SlideShare a Scribd company logo
2
Most read
4
Most read
5
Most read
Page 1 of 5
Flash Steam and Steam Condensates in Return Lines
Jayanthi Vijay Sarathy, M.E, CEng, MIChemE, Chartered Chemical Engineer, IChemE, UK
In power plants, boiler feed water is
subjected to heat thereby producing steam
which acts as a motive force for a steam
turbine. The steam upon doing work loses
energy to form condensate and is
recycled/returned back to reduce the
required make up boiler feed water (BFW).
Recycling steam condensate poses its own
challenges. Flash Steam is defined as steam
generated from steam condensate due to a
drop in pressure. When high pressure and
temperature condensate passes through
process elements such as steam traps or
pressure reducing valves to lose pressure, the
condensate flashes to form steam. Greater the
drop in pressure, greater is the flash steam
generated. This results in a two phase flow in
the condensate return lines
General Notes
1. To size condensate return lines, the
primary input data required to be
estimated is A. Fraction of Flash Steam and
condensate, B. Flow Rates of Flash Steam
& condensate, C. Specific volume of flash
steam & condensates, D. Velocity limits
across the condensate return lines.
2. Sizing condensate return lines also require
lower velocity limits for wet steam since
liquid droplets at higher velocities cause
internal erosion in pipes and excessive
piping vibration. A rule of thumb, for
saturated wet steam is 25 – 40 m/s for
short lines of the order of a few tens of
metres and 15 - 20 m/s for longer lines of
the order of a few hundred metres.
3. Condensate return lines work on the
principle of gravity draining. To effectuate
this, drain lines are to be sloped downward
at a ratio of atleast 1:100.
4. Proper sizing of stem condensate return
lines requires consideration of all
operating scenarios, chiefly start up,
shutdown and during normal running
conditions. During plant start up, steam is
not generated instantly. As a result, the
condensate lines would be filled with
liquids which gradually turn two-phase
until reaching normal running conditions.
During shutdown conditions, with time,
flash steam in the lines condense leaving
behind condensates due to natural cooling.
5. Condensate return line design must also
consider the effects of water hammering.
When multiple steam return lines are
connected to a header pipe that is routed
to a flash drum, flash steam in the presence
of cooler liquid from other streams would
condense rapidly to cause a water hammer.
Fraction of Flash Steam
Taking an example case, condensate flows
across a control valve from an upstream
pressure of 5 bara to 2 bara downstream. The
saturation temperature at 5 bara is 151.84 0C
& 120.20C at 2 bara. The specific volume of
water at 5 bara is 0.001093 m3/kg & 0.00106
m3/kg at 2 bara. The latent heat of saturated
steam upon reaching 2 bara is 2201.56 kJ/kg.
The % flash steam generated is estimated as,
ℎ𝑓,1 = ℎ𝑓,2 + [
% 𝐹𝑙𝑎𝑠ℎ
100
× ℎ𝑓𝑔] (1)
Where,
hf,1 = Upstream specific enthalpy [kJ/kg]
hf,2 = Downstream specific enthalpy [kJ/kg]
hf,g = Latent Heat of Saturated Steam [kJ/kg]
The upstream specific enthalpy, hf1 of
saturated water at 5 bara is 640.185 kJ/kg
and hf2 of 504.684 kJ/kg at 2 bara. The steam
specific volume at 2 bara is 0.8858 m3/kg.
Page 2 of 5
The fraction of flash steam is calculated as,
% 𝐹𝑙𝑎𝑠ℎ =
[640.185−504.684]
2201.56
× 100 = 6.15% (2)
Therefore the condensate fraction is,
% 𝐶𝑜𝑛𝑑 = 100 − 6.15 = 93.85% (3)
The steam volume is calculated as,
𝑉𝑆𝑡𝑒𝑎𝑚 = 0.8858 × 0.0615 = 0.05448
𝑚3
𝑘𝑔
(4)
The condensate volume is calculated as,
𝑉𝐶𝑜𝑛𝑑 = 0.00106 × 0.9385 = 0.000995
𝑚3
𝑘𝑔
(5)
Condensate Return Pipe Sizing
To size the condensate return line, the bulk
properties and mixture properties can be
used to estimate the pipe size. It must be
remembered that as the two-phase mixture
travels through the pipe, there is a pressure
profile that causes the flash % to change
along the pipe length. Additionally due to the
pipe inclination, a certain amount of static
head is added to the total pressure drop.
To estimate the pipe pressure drop across the
pipe length, a homogenous model for
modelling the two phase pressure drop can
be adopted. The homogenous mixture acts as
a pseudo-fluid, that obeys conventional
design based on single phase fluids
characterized by the fluid’s average
properties.
The mixture properties can be estimated as,
𝜌ℎ = 𝜌 𝐿[1 − 𝜀ℎ] + 𝜌 𝑣 𝜀ℎ (6)
Where,
L = Condensate Density [kg/m3]
v = Steam Density [kg/m3]
h = Homogenous void fraction for a given
steam quality [x] [-]
The homogenous void fraction [h] for a given
steam quality [x] can be estimated as,
𝜀ℎ =
1
1+[
𝑢 𝑣
𝑢 𝐿
×
1−𝑥
𝑥
×
𝜌 𝑣
𝜌 𝐿
]
(7)
The dynamic viscosity for calculating the
Reynolds number can be chosen as the
viscosity of the liquid phase or a quality
averaged viscosity, µh.
𝜇ℎ = 𝑥𝜇 𝑣 + [1 − 𝑥]𝜇 𝐿 (8)
The homogenous model for gravitational
pressure drop is applicable for large drop in
pressures and mass velocities < 2000
kg/m2.s, such that sufficient turbulence exists
to cause both phases to mix properly and
ensure the slip ratio (uv/uL) between the
vapour and liquid phase is ~1.0. For more
precise estimates capturing slip ratios and
varying void fraction, correlations such as
Friedal (1979), Chisholm (1973) or Muller-
Steinhagen & Heck (1986) can be used.
The total pressure drop is the sum of the
static head, frictional pressure drop &
pressure drop due to momentum pressure
gradient.
∆𝑃𝑇 = ∆𝑃𝑠𝑡𝑎𝑡𝑖𝑐 + ∆𝑃𝑚𝑜𝑚 + ∆𝑃𝑓𝑟𝑖𝑐 (9)
The Static Head [Pstatic] is computed as,
∆𝑃𝑠𝑡𝑎𝑡𝑖𝑐[𝑏𝑎𝑟] =
𝐻×𝜌ℎ× 𝑔×𝑆𝑖𝑛𝜃
105
(10)
Where,
H = Pipe Elevation [m]
 = Pipe inclination w.r.t horizontal [degrees]
The pressure drop due to momentum
pressure gradient [Pmom] is,
𝑑𝑃
𝑑𝑍
=
𝑑( 𝑚
𝜌ℎ⁄ )
𝑑𝑍
(11)
If the vapour fraction remains constant across
the piping, the pressure drop due to
momentum pressure gradient is negligible.
The frictional pressure drop is calculated as,
∆𝑃𝑓 =
𝑓×𝐿×𝜌ℎ×𝑉2
2𝐷
(12)
Where, P = Pressure drop [bar]
f =Darcy Friction Factor [-]
Page 3 of 5
L = Pipe Length [m]
h = Mixture Density [kg/m3]
V = Bulk fluid Velocity [m/s]
D = Pipe Inner Diameter, ID [m]
Re =
DVρh
µh
(13)
Where, µh = Dynamic Viscosity [kg.m/s]
h = Homogenous Density [kg/m3]
The Darcy Friction Factor [f] depends on the
Reynolds number follows the following
criteria,
If Re <= 2100 ; Hagen Poiseuille’s Equation
If Re <= 4000 ; Churchill Equation
If Re > 4000 ; Colebrook Equation
The Laminar Flow equation also referred to
as the Hagen Poiseuille’s equation is,
f =
64
Re
(14)
The Churchill equation combines both the
expressions for friction factor in both laminar
& turbulent flow regimes. It is accurate to
within the error of the data used to construct
the Moody diagram. This model also provides
an estimate for the intermediate (transition)
region; however this should be used with
caution.
The Churchill equation shows very good
agreement with the Darcy equation for
laminar flow, accuracy through the
transitional flow regime is unknown & in the
turbulent regime a difference of around 0.5-
2% is observed between the Churchill
equation and the Colebrook equation. For
Reynolds number up to ~4000,
f = 8 [(
8
Re
)
12
+
1
(A+B)1.5
]
1
12⁄
(15)
A = [2.457ln (
1
(
7
Re
)
0.9
+0.27
ε
D
)]
16
(16)
B = [(
37,530
Re
)]
16
(17)
The Colebrook equation was developed
taking into account experimental results for
the flow through both smooth and rough pipe.
It is valid only in the turbulent regime for
fluid filled pipes. Due to the implicit nature of
this equation it must be solved iteratively. A
result of suitable accuracy for almost all
industrial applications will be achieved in less
than 10 iterations. For Reynolds number up
greater than ~4000,
1
√f
= −2 log10 [
ε DH⁄
3.7
+
2.51
Re√f
] (18)
Homogenous Property Calculations
The two phase mixture flows through the
condensate return line. The associated
density and viscosity of flash steam and
condensate at 2 bara and 120.20C is,
𝜌 𝑣 =
1
0.8858
= 1.129
𝑘𝑔
𝑚3
(19)
𝜌 𝐿 =
1
0.00106
= 943.4
𝑘𝑔
𝑚3
(20)
𝜇 𝑣 = 0.000229
𝑘𝑔
𝑚.𝑠
(21)
𝜇 𝐿 = 0.0000128
𝑘𝑔
𝑚.𝑠
(22)
The homogenous void fraction [h] for a slip
ratio (uv/uL) of 1.0, i.e., uv = uL, and a steam
quality [x] of 6.15% is,
𝜀ℎ =
1
1+[1×
1−0.0615
0.0615
×
1.129
943.4
]
= 0.9821 (23)
The two phase homogeneous density is,
𝜌ℎ = 943.4 × [1 − 0.9821] + [1.129 × 0.9821] (24)
𝜌ℎ = 18.01
𝑘𝑔
𝑚3
(25)
The two phase homogeneous viscosity is,
𝜇ℎ =
0.0615×1.28
10−5
+
[1−0.9821]×2.29
10−4
(26)
𝜇ℎ = 0.000216
𝑘𝑔
𝑚.𝑠
(27)
Pressure Drop Calculations
The return condensate line from the control
valve discharge is sloped at a ratio of 1:100
for gravity drain. The layout of the return
condensate line is,
Page 4 of 5
Figure 1. Condensate Return Line to Receiver
The condensate receiver operates at 1.1 bara
pressure. The mechanical details of the piping
for a flow rate of 1,000 kg/h, pipe size of 1.5”,
100m length & pipe roughness of 45.2 m is,
Table 1. Condensate Return Line Details
Parameter Value Unit
Mass Flow rate [m] 1000.0 kg/h
Volumetric Flow [Q] 55.51 m3/h
Pipe Length [L] 100 m
Pipe Roughness [ε] 45.2 μm
Pipe Outer Diameter [OD] 48.3 mm
Pipe SMYS [Carbon Steel] 30,000 psi
Pipe Design Pressure [DP] 7 bara
Pipe Wall Thickness [WT] 0.08 mm
Corrosion Allowance [CA] 1.0 mm
Calculated WT 1.08 mm
Selected WT 3.68 mm
Pipe Inner Diameter [ID] 40.94 mm
The pipe wall thickness chosen is based on
ASME/ANSI B36.10M and is calculated based
on the hoop stress created by internal
pressure in a thin wall cylindrical vessel as,
WT =
DP×OD
2×SMYS
=
[7×14.5]×[
48.3
25.4
]
2×30,000
× 25.4 (28)
WT = 0.08𝑚𝑚 (29)
Adding CA of 1 mm, the WT becomes 1.08
mm. Based on ASME/ANSI B36.10M, the
selected WT is 3.68mm. The inner diameter
calculated for the selected WT is 40.94 mm.
The condensate return line mixture fluid
velocity is calculated as,
V =
Q
A
=
4×[
1000
18.01
]×
1
3,600
π×[0.04094]2
= 11.7 𝑚/𝑠 (30)
The Reynolds number is estimated as,
Re =
ID×V×𝜌ℎ
𝜇ℎ
=
0.04094×11.7×18.01
0.000216
(31)
Re = 39,971 (32)
Since the Reynolds number is much higher
than 4,000, the flow is fully turbulent and the
friction factor is calculated based on
Colebrook equation. The friction factor is
estimated as,
𝑓 = 𝑓𝐶𝑜𝑙𝑒𝑏𝑟𝑜𝑜𝑘 = 0.0251 (33)
The frictional pressure drop is now calculated
using the Darcy-Weisbach expression as,
∆𝑃𝑓 =
0.0251×100×18.01×11.72
2×0.04094×105
(34)
∆𝑃𝑓 = 0.757 𝑏𝑎𝑟 (35)
The slope angle is calculated as,
𝜃 = [𝑇𝑎𝑛−1
(
1
100
)] ×
180
𝜋
= 0.6° (36)
The static pressure drop [Pstatic] becomes
∆𝑃𝑠𝑡𝑎𝑡𝑖𝑐 =
18.01×9.81×[(1+5)×𝑠𝑖𝑛(0.6°)]
105
(37)
∆𝑃𝑠𝑡𝑎𝑡𝑖𝑐 = 0.000106 𝑏𝑎𝑟 (38)
Therefore the total P with negligible P due
to momentum pressure gradient [Pmom].
∆𝑃𝑡𝑜𝑡𝑎𝑙 = ∆𝑃𝑠𝑡𝑎𝑡𝑖𝑐 + ∆𝑃𝑓 (39)
∆𝑃𝑡𝑜𝑡𝑎𝑙 = 0.757 + 0.000106 = 0.757 𝑏𝑎𝑟 (40)
The condensate exit pressure is 2 – 0.757 =
1.243 bara which is higher than the receiver’s
operating pressure of 1.1 bara.
References
1. “Engineering Data Book III”, Ch 13, Two
Phase Pressure Drop, Wolverine Tube, Inc.
2. “Steam Handbook”, Dr. Ian Roberts, Philip
Stoor, Michael Carr, Dr. Rainer Hocker,
Oliver Seifert, Endress+Hauser
Page 5 of 5
Appendix A

More Related Content

PDF
HTRI PRESENTATION.pdf
PDF
Design of thermosyphon reboiler
DOCX
1 reboilers types
PPTX
Parts of shell and tube heat exchanger
PDF
Tutorial on Aspen Hysys Dynamics - Separator level controller
PDF
AIR COOLED HEAT EXCHANGER
PDF
Kettle reboilers
PDF
Selection and Design of Condensers
HTRI PRESENTATION.pdf
Design of thermosyphon reboiler
1 reboilers types
Parts of shell and tube heat exchanger
Tutorial on Aspen Hysys Dynamics - Separator level controller
AIR COOLED HEAT EXCHANGER
Kettle reboilers
Selection and Design of Condensers

What's hot (20)

PDF
Selection of Reboilers for Distillation Columns
PDF
Flow assurance using AspenHYSYS Hydraulics
PPTX
Selection of Shell and tube heat exchanger (STHE)
PDF
METHODOLOGY FOR SLUG CATCHER SIZING
PPTX
Heat exchanger: Shell And Tube Heat Exchanger
PDF
Boiler calculations
PPTX
PDF
Distillation Towers (Columns) presentation on Types, governing Equations and ...
PPTX
Shell and tube heat exchanger design comparison
PDF
Gland condensers
PDF
Theory and Operation - Secondary Reformers -
PPTX
Single phase flow line sizing
PPTX
Pressure Relief valve sizing and design
PDF
Amine sweetening unit operation
PDF
Hysys simulation
PDF
CHEMICAL PROCESS EQUIPMENT.pdf
PDF
Control of Continuous Distillation Columns
PDF
Thermosyphon Reboiler & its type with operational parameter.
PPTX
Calculation Of Pump Head
PDF
psv specifications
Selection of Reboilers for Distillation Columns
Flow assurance using AspenHYSYS Hydraulics
Selection of Shell and tube heat exchanger (STHE)
METHODOLOGY FOR SLUG CATCHER SIZING
Heat exchanger: Shell And Tube Heat Exchanger
Boiler calculations
Distillation Towers (Columns) presentation on Types, governing Equations and ...
Shell and tube heat exchanger design comparison
Gland condensers
Theory and Operation - Secondary Reformers -
Single phase flow line sizing
Pressure Relief valve sizing and design
Amine sweetening unit operation
Hysys simulation
CHEMICAL PROCESS EQUIPMENT.pdf
Control of Continuous Distillation Columns
Thermosyphon Reboiler & its type with operational parameter.
Calculation Of Pump Head
psv specifications
Ad

Similar to Flash Steam and Steam Condensates in Return Lines (20)

PDF
Feb. 2016 webinar_condensate_return_piping
PDF
Condensate Collection By Dk Singhal
PDF
Condensate Collection By D K Singhal
PPTX
project ppt
PDF
Preliminary pipe size
DOCX
Methods to determine pressure drop in an evaporator or a condenser
PPT
Design of condenser
PDF
EXPERIMENTAL INVESTIGATION OF FLOW CONDENSATION HEAT TRANSFER IN RECTANGULAR ...
PDF
Designofcondenser 130801223803-phpapp02
PDF
Designofcondenser 130801223803-phpapp02
PPT
LINE SIZING PROCEDURE IN PROCESS INDUSTRY
PDF
Heat transfer enhancement in super heater tube using
PPT
1. BOILERS AND STEAM SYSTEMS ..BOE.ppt
PDF
Experimental investigation of flow condensation heat transfer in rectangular ...
PDF
Diesel Production: Equipments Design
PDF
Experimental investigation of flow condensation in 'v' shaped minichannel
PDF
D013161520
DOCX
Senior Design Final Report (2)
PPTX
Pinch analysis technique to optimize heat exchanger
PPTX
design and analysis of heat exchanger
Feb. 2016 webinar_condensate_return_piping
Condensate Collection By Dk Singhal
Condensate Collection By D K Singhal
project ppt
Preliminary pipe size
Methods to determine pressure drop in an evaporator or a condenser
Design of condenser
EXPERIMENTAL INVESTIGATION OF FLOW CONDENSATION HEAT TRANSFER IN RECTANGULAR ...
Designofcondenser 130801223803-phpapp02
Designofcondenser 130801223803-phpapp02
LINE SIZING PROCEDURE IN PROCESS INDUSTRY
Heat transfer enhancement in super heater tube using
1. BOILERS AND STEAM SYSTEMS ..BOE.ppt
Experimental investigation of flow condensation heat transfer in rectangular ...
Diesel Production: Equipments Design
Experimental investigation of flow condensation in 'v' shaped minichannel
D013161520
Senior Design Final Report (2)
Pinch analysis technique to optimize heat exchanger
design and analysis of heat exchanger
Ad

More from Vijay Sarathy (20)

PDF
Exploring LPG Cylinders for Medical Oxygen - A Preliminary Study
PDF
Heating Value Estimation for Natural Gas Applications
PDF
Empirical Approach to Hydrate Formation in Natural Gas Pipelines
PDF
Evaporation Pond Process Design in Oil & Gas Industry
PDF
Key Thermo-Physical Properties of Light Crude Oils
PDF
Cooling Tower Makeup Water Estimation
PDF
Design Considerations for Antisurge Valve Sizing
PDF
Gas Compression Stages – Process Design & Optimization
PDF
Gas Condensate Separation Stages – Design & Optimization
PDF
ECONOMIC INSULATION FOR INDUSTRIAL PIPING
PDF
Evaluating Pipeline Operational Integrity - Sand Production
PDF
Natural Gas Pipeline Transmission Cost & Economics
PDF
Predicting Performance Curves of Centrifugal Pumps in the Absence of OEM Data
PDF
Load Sharing for Parallel Operation of Gas Compressors
PDF
Basic Unit Conversions for Turbomachinery Calculations
PDF
Single Phase Liquid Vessel Sizing for HYSYS Dynamics
DOCX
Key Process Considerations for Pipeline Design Basis
PDF
Chemical Process Calculations – Short Tutorial
PDF
Process Design for Natural Gas Transmission
PDF
OPERATING ENVELOPES FOR CENTRIFUGAL PUMPS
Exploring LPG Cylinders for Medical Oxygen - A Preliminary Study
Heating Value Estimation for Natural Gas Applications
Empirical Approach to Hydrate Formation in Natural Gas Pipelines
Evaporation Pond Process Design in Oil & Gas Industry
Key Thermo-Physical Properties of Light Crude Oils
Cooling Tower Makeup Water Estimation
Design Considerations for Antisurge Valve Sizing
Gas Compression Stages – Process Design & Optimization
Gas Condensate Separation Stages – Design & Optimization
ECONOMIC INSULATION FOR INDUSTRIAL PIPING
Evaluating Pipeline Operational Integrity - Sand Production
Natural Gas Pipeline Transmission Cost & Economics
Predicting Performance Curves of Centrifugal Pumps in the Absence of OEM Data
Load Sharing for Parallel Operation of Gas Compressors
Basic Unit Conversions for Turbomachinery Calculations
Single Phase Liquid Vessel Sizing for HYSYS Dynamics
Key Process Considerations for Pipeline Design Basis
Chemical Process Calculations – Short Tutorial
Process Design for Natural Gas Transmission
OPERATING ENVELOPES FOR CENTRIFUGAL PUMPS

Recently uploaded (20)

PPTX
CARTOGRAPHY AND GEOINFORMATION VISUALIZATION chapter1 NPTE (2).pptx
PDF
Automation-in-Manufacturing-Chapter-Introduction.pdf
PDF
Mohammad Mahdi Farshadian CV - Prospective PhD Student 2026
PDF
keyrequirementskkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkk
PDF
PPT on Performance Review to get promotions
PPTX
Construction Project Organization Group 2.pptx
PDF
BIO-INSPIRED HORMONAL MODULATION AND ADAPTIVE ORCHESTRATION IN S-AI-GPT
PPTX
UNIT 4 Total Quality Management .pptx
PDF
The CXO Playbook 2025 – Future-Ready Strategies for C-Suite Leaders Cerebrai...
PPTX
Infosys Presentation by1.Riyan Bagwan 2.Samadhan Naiknavare 3.Gaurav Shinde 4...
PPTX
M Tech Sem 1 Civil Engineering Environmental Sciences.pptx
PDF
Model Code of Practice - Construction Work - 21102022 .pdf
PPTX
Current and future trends in Computer Vision.pptx
PPTX
CYBER-CRIMES AND SECURITY A guide to understanding
PPTX
Internet of Things (IOT) - A guide to understanding
DOCX
ASol_English-Language-Literature-Set-1-27-02-2023-converted.docx
PPTX
FINAL REVIEW FOR COPD DIANOSIS FOR PULMONARY DISEASE.pptx
PDF
Enhancing Cyber Defense Against Zero-Day Attacks using Ensemble Neural Networks
DOCX
573137875-Attendance-Management-System-original
PDF
TFEC-4-2020-Design-Guide-for-Timber-Roof-Trusses.pdf
CARTOGRAPHY AND GEOINFORMATION VISUALIZATION chapter1 NPTE (2).pptx
Automation-in-Manufacturing-Chapter-Introduction.pdf
Mohammad Mahdi Farshadian CV - Prospective PhD Student 2026
keyrequirementskkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkk
PPT on Performance Review to get promotions
Construction Project Organization Group 2.pptx
BIO-INSPIRED HORMONAL MODULATION AND ADAPTIVE ORCHESTRATION IN S-AI-GPT
UNIT 4 Total Quality Management .pptx
The CXO Playbook 2025 – Future-Ready Strategies for C-Suite Leaders Cerebrai...
Infosys Presentation by1.Riyan Bagwan 2.Samadhan Naiknavare 3.Gaurav Shinde 4...
M Tech Sem 1 Civil Engineering Environmental Sciences.pptx
Model Code of Practice - Construction Work - 21102022 .pdf
Current and future trends in Computer Vision.pptx
CYBER-CRIMES AND SECURITY A guide to understanding
Internet of Things (IOT) - A guide to understanding
ASol_English-Language-Literature-Set-1-27-02-2023-converted.docx
FINAL REVIEW FOR COPD DIANOSIS FOR PULMONARY DISEASE.pptx
Enhancing Cyber Defense Against Zero-Day Attacks using Ensemble Neural Networks
573137875-Attendance-Management-System-original
TFEC-4-2020-Design-Guide-for-Timber-Roof-Trusses.pdf

Flash Steam and Steam Condensates in Return Lines

  • 1. Page 1 of 5 Flash Steam and Steam Condensates in Return Lines Jayanthi Vijay Sarathy, M.E, CEng, MIChemE, Chartered Chemical Engineer, IChemE, UK In power plants, boiler feed water is subjected to heat thereby producing steam which acts as a motive force for a steam turbine. The steam upon doing work loses energy to form condensate and is recycled/returned back to reduce the required make up boiler feed water (BFW). Recycling steam condensate poses its own challenges. Flash Steam is defined as steam generated from steam condensate due to a drop in pressure. When high pressure and temperature condensate passes through process elements such as steam traps or pressure reducing valves to lose pressure, the condensate flashes to form steam. Greater the drop in pressure, greater is the flash steam generated. This results in a two phase flow in the condensate return lines General Notes 1. To size condensate return lines, the primary input data required to be estimated is A. Fraction of Flash Steam and condensate, B. Flow Rates of Flash Steam & condensate, C. Specific volume of flash steam & condensates, D. Velocity limits across the condensate return lines. 2. Sizing condensate return lines also require lower velocity limits for wet steam since liquid droplets at higher velocities cause internal erosion in pipes and excessive piping vibration. A rule of thumb, for saturated wet steam is 25 – 40 m/s for short lines of the order of a few tens of metres and 15 - 20 m/s for longer lines of the order of a few hundred metres. 3. Condensate return lines work on the principle of gravity draining. To effectuate this, drain lines are to be sloped downward at a ratio of atleast 1:100. 4. Proper sizing of stem condensate return lines requires consideration of all operating scenarios, chiefly start up, shutdown and during normal running conditions. During plant start up, steam is not generated instantly. As a result, the condensate lines would be filled with liquids which gradually turn two-phase until reaching normal running conditions. During shutdown conditions, with time, flash steam in the lines condense leaving behind condensates due to natural cooling. 5. Condensate return line design must also consider the effects of water hammering. When multiple steam return lines are connected to a header pipe that is routed to a flash drum, flash steam in the presence of cooler liquid from other streams would condense rapidly to cause a water hammer. Fraction of Flash Steam Taking an example case, condensate flows across a control valve from an upstream pressure of 5 bara to 2 bara downstream. The saturation temperature at 5 bara is 151.84 0C & 120.20C at 2 bara. The specific volume of water at 5 bara is 0.001093 m3/kg & 0.00106 m3/kg at 2 bara. The latent heat of saturated steam upon reaching 2 bara is 2201.56 kJ/kg. The % flash steam generated is estimated as, ℎ𝑓,1 = ℎ𝑓,2 + [ % 𝐹𝑙𝑎𝑠ℎ 100 × ℎ𝑓𝑔] (1) Where, hf,1 = Upstream specific enthalpy [kJ/kg] hf,2 = Downstream specific enthalpy [kJ/kg] hf,g = Latent Heat of Saturated Steam [kJ/kg] The upstream specific enthalpy, hf1 of saturated water at 5 bara is 640.185 kJ/kg and hf2 of 504.684 kJ/kg at 2 bara. The steam specific volume at 2 bara is 0.8858 m3/kg.
  • 2. Page 2 of 5 The fraction of flash steam is calculated as, % 𝐹𝑙𝑎𝑠ℎ = [640.185−504.684] 2201.56 × 100 = 6.15% (2) Therefore the condensate fraction is, % 𝐶𝑜𝑛𝑑 = 100 − 6.15 = 93.85% (3) The steam volume is calculated as, 𝑉𝑆𝑡𝑒𝑎𝑚 = 0.8858 × 0.0615 = 0.05448 𝑚3 𝑘𝑔 (4) The condensate volume is calculated as, 𝑉𝐶𝑜𝑛𝑑 = 0.00106 × 0.9385 = 0.000995 𝑚3 𝑘𝑔 (5) Condensate Return Pipe Sizing To size the condensate return line, the bulk properties and mixture properties can be used to estimate the pipe size. It must be remembered that as the two-phase mixture travels through the pipe, there is a pressure profile that causes the flash % to change along the pipe length. Additionally due to the pipe inclination, a certain amount of static head is added to the total pressure drop. To estimate the pipe pressure drop across the pipe length, a homogenous model for modelling the two phase pressure drop can be adopted. The homogenous mixture acts as a pseudo-fluid, that obeys conventional design based on single phase fluids characterized by the fluid’s average properties. The mixture properties can be estimated as, 𝜌ℎ = 𝜌 𝐿[1 − 𝜀ℎ] + 𝜌 𝑣 𝜀ℎ (6) Where, L = Condensate Density [kg/m3] v = Steam Density [kg/m3] h = Homogenous void fraction for a given steam quality [x] [-] The homogenous void fraction [h] for a given steam quality [x] can be estimated as, 𝜀ℎ = 1 1+[ 𝑢 𝑣 𝑢 𝐿 × 1−𝑥 𝑥 × 𝜌 𝑣 𝜌 𝐿 ] (7) The dynamic viscosity for calculating the Reynolds number can be chosen as the viscosity of the liquid phase or a quality averaged viscosity, µh. 𝜇ℎ = 𝑥𝜇 𝑣 + [1 − 𝑥]𝜇 𝐿 (8) The homogenous model for gravitational pressure drop is applicable for large drop in pressures and mass velocities < 2000 kg/m2.s, such that sufficient turbulence exists to cause both phases to mix properly and ensure the slip ratio (uv/uL) between the vapour and liquid phase is ~1.0. For more precise estimates capturing slip ratios and varying void fraction, correlations such as Friedal (1979), Chisholm (1973) or Muller- Steinhagen & Heck (1986) can be used. The total pressure drop is the sum of the static head, frictional pressure drop & pressure drop due to momentum pressure gradient. ∆𝑃𝑇 = ∆𝑃𝑠𝑡𝑎𝑡𝑖𝑐 + ∆𝑃𝑚𝑜𝑚 + ∆𝑃𝑓𝑟𝑖𝑐 (9) The Static Head [Pstatic] is computed as, ∆𝑃𝑠𝑡𝑎𝑡𝑖𝑐[𝑏𝑎𝑟] = 𝐻×𝜌ℎ× 𝑔×𝑆𝑖𝑛𝜃 105 (10) Where, H = Pipe Elevation [m]  = Pipe inclination w.r.t horizontal [degrees] The pressure drop due to momentum pressure gradient [Pmom] is, 𝑑𝑃 𝑑𝑍 = 𝑑( 𝑚 𝜌ℎ⁄ ) 𝑑𝑍 (11) If the vapour fraction remains constant across the piping, the pressure drop due to momentum pressure gradient is negligible. The frictional pressure drop is calculated as, ∆𝑃𝑓 = 𝑓×𝐿×𝜌ℎ×𝑉2 2𝐷 (12) Where, P = Pressure drop [bar] f =Darcy Friction Factor [-]
  • 3. Page 3 of 5 L = Pipe Length [m] h = Mixture Density [kg/m3] V = Bulk fluid Velocity [m/s] D = Pipe Inner Diameter, ID [m] Re = DVρh µh (13) Where, µh = Dynamic Viscosity [kg.m/s] h = Homogenous Density [kg/m3] The Darcy Friction Factor [f] depends on the Reynolds number follows the following criteria, If Re <= 2100 ; Hagen Poiseuille’s Equation If Re <= 4000 ; Churchill Equation If Re > 4000 ; Colebrook Equation The Laminar Flow equation also referred to as the Hagen Poiseuille’s equation is, f = 64 Re (14) The Churchill equation combines both the expressions for friction factor in both laminar & turbulent flow regimes. It is accurate to within the error of the data used to construct the Moody diagram. This model also provides an estimate for the intermediate (transition) region; however this should be used with caution. The Churchill equation shows very good agreement with the Darcy equation for laminar flow, accuracy through the transitional flow regime is unknown & in the turbulent regime a difference of around 0.5- 2% is observed between the Churchill equation and the Colebrook equation. For Reynolds number up to ~4000, f = 8 [( 8 Re ) 12 + 1 (A+B)1.5 ] 1 12⁄ (15) A = [2.457ln ( 1 ( 7 Re ) 0.9 +0.27 ε D )] 16 (16) B = [( 37,530 Re )] 16 (17) The Colebrook equation was developed taking into account experimental results for the flow through both smooth and rough pipe. It is valid only in the turbulent regime for fluid filled pipes. Due to the implicit nature of this equation it must be solved iteratively. A result of suitable accuracy for almost all industrial applications will be achieved in less than 10 iterations. For Reynolds number up greater than ~4000, 1 √f = −2 log10 [ ε DH⁄ 3.7 + 2.51 Re√f ] (18) Homogenous Property Calculations The two phase mixture flows through the condensate return line. The associated density and viscosity of flash steam and condensate at 2 bara and 120.20C is, 𝜌 𝑣 = 1 0.8858 = 1.129 𝑘𝑔 𝑚3 (19) 𝜌 𝐿 = 1 0.00106 = 943.4 𝑘𝑔 𝑚3 (20) 𝜇 𝑣 = 0.000229 𝑘𝑔 𝑚.𝑠 (21) 𝜇 𝐿 = 0.0000128 𝑘𝑔 𝑚.𝑠 (22) The homogenous void fraction [h] for a slip ratio (uv/uL) of 1.0, i.e., uv = uL, and a steam quality [x] of 6.15% is, 𝜀ℎ = 1 1+[1× 1−0.0615 0.0615 × 1.129 943.4 ] = 0.9821 (23) The two phase homogeneous density is, 𝜌ℎ = 943.4 × [1 − 0.9821] + [1.129 × 0.9821] (24) 𝜌ℎ = 18.01 𝑘𝑔 𝑚3 (25) The two phase homogeneous viscosity is, 𝜇ℎ = 0.0615×1.28 10−5 + [1−0.9821]×2.29 10−4 (26) 𝜇ℎ = 0.000216 𝑘𝑔 𝑚.𝑠 (27) Pressure Drop Calculations The return condensate line from the control valve discharge is sloped at a ratio of 1:100 for gravity drain. The layout of the return condensate line is,
  • 4. Page 4 of 5 Figure 1. Condensate Return Line to Receiver The condensate receiver operates at 1.1 bara pressure. The mechanical details of the piping for a flow rate of 1,000 kg/h, pipe size of 1.5”, 100m length & pipe roughness of 45.2 m is, Table 1. Condensate Return Line Details Parameter Value Unit Mass Flow rate [m] 1000.0 kg/h Volumetric Flow [Q] 55.51 m3/h Pipe Length [L] 100 m Pipe Roughness [ε] 45.2 μm Pipe Outer Diameter [OD] 48.3 mm Pipe SMYS [Carbon Steel] 30,000 psi Pipe Design Pressure [DP] 7 bara Pipe Wall Thickness [WT] 0.08 mm Corrosion Allowance [CA] 1.0 mm Calculated WT 1.08 mm Selected WT 3.68 mm Pipe Inner Diameter [ID] 40.94 mm The pipe wall thickness chosen is based on ASME/ANSI B36.10M and is calculated based on the hoop stress created by internal pressure in a thin wall cylindrical vessel as, WT = DP×OD 2×SMYS = [7×14.5]×[ 48.3 25.4 ] 2×30,000 × 25.4 (28) WT = 0.08𝑚𝑚 (29) Adding CA of 1 mm, the WT becomes 1.08 mm. Based on ASME/ANSI B36.10M, the selected WT is 3.68mm. The inner diameter calculated for the selected WT is 40.94 mm. The condensate return line mixture fluid velocity is calculated as, V = Q A = 4×[ 1000 18.01 ]× 1 3,600 π×[0.04094]2 = 11.7 𝑚/𝑠 (30) The Reynolds number is estimated as, Re = ID×V×𝜌ℎ 𝜇ℎ = 0.04094×11.7×18.01 0.000216 (31) Re = 39,971 (32) Since the Reynolds number is much higher than 4,000, the flow is fully turbulent and the friction factor is calculated based on Colebrook equation. The friction factor is estimated as, 𝑓 = 𝑓𝐶𝑜𝑙𝑒𝑏𝑟𝑜𝑜𝑘 = 0.0251 (33) The frictional pressure drop is now calculated using the Darcy-Weisbach expression as, ∆𝑃𝑓 = 0.0251×100×18.01×11.72 2×0.04094×105 (34) ∆𝑃𝑓 = 0.757 𝑏𝑎𝑟 (35) The slope angle is calculated as, 𝜃 = [𝑇𝑎𝑛−1 ( 1 100 )] × 180 𝜋 = 0.6° (36) The static pressure drop [Pstatic] becomes ∆𝑃𝑠𝑡𝑎𝑡𝑖𝑐 = 18.01×9.81×[(1+5)×𝑠𝑖𝑛(0.6°)] 105 (37) ∆𝑃𝑠𝑡𝑎𝑡𝑖𝑐 = 0.000106 𝑏𝑎𝑟 (38) Therefore the total P with negligible P due to momentum pressure gradient [Pmom]. ∆𝑃𝑡𝑜𝑡𝑎𝑙 = ∆𝑃𝑠𝑡𝑎𝑡𝑖𝑐 + ∆𝑃𝑓 (39) ∆𝑃𝑡𝑜𝑡𝑎𝑙 = 0.757 + 0.000106 = 0.757 𝑏𝑎𝑟 (40) The condensate exit pressure is 2 – 0.757 = 1.243 bara which is higher than the receiver’s operating pressure of 1.1 bara. References 1. “Engineering Data Book III”, Ch 13, Two Phase Pressure Drop, Wolverine Tube, Inc. 2. “Steam Handbook”, Dr. Ian Roberts, Philip Stoor, Michael Carr, Dr. Rainer Hocker, Oliver Seifert, Endress+Hauser
  • 5. Page 5 of 5 Appendix A