Design, analysis and controlling of an offshore load transfer system Dimuthu Dharshana

Design, Analysis and Control
of an Offshore Load Transfer System

MAS501 Control Theory2- Autumn 2013
University of Agder
Grimstad Norway

Oreste Niyonsaba
Haocheng Su
Dimuthu Dharshana Arachchige
Bernard Sisara Gunawardana
Subodha Tharangi Ireshika
Jagath Sri Lal Senanayaka

Presentation Outline
 Introduction and methodology

 Crane kinematics
 Control architecture in Labview

 HIL Setup
 Control architecture in Step 7
 Results
 Discussion and Conclusion

Introduction and methodology
 Cranes are used in different Engineering activities
including offshore operations for transportation of loads
and personnel.
 Control design needs improvement to meet some of the
main criteria such as automatic tool tip tracking among
others
 Control system is aimed at maintaining the position of the
tool tip within one meter by one meter square relative to
an inertial frame of measurement is developed
 Firstly, the desired tool path is achieved in Labview, then
the control architecture is programmed in Siemens Step7
TIA environment integrated with Siemens ET200 PLC.

 Control Architecture is plugged with crane boom to
achieve the desired movement
 A DLL file is created by modeling HMF crane in
SimulationX in order to use it in Labview software as
shown if the figure below

 The desire is to implement the control architecture with
two P and PI controllers for controlling the tool tip
position

Crane kinematics from matlab
 Kinematics plays a significient role to establish
the traslational relastionships between global
coordinate system and the coordinate system of
crane components.

 Devided in to Forward kinematics, Inverse
kinematics, Forward Jacobian, Inverse Jacobian
and Actuator kinematics

Forward Kinematics
 Utilized to find the change of
tool tip position when the
angular positions are given
 Necessary to develop DH
(Denavit-Hartenberg) table

Inverse Kinematics
 Utilized to sketch the angular
position(q2,q3) when the tool
tip position (x,z)is known.

Forward Jacobian
 Utilized to skecth the
velocities of the actucators
when the angular velocities
are known.
 Maple was used to find the
parameters in the jacobian
matrix.

Inverse Jacobian
 Angular velociteis when the toll tip
velocities are known.
 It is obtained by differentiating the
inverse kinematics.
 This part is also implemented in
both matlab and labview.
 Inverse kinematics and inverse
Jacobian are crucial to convert the
reference positions and velocities of
the desired tool path into reference
angular positions and angular
velocities of two actuator joints.

Actuator kinematics
Used to derive relationships between angular movement of
the actuators and linear movements of the cylinders.
Actuator
Kinematics 2

Actuator
Kinematics 1

L3
L2

Law of Cosine
Analytical procedures are difficult

Polynomail Curve Fitting: ‘Polyfit’
Returns coefficients of the polynomial in descending powers.
L2 and Q2 Relation

L3 and Q3 Relation

140

40

120

20

100

0

80

-20

60

-40

40

-60

20

-80

0

-100

-20

-120

-40

-140

-60

1

1.1

1.2

1.3

1.4

1.5

1.6

1.7

1.8

1.9

2

-160
1.1

Root mean square error
Third order polynomial
Fifth order polynomial

1.2

1.3

1.4

1.5

1.6

0.0067
0.0037

1.7

1.8

1.9

2

2.1

Tool path in Labview
Operated in two segments

The path length (L) between the two locations and
angular position of the velocity vector are
geometrically found in order to guide the tool tip
along the shortest possible path

1. Initial position and target point
2. Reference square path

Tool point along the reference square path
Total Length of path s=3.914 m
s<L1

% First stretch

xDot=v;
zDot=0;
s<L1+L2
% First corner
x=R+L1+R*cos(q-pi/2);
z=R+R*sin(q-pi/2);
xDot=v*cos(q);
zDot=v*sin(q);
s<2*L1+L2)
% Second stretch
s<2*L1+2*L2)
% Second corner
s<3*L1+2*L2)
% Third stretch
s<3*L1+3*L2)
% Third corner
s<4*L1+3*L2)
% Fourth stretch
s<4*L1+4*L2)
% Fourth corner

Reference values for tool tip position and velocities are fed to
inverse Jacobian and inverse kinematics to configure the joint
angular movements .

Implementation of controllers in Labview

P_L2

PI_L2

P_L3
PI_L3

Propotional Gain
(P)

Integral Gain
(I)

P_L2 & P_L3

50

-

PI_L2 & PI_L3

0.05

0.001

Harware-In-Loop;Labview,PLC
& Real Time Target

TCP/IP configuration

Data Communication

Communication between Step7 and Labview

Configuration of the PLC

Siemens ET200S PLC

Tasks of digital switches in PLC
Switch

Addre
Task
ss in
PLC
Manual(Man_Sw) I3.0
Tool tip is moved to any location in the xz plane with
respect to the actuator velocities given by analog inputs
Point(Point_Sw)

I3.2

Tool tip is moved through the shortest path from current
location to any given target location by changing the
velocity given by analog input 1

Auto(Auto_Sw)

I3.1

It is required to enable the point switch before enabling
the auto switch. Due to the limitation in actuator lengths,
it is not possible to run the given reference path starting
from any arbitrary location in the xz plane. Hence, the tool
tip is firstly moved to a defined point (2.0, 1.5) which is
within the scope of controlling the actuator lengths.
Thereafter, the tool tip is moved along the reference tool
tip path by enabling auto switch. Velocity of the tool tip
can be controlled by changing the analog input 1.

Digital switch configuration
Manual

Auto

Point

Output

1

1

1

No operation

1

1

0

No operation

1

0

1

No operation

1

0

0

0

1

1

Manual switch is
enabled
No operation

0

1

0

Auto switch is enabled

0

0

1

Point switch is enabled

0

0

0

No operation

Digital switch configuration in main
controller block with SCL

Manual-1 Auto-1 Point-1


Implementation of the point switch in Step 7
 In order to achieve
the task of the
point switch,
point initializer
block and the
respective SCL
code is developed
 The Initial
Integrator block is
used to calculate
the distance ‘ss’,
travelled by tool
tip along the
linear reference
path from stating
point to target
point

Implementation of auto switch in Step 7

 ‘path’ function is defined to configure the reference tool tip position and
reference velocity.
 The reference tool path is implemented in Step 7 in SCL language
 An integrator is used to calculate the distance travelled by tool tip along
the reference path, ‘s’.

Crane kinematics and PID controllers in Step 7
 Inverse kinematics, Inverse jacobian and
actuator kinematics are developed in Step
seven as FC blocks and the same programs
used in labview were implemented in SCL
language.
 Two P controllers and two PI controllers were
implemented with CONT_C block and
parameters were tuned accordingly.

Controller
P_L2
P_L3
PI_L2Dot
PI_L3Dot

400
400
0.05
0.05

Tuned values
P
I
0
0
0.01
0.01

LabView Results: Path

 Test was sucessful
 Maximum +/- 10 cm
deviation form
expected path
Reference Path

Actual Path

LabView Results: Controllers Operation

Actuator L2 position and velocity controllers

Actuator L3 position and velocity controllers

HiL Test Results: Expected Path

HiL Test Results : Manual and Start Point
Initialization Operations
 Manual and tip point initialize(shortest path finder) control functions
to make the system operation more robust and safe.

Manual Mode Operation

Start Point Initialization Operations

HiL Test Results: Path
 HiL test was very sucessful
 Maximum +/-5 cm deviation form expected path( only in two edges).

Start Point Initialization Operations

Auto mode Actual Path

Real HMF Crane Test Results: Path
 Real crane test was very sucessful
 Maximum +/- 5 cm error in expected path
 Approximate 35 cm offset in x direction.

HiL Test Actual Path

Real Test Actual Path

Why 35 cm Offset ?
There is 34.3 cm offset in x direction
of the coordinate systems used in dll
file and real crane.

Discussion and Conclusions

 Lab view results were observed with +/- 5-10 cm deviation.
 Possible resons could be computer programs were not used dedicated
hardware to run controllers and DLL model.
 Modelling errors in DLL file and our sysem model
 With HIL test we used dedicated hardware to run the controllers and
realtime tartget(DLL model). Maximum +/-5 cm deviation form expected
path( only in two edges) was observed. HiL Test was very sucessful.
 Modelling errors in DLL file and our sysem model
 With real experiment it was observed Maximum +/- 5 cm error in
expected path and 35 cm offset in x direction. Test was very successful.
 Modelling errors in our system model.
 There seems 35 cm offset in x direction of DLL model output than real
crane. We could solve this problem with little modification in system.

Discussion and Conclusions (Cont.)
Solution to the offshore load transfer system
 2D -> 3D Solution
 Reference target is required to get from sensors of floating platform

Tool tip target tracking operation

Discussion and Conclusions (Cont..)
A real engineering experience with 3 steps..

1) Numarial Modeling/Simulation :
 Make the system with more software models(Lab view, dll model)
 Less acuarate but easy to change and think.
2) HiL Method:
 Make the sysem with more real equipments (PLC and switches) and
improved model sysems (DLL model run on real time labview,
Simulink model)
 More acurate results and closer to reality.
 It is cost saving safe method, befoure doing the actual task.
 sucess of this method highly depend on acuracy of real system model.
Here the given DLL model was acurate as the real machine works
accordingly.
3) Real world test with actual devices (PLC, HMF real crane and real switches)
 Solve the real world problem of offshore load transfer system

Design, analysis and controlling of an offshore load transfer system Dimuthu Dharshana

More Related Content

What's hot (18)

Viewers also liked (17)

Similar to Design, analysis and controlling of an offshore load transfer system Dimuthu Dharshana (20)

Recently uploaded (20)

Design, analysis and controlling of an offshore load transfer system Dimuthu Dharshana