence of the structure in the water. The duration of these loads is of the same order of magnitude as the vessel motions and wave periods. They may therefore be combined with other dynamic loads. However, inundation loads are often not in phase with the global inertia loads and may act to reduce these loads. It is recommended to carry out several analyses that include and exclude the inundation loads and design for the worst case. Cargo inundation changes the hydrodynamic stiffness and the motion characteristics of the carrier vessel. If significant inundation is expected, its impact on motions should be considered. While simplified methods are available for this purpose, the non-linear behaviour introduced by cargo inundation is best predicted by model testing or time domain analysis. Slamming is dependent on the encounter relative velocity between the structure and the instantaneous water surface. It also depends on the encounter angle and the shape of the
- 446. Offshore zyxwvutsrqpon Insrallation zyxwvutsrqp 1097 structural member and the amount of entrained air in the water. Most research carried out on slamming corresponds to idealised conditions such as the slamming of wedges or flat plates on flat calm water surfaces. This makes their results conservative. The wave slamming forces can be evaluated on the basis that the impact slamming force is equal to the rate of change of momentum of the water, given by equation (14.10): zy (14.10) where m is the mass amount of water and zyxw I/ is the velocity of an equivalent circular cylinder. There are difficulties in estimating the impacting mass of water and the velocity of the equivalent cylinder which varies with time. The slamming force is given below: F = 0.5C,pV; zyxwvut edlg (14.11) where, C , is the slamming coefficient, typically taken as equal to x for a tubular member. Generally this coefficient has to be agreed upon for shapes other than tubular members is the diameter of the cylinder is the length of the cylinder is the encounter slamming velocity between the member and the wave is the density of water. d l V, p zyxwvutsr 14.8 Platform Installation Methods 14.8.1 Heavy Lift This is the most common method of installation of offshore structures. In this method, the structure is lifted off the transportation vessel by a crane vessel and lowered into position. The lifted structure is equipped with lifting lugs and slings that are connected to the lugs and the hook of the crane vessel. Figure 14.26 shows a typical lifting arrangement. The most common method of attachment of slings to the lifted structure is by the use of shackles connected to padeyes that are welded to the structure. Shackles of up to 1000 tonnes Safe Working Load (SWL) size are available. The slings can be alternatively wrapped around trunnions or cast padears that are tailor designed for the lift. Cast padears are normally used in larger lifts. Given that the structure is supported in the lifting mode in a different configuration from the in-place condition, its integrity in the lift condition needs to be verified. Also, the slings are often attached to the structure at an angle to the vertical, thus imparting horizontal loads in the area of the lifting lugs. Additional bracing is normally required to resist these loads. Where this is not possible, the slings are kept vertical by the use of spreader bars or
- 447. 1098 zyxwvutsrqpo Figure 14.26 Lifting arrangement zyxw Chapter z 14 z spreader frames. For statically indeterminate lifting configurations, the structural integrity also has to be verified against any sling length mis-match that cause the redistribution of loads to individual slings. The main constraint associated with the lifted installation is the capacity and the availability of crane vessels. There are only a few crane vessels with a lifting capacity in excess of 5000 ton. Furthermore, the availability and cost of such vessels is normally a major consideration when planning an offshore installation operation. 14.8.2 Launch This method is typically used for installing jackets with weights that exceed the lifting capacity of available cranes. Launching operations are performed over the stern of the launch barge. The launch barge arrives on site with the launch rigging already attached to the jacket and with the jacket overhanging the barge stern. The launch operation starts by trimming the barge typically by about 4-5" by the stern. In order to initiate the sliding of the jacket over the skid beams, the launch winches pull the jacket towards the stern. As the jacket travels towards the stern the barge trim increases and the sliding process is accelerated till the centre of gravity of the jacket passes over the rocker arm hinges. At this point, the jacket starts to rotate and enters the water. The barge accelerates in the opposite direction of the jacket and a complete separation between the two is achieved. This operation normally lasts several minutes.
- 448. Offshore zyxwvutsrqponm In~rallnriori zyxwvutsrq Figure 14.27 Launching of a 4-Leg Jacket zyxw 1099 z The trajectory of the jacket during the launch should be such that the jacket clears the seabed by a sufficient margin. Launch trajectory is predicted by a launch analysis. In the launch analysis, the equations of motion of the barge and the jacket are solved at small discrete intervals during the launch sequence. The jacket-entrance into the water introduces drag and inertia loads onto the jacket members that resist the jacket motion. The launch trajectory is dictated by the relative magnitude of weight and buoyancy of the jacket, the relative positions of their respective centres and by damping introduced through the drag loads on the jacket members. Figure 14.27 shows a launched jacket entering the water. zyxwvut 14.8.3 Mating Also known as “deck mating” and “floatover”, this method is used for a deck installation when the weight of the deck exceeds the available crane capacity. The mating operation is executed using the transporting vessel which may be a flat top cargo barge or a heavy lift ship. The most common deck mating method in offshore environments is the internal floatover where the transportation vessel is maneuvered between the legs of a fixed platform jacket. Deck mating is also used for installation of decks of the semi-submersible vessels over their hulls. The weight of the deck is transferred to the jacket, or floating substructure, solely or largely by ballasting the barge down until contact is made between the jacket and the deck and the load is transferred completely from the barge to the jacket. Figure 14.28 shows an internal floatover operation.
- 449. 1100 zyxwvutsrqp Figure 14.28 Floatover of the Malampaya topsides (Heerema) zyx Chapter z 14 z Deck mating requires the deck to be supported during transportation and installation at locations other than its normal in-place supports. Additional steel trusses are therefore required to transfer the weight of the deck into these temporary supports. The jacket also has to be designed for an internal floatover. The distance between the jacket legs has to be greater than the width of the barge by a sufficient margin to allow safe entrance and withdrawal. In addition to increasing the size and weight of the jacket to accommodate the barge entrance, this has a knock-on effect on the deck design since the deck now has to span over a larger distance between supports in its in-place condition. Furthermore, no jacket braces can be installed in the area of barge operations. The leg structures have to be made stronger to compensate for diagonal bracing. This can be a significant factor in locations subject to seismic loads. While the main mechanism for a load transfer is through changing the draft of the transportation vessel, there are several proprietary systems that speed the load transfer operation or increase its operating envelope. The most common systems use: Sand jacks. In this system, the top of the jacket legs are turned into enclosures that are filled with fine sand. The weight of the deck is first transferred to the sand “jacks” by ballasting the barge down. Once the deck separates from the barge, the sand is allowed to drain from the bottom of its compartments and the deck settles into its final position. This method requires strict quality control of the sand, moisture-content and drain valves. Hydraulic Jacks. The deck weight is transferred to the jacket by a combination of ballasting and hydraulic jacks. Given their ability to change the elevation of the deck rapidly the jacks help shorten the installation period. They also allow the operation to be carried out using shallower barges which helps reduce the impact on the jacket design.
- 450. Offsshure zyxwvutsr InFtallarioii zyxwvutsrq 1101 z External floatovers are less common than internal ones and are more sensitive to environmental loading than internal floatovers. Historically they have been used in the installation of modules over gravity based platforms in the North Sea. In this type of installation, two barges support the extremeties of the deck while the area under the middle section of the deck is kept free to avoid clashing with the platform structure during the floatover operation. The deck structure has to be capable of being supported at the extreme ends. The most critical type of loads are the racking loads that could result from relative motions, particularly the pitch motions, of the two barges. Such loads have historically limited the application of this method to inshore sheltered locations, lakes and fjords. z 14.8.4 Hook-up to Pre-Installed Mooring Lines Moored platforms are installed on station by hooking up the pre-installed mooring lines. This operation requires the use of vessels with sufficient winching or lifting capacity to handle the weight of pre-installed mooring lines. Equipment is also required on board the vessels such as winches or chain jacks. 14.8.5 Heavy Lift Vessels 14.8.5.1 Types The lifting capacities of the floating crane vessels have increased over the years in parallel with the increase in platform sizes. Lifting topsides in larger modules reduces the cost of offshore hook-up and commissioning. The current offshore lifting record stands at 12,000 ton. Heavy lift vessels can be categorised as follows: Semi-submersible crane vessels (SSCVs) with dual cranes such as the two largest lift vessels in the world, Saipem’s S7000 and Heerema’s Thialf. Ship-shaped monohull lift vessels with slewing cranes. Seaway’s Stanislav Yudin is an example of this type of heavy lift vessel. Flat bottom monohull lift vessels with slewing cranes. Saipem’s S3000 is an example of such a vessel. Sheer leg crane barges. These are flat bottom barges with an A-frame type boom that can boom up and down. Often, the position of the boom can be adjusted along tracks on the deck of the barge for given lift configurations. Smit’s Taklift 4 is an example of this type of lift vessels. Some heavy lift vessels have dual lifting and pipelay capabilities and are referred to as derrick lay barges. 14.8.5.2 Lift Capacities Table 14.4 shows a listing of heavy lift crane vessels with capacities exceeding 300 ton. The capacity referenced in the table show the maximum lift capacity of the main block at the minimum radius and with the booms tied back, where applicable.
- 451. Table 14.4 Capacities of heavy lift vessels zyxwvu Vessel name Nan Tian Ma Cairo Mohawk Thor zyxwvutsrqpo ~- -~~ ~ ~ Q4000 ~~ Southern Hercule5 ~- Asian Helping Hand zyxwvutsrq 111 Smit Typhoon Taklift zyxwvutsrqpon 3 Taklift zyxwvutsrqponm 5 Nan Tian Long Illuminator DB Sara Maria Sara Maria Atlantic hori7on _ _ _ _ _ _~ Max Vessel name capacity (ton) 300 DLB 750 318 318 350 DB 3 359 362 Arapaho ~~~~ HD-423 ~~ ~~ Yamato ~~ -~ _ _ _ ~ ~~ Kuroshio 11 _ _ ~ ~ 400 DLB-KPI Teknik Perdana (DLB 332) Ocean Builder ~~~~ 400 400 400 DB 16 450 DB Raeford 465 DB16 476 Comanche Eide lift 2 476 499 Pacific Horizon ~~~~ -~~ -~ Max. capacity (ton) 680 680 Vessel name Semco L-1501 Max. capacity 700 IYamashiro I1626 725 IToltika 11814 725 ~ D L B1601 I1814 726 IHercules I1815 730 ITaklift 1 11900 750 IHuasteco I2032 755 1Kongo 12050 Castoro Otto ______ HD2500 784 Kuroshio 2272 800 (DB30 I2300 800 (Master Mind I2400
- 452. 2500 562 590 599 600 600 600 HD-1000 907 Shawnee 909 Castoro 2 998 Chesapeake 1000 Mnicopcri 30 1000 zyxwv - Roland 1000 Field dcvclopmcnt ship Courageous DB zyxwvutsrqp 1 Seminolc DB General 605 Smit Cyclone 1000 615 Tcknik Padu (DLB 264) I000 620 Nan Tian Long-900 1200 632 DB 27 1260 635 DLB 1000 1290 Nagato DLB Polaris 1300 1500 IL M Balder I500 Stanislav Yudin Taklift 6 Mixteco DLB 801 Cherokee Enak 2800 IMexica I500 INanyang I500 2800 ICrawler I540 3000 S3000 Taklift 8 ILili Bisso I544 3000 DB 17 BOS 355 ~ __ DBlOl 3175 3200 zyx Asian Herculcs I1 OFSI DB-I 3600 Musahi DB 50 4000 Avon Senior Rambiz 4000 Koeigo Uglcn Taklift 4 4000 SLC 5000 4536 4600 ~ Columbia Balder 7200 9000 Hermod Thialf 12,000 ICappy Bisso I635 zyxwvutsr S 7000 14,000 1OFSI DB RAREFORD 1635
- 453. 1104 zyxwvutsrqpo Figure 14.29 Capacity distribution of heavy lift vessels zyxw Chapter z 14 z Figure 14.29 presents the distribution of heavy lift installation vessels in terms of their maximum lifting capacities. The plot is presented in such a way that vessels are grouped in the ranges of their lift capacities. 14.8.5.3 Station Keeping of Heavy Lift Vessels The station keeping of a heavy lift vessel consists of a conventional spread mooring or a dynamic positioning (DP) system. On some vessels, the dynamic positioning system is used to assist the conventional spread mooring system. Spread mooring systems are used to hold the crane vessel on location, as well as moving the vessel slowly during installation operations, between the load pick-up and the load set down locations. The mooring lines are typically made of wire rope and the anchors are conventional drag-embedded anchors. The mooring system has to be designed for typical operating conditions such as the 1-yr return storm and should prudently be checked for the single damaged line case. Anchors are normally pre-loaded to ensure no further anchor slippage takes place during installation. In deepwater installation, the time required for the deployment of the mooring lines becomes impractically long in relation to the duration of the overall installation campaign. Some heavy lift vessels have been fitted with DP systems to meet the requirements of deepwater installation. Vesselsequipped with DP systems can set up on location and depart in short time periods. The dynamically positioned vessels can also change headings quickly to reduce the environmental loads and motions. A DP system consists of a control, a sensor, a thruster and a power system. Its positioning may be accomplished through the use of an acoustic, mechanical, satellite or a radar
- 454. Offshore Instullation zyxwvutsrq 1105 z positioning system. Vessels are given the notation of DP1, DP2 or DP3 depending of the levels of redundancy and segregation of their DP systems. zyxw 14.9 Platform Installation Criteria 14.9.1 Environmental Criteria Defining the limiting environmental criteria is an essential part of the installation planning. The limits set for a particular operation are a function of the duration of the operation, the ability to forecast weather and/or change the course of the operation once it starts. The following summarises the industry’s approach to the environmental criteria: Low environmental limits are set for operations that can be completed within a short period of time and can be carried out in a relatively sheltered environment and under controlled conditions. Most loadout operations, for example can be completed in less than one day and are started upon receipt of a good weather forecast. Typically loadout operations are designed for wind speeds of 20 knots. A lifting operation is another such example. For operations that require several days to complete, such as a jacket lifting and piling operation, the jacket is designed to meet a design storm of 1 yr or greater. Operations that require longer than three days, but less than 30 days to complete, such as most of the towage operations, are designed for the 10-yr return storm. Adjustments for limited exposure to certain sectors of the tow route are possible such that the effective return period for the storm is reduced to somewhere between 1 and 10 yr. Operations that require longer than 30 days, such as mooring of FPSO and TLP at the quayside during hook-up and commissioning work, are designed for storms with return periods that range from 20 to 50 yr. When determining the duration of the operations, it is important to take into consideration the possibility of equipment-failure, slowing down the progress of the operation. While a particular operation may be designed for a storm of a certain return period, the operation does not normally commence in weather of that magnitude. All the installation- related operations should be started in good weather. It is also important to note that the operability of the installation equipment is often the most limiting factor. Lifting operations are often limited to seas of less than 5 ft while the piling operations that require crane assistance are limited to seas of less than 8 ft. Where it is impractical, or uneconomical, to design against a storm of a particular return period, it is possible to plan the operation for a lesser limit provided: A typical example of such a case is the weather routing of transportation operations with sensitive cargoes carried on heavy lift ships with speeds of 12 knots or greater. Regular Good weather forecasting is available on a frequent basis. It is possible to avoid the inclement weather, given sufficient notice.
- 455. 1106 zyxwvutsrqpon Chapter z 14 z forecasts of at least once daily are relayed to the ship. When an inclement weather is forecast, the ship changes course to avoid the weather. zyxw 14.9.2 Heavy Lift zyxwvuts Hook Load The heavy lift operation consists of three distinct systems, the crane vessel, the lifting slings and the structure being lifted. The load experienced by the crane vessel is referred to as the hook load. This consists of the weight of the structure, the weight of the rigging and any dynamic factors caused by the dynamic motions of the crane vessel and or the transportation barge. Weight contingencies are a function of the uncertainty in the weight estimate. For example, a structure at preliminary stage design could have a weight factor of 1.10-1.15 while the factor drops to 1.03 for a structure that has been weighed. Dynamic amplification factors (DAF) can be derived analytically from motion studies of the crane vessel where the boom tip accelerations are calculated using an appropriate hydrodynamic theory. Coupled body dynamic studies can also be used to calculate the DAF where the dynamics behaviour of the transportation barge is significant or in deep water installations where the lifted structure is under water. Standard DAFs are often used in lieu of such studies. These are covered by the guidelines of the marine warranty surveyors (MWS) or the classification societies. Typically the DAFs are larger for the smaller structures and for offshore lifts, as shown in Table 14.5. Additional load factors are applied for structures lifted by two or more cranes. These account for the inaccuracies in the location of the centre of gravity (COG), the tilt of the lifted structure and the relative movement of the crane vessels. The hook load should not exceed the capacity of the crane at the installation radius as given by the crane chart. Where the crane chart includes an allowance for the dynamics, the hook load need not include a dynamic amplification factor unless the dynamic response of the lift is expected to be excessive. For subsea installation work in deepwater, the DAF may be in excess of the values listed in table 14.5 and needs to be calculated explicitly. In order to carry out a dynamic lift analysis for deepwater installation, the following information is required: The equation of the dynamic equilibrium is: Response amplitude operators of the installation vessel. Mass and stiffness of crane hoisting wires. Installation wave height and range of periods. Mass of the lifted structure, added mass and the drag coefficient. (14.12) where, Z(t) is the heave of the lifted structure
- 456. Ofshore InsiaNarion zyxwvutsrq Table 14.5 Dynamic amplificationfactors zyxw 1107 z Weight (ton) Offshore Inshore -~ ~ 1100-1000 11.20 11000-2500 11.15 ~ 2500 11.10 11.05 1 Z,(t] is the vessel heave. M Mu niWire C K is the structure dry weight. is the structure added mass. is the dry weight of the hoisting wires. is the vertical drag coefficient of the structure. is the stiffness of the hoisting wires. The dynamic amplification factor due to the heave motions is given by: (14.13) K [((M+Mu +( ~ i r e / 2 . 7 1 8 ) ) ~ ~ - K)2+(C~)2]o'5 DAF = where w is the natural frequency of the spring (rigging) and the mass (structure) system. Load in Rigging For lifts that are statically indeterminate. such as a deck lifted by four slings connected to a single hook, the load distribution of the individual sling is affected by variations in the lengths of the slings as well as the relative distance of their attachment points from the centre of gravity (COG). This mis-match causes a skew load which is accounted for using one of the several methods available in the industry: Assume that one diagonally opposite pair of slings carries 75% of the weight, while the other carries 25%. A similar approach is to apply a skew load factor equal to 1.25 to the worst loaded sling. For a symmetrical lifting arrangement with a COG located in the centre, this approach is equivalent to a 62.5-37.5% distribution, which is less onerous than the first one [Noble Denton Guidelines 00271. Provided the mis-match does not exceed 1.5 in. or 0.25% of the length of any one sling or 3 in. or 0.5% difference between the shortest and the longest slings, it can be ignored as recommended by the API RP2A. Mis-matches of greater magnitude need to be accounted for analytically. The skew load factor can be calculated analytically as recommended by the Det Norske Veritas Rules for Marine Operations (1996). For a 4-point lift with double symmetrical sling arrangement, the skew load factor, SKF, is given by equation (14.14). (14.14) E O SKF= 1 +- zyxw E
- 457. 1108 zyxwvutsrqponm Chapter z 14 z where, E, E ~ , is the average strain in the sling at a load 30% greater than the dynamic hook load. is the strain associated with the total sling and the padeye fabrication tolerance. Shackles are normally designed to a factor of safety of 4.0, if the loading includes dynamics and 5, if it does not. The factor of safety is defined as the ratio of the minimum break load (MBL) to the safe working load (SWL). Slings are typically designed to a minimum factor of safety of 3 which applies to the hand-spliced terminations. If the ratio of the pin (or trunnion) diameter to the sling diameter is less than 4, sling bending efficiency may become critical and a higher safety factor would apply. The required sling minimum break load is given by equation (14.15) [Noble Denton Guidelines 00271. Fsling x 2.25 rl MBL = where, (14.15) Fsllng: q, is the dynamic force in the sling. is the lesser of the sling termination factor, STF, or the bending efficiency factor, BEF. STF = 0.75 for hand-spliced slings. BEF = is given by equation (14.16). 1.0 for resin sockets or swage fittings. BEF = 1 - 0.5,/- +sling +pin (14.16) where, c$pln and Load in Lifted Structuve The global integrity of the lifted structure has to be assessed against the following loads: Given their criticality, the structural members that frame into the lifting points are assessed against higher loads using a “consequence factor”. The API RP2A lumps the three factors listed above into a single factor of 1.35 while for the members framing into the lift point, the factor is increased to 2.0. A typical lift analysis carried out to the requirements of the API RP2A would consist of three load cases as follows: are the diameters of the pin and sling respectively. Self-weight including any weight contingencies. Dynamic loads typically applied as a DAF. Skew loads due to sling mis-match. Static load cases to calibrate the model weight against the weight report. Static load case x 2.0 to investigate the integrity of members framing into the lift point. A factor of 1.5 can be used instead for lifts in sheltered conditions.
- 458. Offshore zyxwvutsrqponm Insrallarion zyxwvutsrq 1109 Static load case x 1.35 to investigate the integrity of other members. A factor of 1.15 can be used instead for lifts in sheltered conditions. The approach described earlier for the development of the skew load in rigging components can also be used. In this case, the effects of dynamics and weight contingencies are applied as a load factor while the skew load effects (if the lift is statically indeterminate) are accounted for by forcing the diagonally opposite slings to carry 75% (or 67.5%) of the load. Loads in Lifting Points Lifting points are normally designed using a load factor of 2.0 on the statically resolved rigging load [API RP2Al. To allow for the effects of uncertainty in the alignment of the rigging, a load equal to 5% of the lifting point design load is applied so as to cause bending in the weak axis of the lifting attachment. The load factor of 2.0 is consistent with the API RP2A’s load factor for members framing into the lift point. Where the skew load factors are used explicitly, such as in the 67.5-32.5% distribution, an additional consequence factor of 1.35 is applied to the load in the lift point. The lift point load factor is therefore built up from: Lift point load factor = 1.1 (weight contingency) x 1.1(DAF) x 1.25 (skew load factor) x 1.35 (consequence factor) = 2.05. This is close to the load factor stipulated by the API RP2A of 2.0. Structural Design Requirements In assessing the global integrity of the lifted structure and the strength of lifting points, the steel components have to meet the requirements of the API RP2A for tubular members or the American Society for Steel Construction, the Allowable Stress Design (AISC-ASD) for non-tubular members. The 1.33 increase on allowable stress is not permitted in the design and analysis of steel for lifting operations. The load factors listed above are consistent with the working stress design (WSD). Lifting operations can also be designed to Load Resistance Factor Design (LRFD) codes such as the API RP2A LRFD where different load factors are stipulated. In the design of lifting points, it is normally preferable to rely on load transmission to the primary steel through shear rather than tension. Full penetration welds and plates made of steel with through thickness properties (Z-quality) are also preferred. Pin holes in the padeyes are normally line bored after the cheek plates are installed in order to ensure an even bearing surface against the pin. Lifting Point Inspection and Re-Use A suitable scope of non-destructive testing is normally specified including an ultrasonic testing (UT) of full penetration welds. Where attachments are used for more than one lift, the critical welds and the inside of the pin hole should be inspected for cracks using a suitable technique, such as magnetic particle inspection (MPI). A visual inspection for mechanical damage should also be carried out.