This action might not be possible to undo. Are you sure you want to continue?
COUPLED WITH MOORING LINES AND RISERS
A Dissertation
by
YOUNGBOK KIM
Submitted to the Office of Graduate Studies of
Texas A&M University
in partial fulfillment of the requirements for the degree of
DOCTOR OF PHILOSOPHY
May 2003
Major Subject: Ocean Engineering
DYNAMIC ANALYSIS OF MULTIPLEBODY FLOATING PLATFORMS
COUPLED WITH MOORING LINES AND RISERS
A Dissertation
by
YOUNGBOK KIM
Submitted to the Office of Graduate Studies of
Texas A&M University
in partial fulfillment of the requirements for the degree of
DOCTOR OF PHILOSOPHY
Approved as to style and content by:
MooHyun Kim Cheung H. Kim
(CoChair of Committee) (CoChair of Committee)
Jun Zhang Robert H. Stewart
(Member) (Member)
Paul N. Roschke
(Head of Department)
May 2003
Major Subject: Ocean Engineering
iii
ABSTRACT
Dynamic Analysis of MultipleBody Floating Platforms Coupled with Mooring
Lines and Risers. (May 2003)
YoungBok Kim, B.S., Inha University;
M.S., Seoul National University
CoChairs of Advisory Committee: Dr. MooHyun Kim
Dr. Cheung H. Kim
A computer program, WINPOSTMULT, is developed for the dynamic analysis
of a multiplebody floating system coupled with mooring lines and risers in the presence
of waves, winds and currents. The coupled dynamics program for a single platform is
extended for analyzing multiplebody systems by including all the platforms, mooring
lines and risers in a combined matrix equation in the time domain. Compared to the
iteration method between multiple bodies, the combined matrix method can include the
N N 6 6 × full hydrodynamic interactions among N bodies. The floating platform is
modeled as a rigid body with six degrees of freedom. The first and secondorder wave
forces, added mass coefficients, and radiation damping coefficients are calculated from
the hydrodynamics program WAMIT for multiple bodies. Then, the time series of wave
forces are generated in the time domain based on the twoterm Volterra model. The wind
forces are separately generated from the input wind spectrum and wind force formula.
The current is included in Morison’s drag force formula. In the case of FPSO, the wind
and current forces are generated using the respective coefficients given in the OCIMF
iv
data sheet. A finite element method is derived for the long elastic element of an arbitrary
shape and material. This newly developed computer program is first applied to the
system of a turretmoored FPSO and a shuttle tanker in tandem mooring. The dynamics
of the turretmoored FPSO in waves, winds and currents are verified against independent
computation and OTRC experiment. Then, the simulations for the FPSOshuttle system
with a hawser connection are carried out and the results are compared with the
simplified methods without considering or partially including hydrodynamic interactions.
v
ACKNOWLEDGEMENTS
This work was completed only because of the financial support of the OTRC and
JIP (Joint Industry Project) for over four years. I deeply thank the sponsors for this
support. I would like to express my sincere gratitude to my advisors, Dr. M. H. Kim and
Dr. C. H. Kim, for their continuous encouragement and guidance during my studies. I
also would like to thank Dr. Zhihuang Ran (Alex) and Dr. Arcandra Tahar for sharing
their efforts to review the programming and to discuss the problem. I greatly appreciate
Dr. J. Zhang and Dr. R. H. Stewart for serving as advisory committee members, Dr. R.
Mercier for releasing the OTRC experiment data, and Dr. E. B. Portis for supervising the
procedure of the final defense as a GCR.
Finally, I would like to thank my wife, DeockSeung Seo, for her support and
encouragement during the period of this study.
This work could only be done under the merciful guidance and the tender love of
God. I would like to devote this work to His Glory.
vi
TABLE OF CONTENTS
Page
ABSTRACT ………………………………………………………………………… iii
ACKNOWLEDGEMENTS ……………………………….….…………………… v
TABLE OF CONTENTS …………………………………….……….…………….. vi
LIST OF FIGURES ………………………………………….……….……………... x
LIST OF TABLES ………………………………………….……….…………….… xiv
CHAPTER
I INTRODUCTION …………………………………….………………… 1
1.1 Background…………..…………………..…………….………….……. 1
1.2 Literature Review …………..……………..…………….………….…... 3
1.3 Objective and Scope ………..……………..…………….………….…... 5
1.4 Procedure …………………………………..…………….………….….. 7
1.4.1 Interpretation and Preparation of WAMIT Results and Wind/
Current Forces ………………………….………………….……. 7
1.4.2 Developing the Coupled Dynamic Program ………….………..… 8
1.4.3 Comparative Studies …………………………….……………….. 10
II DYNAMICS OF THE FLOATING PLATFORM ………..…….……….. 12
2.1 Introduction ……………………………..…………….………….……. 12
2.2 Formulation of Surface Wave ………………….………….…….…….. 12
2.2.1 Boundary Value Problem (BVP) of Surface Wave ……..………. 12
2.2.2 Wave Theory ……………………………………………………. 14
2.2.3 Diffraction and Radiation Theory …………….………………… 16
2.2.3.1 FirstOrder Boundary Value Problem ……………………. 17
2.2.3.2 SecondOrder Boundary Value Problem ……..…………… 19
2.3 Hydrodynamic Forces ……………………………….…………………. 23
2.3.1 The FirstOrder Hydrodynamic Forces and Moments …………... 23
2.3.2 The SecondOrder Hydrodynamic Forces and Moments …….….. 26
2.4 MultipleBody Interaction of Fluid …………….………………………. 28
2.5 Boundary Element Method …………………………………………….. 30
2.6 Motions of the Floating Platform ………………………………………. 33
2.6.1 Wave Loads ……………………………………..……………….. 33
2.6.2 Morison’s Equation ……………..……………………………….. 36
vii
CHAPTER Page
2.6.3 Single Body Motion …………..………………………………… 37
2.6.4 Multiple Body Motion ………………….……………………….. 38
2.6.5 Time Domain Solution of the Platform Motions …….…………. 40
III DYNAMICS OF MOORING LINES AND RISERS …………….……. 44
3.1 Introduction ……………………….…………………………………… 44
3.2 Theory of the Rod ……………………………………………………… 46
3.3 Finite Element Modeling ………………………………………………. 50
3.4 Formulation of Static Problem …………………………………………. 55
3.5 Formulation for Dynamic ProblemTime Domain Integration …….…… 59
3.6 Modeling of the Seafloor ……………………………………………….. 63
IV COUPLED ANALYSIS OF INTEGRATED PLATFORM AND
MOORING SYSTEM …………………………………………………... 66
4.1 Introduction …………………………………..………………………… 66
4.2 The Spring to Connect the Platform and the Mooring System…………. 67
4.2.1 Static Analysis …………………………………………………... 69
4.2.2 TimeDomain Analysis ………………………………………….. 71
4.3 Modeling of Damper on the Connection ………….……………………. 72
4.4 Modeling of Connection between Lines and Seafloor ……..…………… 74
4.5 Formulation for the Multiple Body System ……………………..……… 75
V CASE STUDY 1: DYNAMIC ANALYSIS OF A TANKER BASED
FPSO ………………………………………………………………….…. 79
5.1 Introduction ……………………………………………………………. 79
5.2 Design Premise Data of FPSO and Mooring Systems ……..………….. 80
5.3 Environmental Data …………………………………………………… 85
5.3.1 Wave Force ……………………………………..………………. 87
5.3.2 Wind Force ………………………………………………………. 88
5.3.3 Wind and Current Forces by OCIMF …………….……………… 90
5.4 Hydrodynamic Coefficients ……………………………………………. 93
5.5 Coupled Analysis of FPSO …………..………………………………… 95
5.6 Results and Discussion ..……..…………..……………………………... 98
5.6.1 Static Offset Test (in Calm Water without Current) ………..…… 99
5.6.2 Freedecay Tests (in Calm Water without Current) ……………. 101
5.6.3 Timedomain Simulation for Hurricane Condition ………..……. 103
5.7 Summary and Conclusions …………………………………………….. 106
viii
CHAPTER Page
VI CASE STUDY 2: DYNAMIC ANALYSIS OF A TANKER BASED
FPSO COMPARED WITH THE OTRC EXPERIMENT ………………. 108
6.1 Introduction …………………………………………………………... 108
6.2 OTRC Experimental Results and Design Premise Data ………….….. 109
6.3 Environmental Data ………………………………………………….. 114
6.4 Regeneration of the Experimental Model ………………………….. 116
6.5 Results and Discussion ……..…………….……………………..…… 119
6.5.1 Static Offset Test with Regenerated Model Data …………….. 119
6.5.2 FreeDecay Test with Regenerated Model Data ….………….. 120
6.5.3 Time Simulation Results …………………………...…………. 123
6.6 Summary and Conclusions ………………………………………….. 125
VII CASE STUDY 3: CALCULATION OF HYDRODYNAMIC
COEFFICIENTS FOR TWO BODY SYSTEM OF FPSO AND
SHUTTLE TANKER ……………………………………………….…. 126
7.1 Introduction …………………………..……………………………… 126
7.2 Particulars of Models and Arrangements for the Tests ……………… 128
7.3 Environmental Conditions …………………………………………… 132
7.4 Results and Discussion ………..…………………….…..…………… 133
7.5 Summary and Conclusions …………………………………………... 141
VIII CASE STUDY 4: DYNAMIC ANALYSIS FOR TWOBODY
SYSTEM COMPOSED OF SPAR AND SPAR …………………….….. 142
8.1 Introduction ………………………………………………………….. 142
8.2 Particulars of Models and Arrangements for the Analyses …….……. 143
8.3 Environmental Conditions …………………………………………… 146
8.4 Calculation of Hydrodynamic Coefficients Using WAMIT 1
st
and
2
nd
Order ………….……………………………………….………... 147
8.5 Linear Spring Modeling ………..………………….…….….……….. 149
8.6 Results and Discussion ………..……………….……..………….…... 149
8.7 Summary and Conclusions ……………………………….………..… 154
IX CASE STUDY 5: DYNAMIC ANALYSIS FOR TWOBODY
SYSTEM COMPOSED OF AN FPSOFPSO AND AN FPSO
SHUTTLE TANKER …………………………………………..…….… 155
9.1 Introduction …………………………………………………………. 155
9.2 Particulars of Models and Arrangements for the Analyses ….……… 156
9.3 Environmental Conditions …………………………………………... 160
ix
CHAPTER Page
9.4 Calculation of Hydrodynamic Coefficients Using WAMIT ..……….. 162
9.5 TwoMassSpring Modeling …..………………….………………….. 164
9.6 Results and Discussion ………..………………..……………………. 174
9.7 Summary and Conclusions …………………………………………… 200
X CONCLUSIONS FOR ALL CASE STUDIES ……………….…….…… 201
REFERENCES ………………………..…………………….……….……………… 203
VITA …………………………………..…………………….……….……………… 208
x
LIST OF FIGURES
FIGURE Page
3.1 Coordinate system of rod ………………………………………………………. 46
5.1 The body plan and the isotropic view of FPSO 6,000 ft ………………………. 82
5.2 Arrangement of the mooring lines for FPSO 6,000 ft. ………………………… 84
5.3 Arrangement of the risers for FPSO 6,000 ft. ………………………………….. 85
5.4 JONSWAP wave spectrum ……………………………………………………. 88
5.5 API wind spectrum ……………………………………………………………. 89
5.6 Modeling of body surface of FPSO …………………………………………… 94
5.7 Modeling of body surface and free surface of the water ……………………… 95
5.8 Hull drag damping coefficients (Wichers, 1996) ……………………………… 97
5.9 Static offset test results for surge motion …………………… ………………… 100
5.10 Freedecay test results for surge, heave and roll motions …………..………... 102
6.1 General arrangement and body plan of FPSO 6,000 ft ………………………... 110
6.2 Arrangement of mooring lines for turretmoored FPSO ……………………… . 113
6.3 NPD wind spectrum curve .……………………………………………………. 115
6.4 Comparison of the static offset test results ……………………………………. 121
6.5 Hull drag coefficients proposed by Wichers (1998 & 2001) ………………….. 122
7.1 Configuration of the mooring system …………………………………………. 131
7.2 Roughmeshed numerical modeling for a LNG FPSO and a shuttle tanker …… 132
7.3 Finemeshed numerical modeling for a LNG FPSO and a shuttle tanker ……… 132
xi
FIGURE Page
7.4 Heave response operators of sidebyside moored vessels in the beam sea …… 134
7.5 Roll response operators of sidebyside moored vessels in the beam sea ……... 135
7.6 Longitudinal wave drift force of tandem moored vessels in the head sea …….. 136
7.7 Longitudinal wave drift force of sidebyside moored vessels in the head sea … 137
7.8 The distance effect on the longitudinal wave drift force for a twobody
and a single body model in the head sea ……………………………………….. 138
7.9 Lateral wave drift force of sidebyside moored vessels in the head sea …….… 139
7.10 Lateral wave drift force of sidebyside moored vessels in the beam sea ……. 140
8.1 Configuration of the mooring system and the environmental loads
(Tandem arrangement, d=30m)………………………………………………... 144
8.2 Configuration of the modeling of a single spar ……………………………….. 148
8.3 Configuration of the modeling of a twobody spar …………………………… 148
8.4.a Comparison of the surge motion RAOs …………………………………….. 151
8.4.b Comparison of the heave motion RAOs …………………………………….. 151
8.4.c Comparison of the roll motion RAOs ……………………………………….. 152
8.5 Comparison of the surge drift force ……………………………………………. 152
9.1 Configuration of the mooring systems (Tandem mooring system)…………….. 158
9.2 Configuration of the arrangement of the mooring line groups ………………… 159
9.3 Configuration of singlebody, twobody models and the system ……………… 163
9.4 Twomassspring model ……………………………………………………….. 165
9.5 The diagram of the time simulation in SIMULINK of MATLAB ……………. 168
xii
FIGURE Page
9.6 The surge motion of the FPSO and FPSO model by MATLAB for massspring
model and by WINPOSTMULT for twobody model ………………………. 169
9.7 The time simulation results of the FPSO and shuttle tanker model ………….. 172
9.8.a Time simulation for the two body model of the FPSO and shuttle tanker
(at body #1=FPSO; tandem; without interaction effect) …………………….. 176
9.8.b Time simulation for the two body model of the FPSO and shuttle tanker
(at body #2=shuttle tanker; tandem; without interaction effect) …..………. 178
9.8.c Amplitude spectrum density curve of the motion responses
for the two body model of the FPSO and shuttle tanker
(at body #1=FPSO; tandem; without interaction effect) …………………….. 180
9.8.d Amplitude spectrum density curve of the motion responses
for the two body model of the FPSO and shuttle tanker
(at body #2=shuttle tanker; tandem; without interaction effect) …………….. 182
9.9.a Time simulation for the two body model of the FPSO and shuttle tanker
(at body #1=FPSO; tandem; with interaction effect and by iteration method).. 184
9.9.b Time simulation for the two body model of the FPSO and shuttle tanker
(at body #2=shuttle tanker; tandem; with interaction effect
by iteration method) ………………………………………………………….. 186
9.9.c Amplitude spectrum density curve of the motion responses
for the two body model of FPSO and shuttle tanker
(at body #1=FPSO; tandem; with interaction effect by iteration method) …… 188
9.9.d Amplitude spectrum density curve of the motion responses
for the two body model of FPSO and shuttle tanker
(at body #2=shuttle tanker; tandem; with interaction effect
by combined method) ………………………………………………………… 190
9.10.a Time simulation for the two body model of the FPSO and shuttle tanker
(at body #1=FPSO; tandem; with interaction effect by combined method) … 192
9.10.b Time simulation for the two body model of the FPSO and shuttle tanker
(at body #2=shuttle tanker; tandem; with interaction effect
by combined method) ………………………………………………………... 194
xiii
FIGURE Page
9.10.c Amplitude spectrum density curve of the motion responses
for the two body model of the FPSO and shuttle tanker
(at body #1=FPSO; tandem; with interaction effect by combined method) … 196
9.10.d Amplitude spectrum density curve of the motion responses
for two body model of FPSO and shuttle tanker
(at body #2=shuttle tanker; tandem; with interaction effect
by combined method) ……………………………………………………….. 198
xiv
LIST OF TABLES
TABLE Page
5.1 Main particulars of the turret moored FPSO 6,000 ft …………………………. 81
5.2 Main particulars of mooring systems …………………………………………. 83
5.3 Hydrodynamic coefficients of the chain, rope and polyester …………………. 83
5.4 Main particulars of risers ……………………………………………………… 84
5.5 Hydrodynamic coefficients of risers …………………………………………... 84
5.6 Azimuth angles of risers bounded on the earth ………………………………... 85
5.7 Environmental loading condition ……………………………………………… 86
5.8 Natural periods from freedecay tests …………………………………………. 103
5.9 Damping from freedecay tests estimated from the first 4 peaks
assuming linear damping ……………………………………………………… 103
5.10 Timedomain simulation results ……………….……………………………... 104
5.11 The results of tensions on the mooring lines and risers …………………….... 105
6.1 Main particulars of the turret moored for the OTRC FPSO ……………………. 111
6.2 Main particulars of mooring systems for the OTRC FPSO ……………….……. 112
6.3 Hydrodynamic coefficients of the chain, rope and wire for the OTRC FPSO
…………………………………………………………………………………… 112
6.4 Environmental loading condition for the OTRC FPSO ………………………… 114
6.5 WAMIT output and handcalculation …………………………………………. 117
6.6 Reestimated data from WAMIT output and handcalculation ……………….. 119
6.7 Comparison of the free decay test results ……………………………………… 122
xv
TABLE Page
6.8 Comparison of time simulation results ………………………………………… 124
7.1 Main particulars of two vessels ………………………………………………... 129
7.2 Freedecay test results for a LNG FPSO and a shuttle tanker
(heave and roll) ……………………………………………………………….... 130
7.3 Comparison of the hydrodynamic coefficients obtained from the rough model
and the fine models …………………………………………………………….. 131
8.1 Main particulars of moored spar ………………………………………………. 144
8.2 Particulars of the mooring systems ……………………………………………. 145
8.3 Environmental conditions ……………………………………………………… 146
8.4 The analysis results for twobody model composed of two spars …………….. 153
9.1 Main particulars of the turret moored FPSO …………………………………… 157
9.2 Main particulars of the mooring systems ………………………………………. 158
9.3.a Environmental conditions (100year storm condition at GoM) ……………… 161
9.3.b Environmental conditions (west Africa sea condition) ………….…………… 161
9.4 The system parameters for twomassspring model …………………………… 168
9.5 Analysis results of massspring model: displacement at mass #1 and #2 ……… 170
9.6 Summary of the analysis results for two body FPSO+FPSO ………………….. 171
9.7 Summary of the analysis results for the twobody FPSO+shuttle tanker ……….. 173
1
CHAPTER I
INTRODUCTION
1.1 Background
Recently, floating structures have been invented and their installation has been
attempted worldwide because of cost effectiveness, in an attempt to replace traditional
fixed jacket platforms. These structures include the shipshaped vessel called an
FPSO(Floating Production Storage and Offloading Unit), the column stabilized semi
submergible platform, the spar platform, and the tension leg platform(TLP). The last two
types have been designed and installed in the Gulf of Mexico(GoM) for the last decade.
In the case of TLPs, there were several built and installed in GoM, of which Auger, Mars,
Ursa, and Marlin were fixed in position by means of the mooring lines or risers in 2,800
ft to 4,000 ft of water depth. In the case of spars, Neptune, Genesis, and Diana were
installed in 2,000 ft, 2,590 ft, 4,300 ft of water depth, respectively. These installations
were made from 1996 to 1999. Nowadays, the truss spar is being considered more cost
effective. The recent trend in the installation of floating structures shows the water depth
getting deeper and deeper since the oil and gas fields are expedited and discovered in the
deeper sea. This means the more developed designs should be invented and studied
realistically for the installation of the floating structures in deep water of 6,000 ft or
more. Floating structures are more attractive to the industrial companies
This dissertation follows the style and the format of the Journal of Fluid Mechanics.
2
because they can allow for environmental conditions more flexibly than the fixed
structures.
They have more advantages in that they have been designed under the concept of
optimization and minimization against the responses to environmental conditions. For
the spars, they have small water plane areas compared with other floating structures.
This results in reducing the heave response by decreasing the vertical wave load and
shifting the heave natural frequency in the low part far apart from the wavedominant
frequency. The surfaceproduction trees and rigid risers are allowed due to the above
mentioned aspect of design, instead of the subsea trees and flexible risers that are more
expensive. For the TLPs, the highstrength vertical tethers are normally used. It results in
avoiding the resonance between the motion of TLPs and the wave excitations so that it is
able to stay more stable while operating during oil or gas extraction, and it allows using
the surfaceproduction trees. For the floating structures in deep water, many researchers
have proved that coupled dynamic analyses are indispensable to get more convincing
results from the platform responses and the line tensions than those of conventional
uncoupled analysis methods (Pauling and Webster, 1986; Kim et al., 1994; Ran and Kim,
1997; Ran, Kim and Zheng, 1999a; Ran, Kim and Zheng, 1999b; Ma et al., 2000). Since
the shipshaped floating structures called FPSOs have more advantages as the solutions
to comparably large deck space, costsaving problems and less risk of oil spills, they will
have to be potentially attractive production systems in ultra deep water of the GOM.
Nowadays, the Mineral Management System (MMS) has approved the installation of an
FPSO under the condition that the vessel has the construction of a double hull tanker in
3
the GOM. The large storage capacity is the biggest advantage because no pipeline has to
be laid out from the sea floor to the land. A kind of LNG carrier or oil shuttle tanker is
substituted for the pipelines for the purpose of turning over the oil and gas. For the
installation of FPSO in deep water such as GoM, the development of a coupled dynamic
analysis code for solving the large yaw motion and the interaction problem of multiple
body system becomes indispensable.
1.2 Literature Review
The comprehensive studies about the viscous dampings for dynamic motion
analysis of the turretmoored FPSO were performed by Wichers(1988). He derived the
equation of the motions of a singlepointmoored FPSO exposed to current, wind and
longcrested irregular waves, and carried out the nonlinear analysis by uncoupled
method, which solves the motions of body and mooring lines, separately. The coupling
effects of the low frequency component of a viscous reaction force were studied by
Wichers and Chunqun Ji (2000). By conducting a series of experimental studies, they
examined the coupling terms due to the combined modes of motion in still water and in
the current. They proved the viscous part in a normal direction contributes significantly
to the hull dynamics, so that it cannot be neglected. In addition, the coupling effect of
rigid body motion and the motions of the mooring lines and risers was investigated by
Wichers and Devlin(2001). The fully coupled dynamic mathematical model is necessary
to estimate realistic motion responses and line tensions.
4
The extreme response of a turret moored FPSO in GoM was studied by Baar et al.
(2000). The dynamic motion of FPSO on collinear, noncollinear wind, the wave and
current of a 100year return period storm was investigated so that it was verified that the
response of a turret FPSO is sensitive to noncollinear environmental conditions. Ward
et al.(2001) presented the results of experiments conducted in OTRC(Offshore
Technology of Research Center in Texas A&M University) for a turretmoored FPSO in
collinear and noncollinear environmental conditions. The hull/mooring/riser coupled
analyses of a tankerbased turretmoored FPSO was carried out by Arcandra et al. (2002)
using a coupled dynamic analysis tool for floating structures, developed by him. They
investigated two types of mooring system of the polyester mooring lines and buoy type
mooring lines through the time simulation of FPSO 6,000 ft under the conditions of 100
year hurricane.
The aspects of the hydrodynamic characteristics of the multiplebody structure
combined with a barge and a mini TLP were studied by Teigen (2000). He compared the
hydrodynamic coefficients of the multiplebody and the singlebody and also conducted
the convergence tests according to the mesh size of the multiple body. He emphasized
the importance of hydrodynamic interaction for the motion response of two bodies and
indicated the fact that neglecting the fluidcoupling effect may result in an erroneous and
nonconservative prediction. Using a threedimensional source technique, Inoue et al.
(2001) solved the drift force for a multiplebody system of the FPSOLNG carrier in
parallel arrangement with zero forward speed waves. By adding the viscous roll damping
to the potential damping, the study was attempted to compare the effect on drift forces
5
with experimental results in regular and irregular waves. For a multibody system with a
sidebyside mooring of an FPSO and an LNG carrier, a linear potential solver was
developed by Huijsmans (2001), and the mean and lowfrequency wave drift forces were
calculated by using it. For the same model, Buchner et al. (2001) conducted the
numerical simulation for the prediction of hydrodynamic responses of an LNG FPSO
with alongside moored an LNG carrier. They used a free surface lid in this multiple
body diffraction analysis for the calculation of drift forces and a relative viscous
damping in a horizontal plane, and the composition of the complete matrix of retardation
function for the correct prediction of heave and pitch motions. The hydrodynamic
interaction of forces and motions of the floating multiplebody was investigated using
the WAMIT program (Clauss et al., 2002) and the higherorder boundary element
method (Choi et al., 2002).
1.3 Objective and Scope
The main objective of this research is to develop a numerical program to analyze the
hydrodynamic interaction responses of multiple bodies, mooring lines and risers based
on the hull/mooring/riser coupled dynamic program called WINPOSTFPSO(Arcandra,
2001), using the hydrodynamic coefficients calculated by WAMIT (Lee, 1999)
considering the interaction effects of the multiplebody.
The first stage consists of the evaluation and interpretation of the hydrodynamic
interaction analysis results with WAMIT and the preparation of the wind and current
force data (OCIMF, 1994) for performing the coupled dynamic analysis program newly
6
developed (WINPOSTMULT) for the shipshaped multiplebody system (FPSO, LNG
carrier etc.). The interpretation program (WAMPOSTMULT) of the WAMIT results
will be made for the preparing the properly formatted data for WINPOSTMULTI. For
the wind and current forces, a modification in some parts of the original program
(WINPOSTFPSO) will be needed.
In the second stage of this research, the original program (WINPOSTFPSO) will be
developed to be able to perform the hull/mooring/riser coupled dynamic analysis for
general multiple floating bodies. In the new program, it will be considered that the
multiple bodies can be laid in any relative position to the open sea. The wave heading
angle will be considered separately for each body at every small degree of angle and the
relative angles between multiple bodies will be considered at every span in the same
manner as for the wave heading angle.
The third stage is to prove the validity of the newly developed program through
carrying out the numerical simulation after the proper models are selected. Buchner’s
model (2001) and Choi’s model (2002) may be used for a comparative study about the
results to be obtained from WINPOSTMULT. The former has the characteristics to deal
with the close proximity problem of a sidebyside offloading system. The latter took
two, same sized vessels of an FPSO and a shuttle tanker to tackle the problems of both
cases of the sidebyside system and the tandem system, and used the higherorder
boundary element method (HOBEM) while the constant panel method(CPM) was used
in WAMIT. The coupled dynamic analysis scheme adopted in the program WINPOST
7
MULT will be proved as the robust tool for analyzing the interaction problem of the
multiplebody floating structure.
1.4 Procedure
1.4.1 Interpretation and Preparation of WAMIT Results and Wind/Current
Forces
For the calculation of the hydrodynamic coefficients and wave forces, WAMIT
(1999) will be used. WAMIT will give the results of N × 6 degree of freedoms (DOFs)
for N bodies in consideration of the N body interaction. WAMIT should be run for each
contacting angle between N bodies at every small angle. It will give the hydrodynamic
interaction coefficients of added mass and damping and wave forces. The added mass
and wave drift damping will be given as a matrix sized by (NFREQ x 6N x 6N), where
NFREQ means the number of frequencies of the wave. The wave forces will be given as
the linear wave force transfer function (LTF), sized by (NFREQ x 6N) and as the sum
and differencefrequency components sized by (NFREQ x NFREQ x 6N). WAMIT
should be prerun for each contacting angle between Nbodies at every small angle of
wave heading and at every small amount angle of contact with each body for the
expected positions. These results will be converted as the input data (each input data file
will be named as data000.wv) for WINPOSTMULT. For the preparation of the input
data, one converting program (WAMPOSTMULT) will be made.
The wind and current forces subject to any shipshaped floating structures can be
referred to the OCIMF (1994). For the full loading and the ballast condition, wind and
8
current forces and moments can be read from the tables in the booklet published by
OCIMF (1994). They also will be prepared prior to running the WINPOSTMULT. In
the WINPOSTMULT, the two data files will be read, and the real drafts of the subjected
vessels will be recognized as the draft ratio to the full draft. During the running of the
program WINPOSTMULT, the angles against wave headings and the relative angles
between multiple bodies will be checked at every time step. If the angles exceed the
initial angle, the wind/current forces and moments for the updated angle will be read
from the files of the hydrodynamic coefficients precalculated for every 5 degree of yaw
angle.
1.4.2 Developing the Coupled Dynamic Program
The backborn program, WINPOSTFPSO, is already developed by Arcandra(2001).
For the N bodies, the dealing DOF number should be set up as 6N and the related
subroutines should be modified. WINPOSTFPSO is a coupled dynamic program that
can treat the body and rods(mooring lines and risers). For N bodies, the total equations
of motion for the total system will be combined with the mooring line dynamic
equations. For a single body system, the final equation of motion with a combination of
the coupling terms of a single body and mooring lines/risers is obtained as:
)
`
¹
¹
´
¦
=
)
`
¹
¹
´
¦
(
¸
(
¸
B
L
B
L
B C
C L
F
F
U
U
K ) (K
K K
T
where, subscripts of r, c and b mean the rod, the coupled term and the body, respectively.
If the total number of mooring lines and risers of the system is defined as
L
n , the
9
matrices in the above equation, where the equations and figures in the parentheses after
the matrix name mean the matrix size, are defined as follows:
L
K ( ) bandwidth ( ]) 1 ) 1 ( 8 [ ( × − + × ×
E L
n n ) = the stiffness matrix of mooring lines and
risers
C
K ( ) 6 ( ] 1 ) 1 ( 8 [ N n n
E L
× × − + × × ) = the stiffness matrix coupled with the body and
mooring lines/risers
B
K ( N N 6 6 × ) = the motion matrix of the body
L
U ( 1 ]) 1 ) 1 ( 8 [ ( × − + × ×
E L
n n ) = the motion vector of mooring lines and risers
B
U ( 1 6 × N ) = the motion vector of the body
L
F ( 1 ]) 1 ) 1 ( 8 [ ( × − + × ×
E L
n n ) = the external force vector subject to mooring lines
and
risers
B
F ( 1 6 × N ) = the external force vector subject to the body
where
E
n is the number of elements per one line, the bandwidth is 15, and N denotes
the number of bodies to be considered. For the multiple body system of N bodies, the
rigid bodies are lumped at N points with N 6 DOFs, which are connected with springs
and dampers to the mooring lines and risers. The number of DOFs of
B
U will be
enlarged to N 6 as much as the number of DOFs for multiple bodies. Furthermore, the
part of the program to deal with multiplebody systems needs to be modified for reading
10
the hydrodynamic coefficients and wave forces for the proper contacting angle at every
time step, and for evaluating and assigning to the external forces of the wind and current
forces for the loading conditions of the subject vessels. At every time step, the program
will check the yaw angle for each body, so that if the angle exceeds a certain amount, the
proper wave data file will be read and used for next time step.
The existing program is implemented to consider the connecting part of the vessel to
the mooring lines and risers as stiff linear rotational springs, or dampers only at the
position of starting points of mooring lines and risers. On the contrary, the ending points
of the mooring lines and risers are to be regarded as jointing to the sea floor with
assumed very huge stiffness of the seabed foundation. Some parts of the future
developed program will be modified so that the flexible connections at both ends of the
mooring lines and risers are available. The program will use the existing output format
of the previous program except extending the columns of output file for N 6 DOF
motions.
1.4.3 Comparative Studies
In this stage, the Buchner’s model(2001) and Choi’s model(2002) may be taken for
the comparative study about the results to be obtained from WINPOSTMULT. The
former is the multiple body system composed of the LNG FPSO tanker and the LNG
carrier. The two vessels are located each at very close proximity to the other in the open
sea. Buchner et al. (2001) has performed the calculation of hydrodynamic interaction
coefficients, wave load coefficients with the linear potential program using a lid
11
technique and the motion analysis of a multiplebody system using the above results as
input data. The results will be good for comparison with WINPOSTMULT’s. The latter
used the combining model of the FPSO and shuttle tanker located at close proximity
with the sidebyside arrangement and also at a distance with the tandem arrangement.
Choi et al. (2002) used the higherorder boundary element method not CPM(Constant
Panel Method) used in WAMIT.
Some examples are taken for verification of the hull/mooring/riser coupled dynamic
analyses of twobody system using the WINPOSTMULT program, for which two
identical SPARs, two identical FPSOs and also an FPSO and a shuttle tanker are selected
as the test models. The analysis results for those models are compared with the
simplified springmass models. For the environmental conditions, the 100year storm
condition in GOM and the sea condition in West Africa are taken.
12
CHAPTER II
DYNAMICS OF THE FLOATING PLATFORM
2.1 Introduction
In this chapter, the wave loads and dynamic responses of floating structures are
discussed. First, linear and secondorder wave theories are reviewed in the consideration
of the free surface boundary value problem, and then the boundary element method is
discussed as one of the solution schemes for the free surface boundary value problem,
and Morison’s equation and the wave drift damping are considered. Finally, the
multiplebody interaction of fluid is reviewed, and then the dynamic motions for single
body and multiple body systems of the floating structure are described, sequentially.
2.2 Formulation of Surface Wave
2.2.1 Boundary Value Problem (BVP) of Surface Wave
The fluid in the region surrounding the free surface boundary can be expressed as a
boundary value problem in the domain. The surface wave theory is derived from the
solution of the BVP with the free surface. The fluid motion can be expressed by the
Laplace equation of a velocity potential with the assumption of irrotational motion and
an incompressible fluid.
0 = ∇u (2.1)
or 0
2
2
2
2
2
2
2
=
∂
Φ ∂
+
∂
Φ ∂
+
∂
Φ ∂
= Φ ∇
z y x
(2.2)
13
where u is the velocity in x, y or z direction of fluid, so it becomes k j i
z y x ∂
Φ ∂
+
∂
Φ ∂
+
∂
Φ ∂
.
φ is the velocity potential. In order to solve the equation (2.2), the boundary condition
should be considered, specifically. The bottom boundary condition is to be considered.
In addition, there are two free surface conditions, which are the dynamic free surface
condition and the kinematic free surface condition. The bottom boundary condition is
given by the condition that the sea bed is impermeable:
0 =
∂
Φ ∂
z
at d z − = (2.3)
where d is the water depth. The kinematic condition is to represent that the fluid particle
on the free surface at any instance retains at one position of the free surface. The
equation of the kinematic free surface condition can be given by:
0 =
∂
Φ ∂
−
∂
∂
+
∂
∂
+
∂
∂
z y
v
x
u
t
η η η
at η − = z (2.4)
where ) , , ( t y x η is the displacement on the plane of the free surface to be varied in space
and time. The dynamic free surface condition defines that the pressure on the free
surface is constant as the equal value to the atmospheric pressure and normally the
atmospheric pressure is assumed to be zero. Thus, the condition can be described as
follows:
0 ) (
2
1
= + Φ ∇ ⋅ Φ ∇ +
∂
Φ ∂
gz
t
at η − = z (2.5)
where g is the gravitational acceleration. The most popular approach to solve the
equation (2.1) is known as the perturbation method under the assumption that the wave
14
amplitude is very small, which can give the approximated solution to satisfy partially the
free surface boundary conditions. In the method, the wave elevation (wave particle
displacement) and the velocity potential are to be taken as the power series forms a very
small nondimensional perturbation parameter. The linear wave and the second order or
higher order wave can be derived from the perturbation formula of the wave equation, to
be represented by the wave elevation and the velocity potential in terms of the
perturbation parameter.
2.2.2 Wave Theory
The perturbation formulation of the BVP with the first and secondorder
parameters can give the firstorder solution and the secondorder solution. The first
order solution leads the linear wave theory and the secondorder solution leads the
second order wave theory. The velocity potential is represented by the summation of all
perturbation terms and the wave elevation by summation of the perturbative wave
elevations. Finally, the total velocity potential and the wave elevation are written in the
following forms:
∑
Φ = Φ
) ( ) ( n n
ε (2.6)
) (
) (
n
n
∑
= η ε η (2.7)
The linear wave equations are obtained by solving the perturbation formulation
formed with the velocity potential and that with the wave elevation are obtained by:
The firstorder potential:
15
(
¸
(
¸
+ −
= Φ
− + ) sin cos ( ) 1 (
cosh
) ( cosh
Re
t ky kx i
e
kd
d z k igA
ω θ θ
ω
(2.8)
The firstorder wave elevation:
) sin cos cos(
) 1 (
t ky kx A ω θ θ η − + = (2.9)
where k is the wave number expressed by
L
π 2
when L is the wave length, ω is the
wave frequency, A is the wave amplitude, and θ is the incident wave angle. The
secondorder potential and the secondorder wave elevation are obtained by solving the
perturbation formulations formed with the secondorder potential and the secondorder
wave elevation are obtained as follows:
The secondorder potential:
(
¸
(
¸
+
= Φ
− + ) 2 sin 2 cos 2 (
4
2 ) 2 (
sinh
) ( 2 cosh
8
3
Re
t ky kx i
e
kd
d z k
A
ω θ θ
ω (2.10)
The secondorder wave elevation:
) 2 sin 2 cos 2 cos( ) 2 cosh 2 (
sinh
cosh
3
2 ) 2 (
t ky kx kd
kd
kd
k A ω θ θ η − + + = (2.11)
In the real sea, the wave is irregular and random. A fully developed wave is
normally modeled in terms of energy spectra combined with ensembles of wave trains
generated by random phases. Wellknown spectra in common usage, such as the
Pierson Moskowitz and the JONSWAP spectra, are established. The time series for a
given input amplitude spectrum ) (ω S is obtained by combining a reasonably large
number N of linear wave components with random phases:
(
¸
(
¸
= + − + =
∑ ∑
=
+ − +
=
N
i
t y k x k i
i
N
i
i i i i i
i i i i
e A t y k x k A t y x
1
) sin cos (
1
Re ) sin cos cos( ) , , (
ε ω θ θ
ε ω θ θ η (2.12)
16
where ω ω ∆ = ) ( 2
i i
S A is the wave amplitude of the i th wave, ω ∆ is the interval of
wave frequency, and
i
ε is the random phase angle. To avoid the increase of wave
components and to increase the computational efficiency for a long time simulation, the
following modified formula is used:
(
¸
(
¸
=
∑
=
+ ′ − +
N
j
t y k x k i
i
j j j j
e A t y x
1
) sin cos (
Re ) , , (
ε ω θ θ
η (2.13)
where
j j j
δω ω ω + = ′ and
j
δω is a random perturbation number uniformly determined
between
2
ω ∆
− and
2
ω ∆
. The total potential and the wave elevation are given by adding
every solution of each order equation, including the diffraction and the radiation.
2.2.3 Diffraction and Radiation Theory
The total velocity potential is decomposed into the incident potential
I
Φ , the
diffraction potential
D
Φ , and the radiation potential
R
Φ . By applying the perturbation
method, the total potential can be written by:
) (
) ( ) ( ) ( ) ( n
R
n
D
n
I
n
Φ + Φ + Φ = Φ
∑
ε (2.14)
The diffraction wave force and the radiation wave force have a significant effect on a
floating platform in deep water. The diffraction wave represents the scattered term from
the fixed body due to the presence of the incident wave. On the other hand, the radiation
wave means the wave to be propagated by the oscillating body in calm water. The forces
17
induced by them are evaluated by integration of the pressure around the surface of the
floating structure using the diffraction and the radiation potential, which can be obtained
by solving the BVPs of them.
2.2.3.1 FirstOrder Boundary Value Problem
By separation of variable for the firstorder component, the firstorder potential can
be written by:
{ }  
t i
R D I
R D I
e z y x z y x z y x
ω
φ φ φ
ε
−
⋅ + + =
Φ + Φ + Φ = Φ
) , , ( ) , , ( ) , , ( Re
) (
) 1 ( ) 1 ( ) 1 (
) 1 ( ) 1 ( ) 1 ( ) 1 (
(2.15)
By referring to the equation (2.8), the solution of incident wave velocity potential is
inferred as follows:
(
¸
(
¸
+ −
=
kd
d z k igA
I
cosh
) ( cosh
Re
) 1 (
ω
φ (2.16)
The BVPs for the firstorder potential of diffraction and radiation are defined as the
following formula:
0
) 1 (
,
2
= ∇
R D
φ in the fluid ( 0 < z ) (2.17)
0
) 1 (
,
2
=

.

\

∂
∂
+ −
R D
z
φ ω on the free surface ( 0 = z ) (2.18)
0
) 1 (
,
=
∂
∂
z
R D
φ
on the bottom ( d z − = ) (2.19)
¦
¦
)
¦
¦
`
¹
× + ⋅ − =
∂
∂
∂
∂
− =
∂
∂
) r α ξ ( n
1 ) 1 (
) 1 (
) 1 ( ) 1 (
) ( R
I D
i
n
n n
ω
φ
φ φ
on the body surface (2.20)
18
0 ) ( lim
) 1 (
,
= ±
∂
∂
∞ →
R D
ik r φ
ζ
ζ
at far field (2.21)
where r is the position vector on the body surface, R is the radial distance from the
origin (
2 2 2
y x r + = ), ) , , ( n
z y x
n n n = is the outward unit normal vector on the body
surface,
) 1 (
Ξ is the firstorder translational motion of the body, and
) 1 (
A is the firstorder
rotational motion of body. The
) 1 (
Ξ and
) 1 (
A can be expressed as follows:
 
t i
e
ω −
=
) 1 ( ) 1 (
ξ Re Ξ , ) , , ( Ξ
) 1 (
3
) 1 (
2
) 1 (
1
) 1 (
ξ ξ ξ = (2.22)
 
t i
e
ω −
=
) 1 ( ) 1 (
α Re A , ) , , ( α
) 1 (
3
) 1 (
2
) 1 (
1
) 1 (
α α α = (2.23)
where , , 3 2 1 means the x , y , z  axis, respectively. Thus,
) 1 (
3
) 1 (
2
) 1 (
1
, , ξ ξ ξ are defined as
the amplitude of surge, sway and heave motion, while
) 1 (
3
) 1 (
2
) 1 (
1
, , α α α are defined as the
amplitude of roll, pitch and yaw motion. The six degrees of freedom of the first order
motion are rewritten as:
¦
¹
¦
´
¦
=
=
=
−
6 , 5 , 4 for
3 , 2 , 1 for
) 1 (
3
) 1 (
j
j
j
j
j
α
ξ
ς (2.24)
The radiation potential can be decomposed as follows:
∑
=
=
6
1
) 1 ( ) 1 (
j
j j R
φ ς φ (2.25)
where
) 1 (
j
φ represents the velocity potential of rigid body motion with unit amplitude in
the j th mode when the incident wave does not exist. Equation (2.25) should satisfy the
boundary conditions of equation (2.18) to (2.21). The body boundary condition of
) 1 (
j
φ is
written as:
19
j
j
n i
n
ω
φ
− =
∂
∂
) 1 (
for 3 , 2 , 1 = j (2.26)
3
) 1 (
) n r (
−
× − =
∂
∂
j
j
i
n
ω
φ
for 6 , 5 , 4 = j (2.27)
These boundary conditions are valid on the body surface. The diffraction potential
problem, equation (2.17), can be solved numerically in consideration of the boundary
conditions (equation (2.18)(2.21)).
2.2.3.2 SecondOrder Boundary Value Problem
The secondorder boundary value problem is made by considering the interaction of
bichromatic incident waves of frequency
m
ω and
n
ω with a floating body. The Volterra
series method will be applied to solve the secondorder BVP. If the secondorder terms
are taken from the perturbation formulation (2.14) and the separation of variable is
applied, the secondorder potential is derived by:
{ } 
{ } 
t i
R D I
t i
R D I
R D I
e z y x z y x z y x
e z y x z y x z y x
t z y x
+
−
− + + +
− − − −
⋅ + + +
⋅ + + =
Φ + Φ + Φ = Φ
ω
ω
φ φ φ
φ φ φ
ε
) , , ( ) , , ( ) , , (
) , , ( ) , , ( ) , , ( Re
) ( ) , , , (
) 2 ( ) 2 ( ) 2 ( 2 ) 2 (
(2.28)
where
n m
ω ω ω − =
−
is the differencefrequency,
n m
ω ω ω + =
+
is the sum frequency,
−
φ is
the differencefrequency potential, and
+
φ is the sumfrequency potential. The
differencepotential and sumfrequency potential can be solved independently. The
governing equation (2.1) or (2.2) can be solved for each potential component of equation
(2.28) considering the boundary conditions, equation (2.3) to (2.5) as follows:
20
( )
x
cosh
) ( cosh
2
1 +
+
+
+ + +
+
+ =
ik
nm mn I
e
d k
d z k
γ γ φ (2.29)
( )
x *
cosh
) ( cosh
2
1 −
−
−
− − −
+
+ =
ik
nm mn I
e
d k
d z k
γ γ φ (2.30)
where
( ) ( )
d k k
d k d k k k d k k A igA
n m n m m m
m
n m
mn + + +
+
−
− + −
− =
tanh
tanh tanh 1 2 tanh 1
2
2 2
ν ω
γ (2.31)
and
( ) ( )
d k k
d k d k k k d k k A igA
n m n m m m
m
n m
mn − − −
−
−
+ − −
− =
tanh
tanh tanh 1 2 tanh 1
2
2 2 *
*
ν ω
γ (2.32)
and the asterisk represents a complex conjugate, and
±
ν and
±
k are defined respectively
by:
g
2
) (
±
±
=
ω
ν ,
n m
k k k ± =
±
(2.33)
The secondorder diffraction and radiation potential,
) 2 (
,R D
φ , deal with the second
interaction of plane bichromatic incident waves. The secondorder diffraction potential,
) 2 (
D
φ , contains the contributions of the secondorder incident potential and the firstorder
potential. The governing equation of the secondorder radiation potential is only
expressed by the outgoing waves propagated by the secondorder body motion. Thus, the
governing equation of the secondorder diffraction potential is defined by:
0
2
= ∇
±
D
φ in the quiescent fluid volume ( 0 < z ) (2.34)
( )
± ± ±
=
(
¸
(
¸
∂
∂
+ − Q
z
g
D
φ ω
2
on the free surface ( 0 = z ) (2.35)
21
0 =
∂
∂
±
z
D
φ
on the bottom ( d z − = ) (2.36)
±
± ±
+
∂
∂
− =
∂
∂
B
n n
I D
φ φ
on the body surface (2.37)
Boundary condition at far field (2.38)
where
±
Q are the sum and difference frequency components of the free surface force
and
±
B are the sum and difference frequency components of the body surface force. The
±
Q are symmetric and expressed as follows:
( )
+ + +
+ =
nm mn
q q Q
2
1
, ( )
*
2
1
− − −
+ =
nm mn
q q Q (2.39)
and,
+ +
− ∇ ∇ +


.

\

∂
∂
+
∂
∂
− − =
II n m n
m m
n
m
mn
q i
z
g
z g
i
q
) 1 ( ) 1 (
2
) 1 ( 2 ) 1 (
2 ) 1 (
φ φ ω
φ φ
ω φ
ω
(2.40)
− −
− ∇ ∇ +


.

\

∂
∂
+
∂
∂
− − =
II n m n
m m
n
m
mn
q i
z
g
z g
i
q
* ) 1 ( ) 1 (
2
) 1 ( 2 ) 1 (
2 * ) 1 (
φ φ ω
φ φ
ω φ
ω
(2.41)
The
±
B are also symmetric and expressed as follows:
( )
+ + +
+ =
nm mn
b b B
2
1
, ( )
*
2
1
− − −
+ =
nm mn
b b B (2.42)
and,
( )
) 1 ( ) 1 (
n
2
1
m n mn
b φ ς ∇ ∇ ⋅ ⋅ − =
+
(2.43)
( )
) 1 ( * ) 1 (
n
2
1
m n mn
b φ ς ∇ ∇ ⋅ ⋅ − =
−
(2.44)
22
The boundary condition (2.37) for the secondorder diffraction potential needs to be
applied to the decomposed diffraction potential into a homogenous term and a particular
solution term due to the complication. The homogeneous term of the secondorder
diffraction potential has the farfield propagating behavior, while the free surface force
±
Q are dominant in the particular equation term.
The governing equation and boundary conditions for the secondorder radiation
potential
±
R
φ are defined as the firstorder radiation BVP, since the boundary conditions
for the radiation potential do not contain any other potentials:
0
2
= ∇
±
R
φ in the fluid ( 0 < z ) (2.45)
0
2
=

.

\

∂
∂
+ −
±
R
z
φ ω on the free surface ( 0 = z ) (2.46)
0 =
∂
∂
±
z
R
φ
on the bottom ( d z − = ) (2.47)
) r α ξ ( n × + ⋅ − =
∂
∂
± ±
±
ω
φ
i
n
R
on the body surface (2.48)
0 ) ( lim = ±
∂
∂
±
∞ →
R
R
ik
R
R φ at far field (2.49)
where
±
ξ and
±
α are the second order translations and rotational motions of the body at
the sum and difference frequencies. Therefore, the secondorder radiation potential has
the same formula as the firstorder radiation potential.
23
2.3 Hydrodynamic Forces
2.3.1 The FirstOrder Hydrodynamic Forces and Moments
If all of the potentials are solved, the firstorder force and moment can be obtained
from the integration over the whole surface pressure on the body. The pressure on the
body surface (
B
Ω ∂ ) is obtained from the potential as follows:

.

\

+
∂
Φ ∂
− = gz
t
P
) 1 (
) 1 (
ρ (2.50)
where ρ is the fluid density. The six components of forces and moments are calculated
as follows:
(
(
¸
(
¸
+ −
(
(
¸
(
¸
−
− =
∫∫ ∫∫
∫∫
Ω ∂
−
Ω ∂
−
Ω ∂
B B
B
dS n Ae i dS n e i
dS zn g t F
j D I
t i
j j
t i
j
j j
) ( Re Re
) (
) 1 (
φ φ ω ρ φ ως ρ
ρ
ω ω
, 6 ... 1 = j (2.51)
where,
¹
´
¦
× =
=
n r ) 6 , 5 , 4 (
) , , (
n
3 2 1
n n n
n n n
6 , 5 , 4 for
3 , 2 , 1 for
=
=
j
j
(2.52)
In the above equation (2.51), the three terms represent the different contributions to the
body forces and moments. The first term (
) 1 (
F
S
) is the hydrostatic restoring force, the
second term (
) 1 (
F
R
) is the force term due to the radiation potential, and the last term (
) 1 (
F
E
)
is the exciting forces generated by the incident and the diffraction potentials. The
hydrostatic restoring forces are defined as the multiplication of the restoring stiffness
and the motion responses, and the components of restoring stiffness are defined as the
24
following surfaceintegral form over the wetted body surface at the mean position
(
B
Ω ∂ ):
 { }
) (
S
1 ) 1 (
ς K F − = (2.53)
where
cg b
cg b
cg b
cg b
f wp
f wp
wp
mgy y g K
mgz z g dS n x g K
mgx x g K
dS xyn g K
mgz z g dS n y g K
x gA dS xn g K
y gA dS yn g K
gA dS n g K
B
B
B
B
B
B
+ ∀ − =
− ∀ + =
+ ∀ − =
− =
− ∀ + =
= − =
= =
= =
∫∫
∫∫
∫∫
∫∫
∫∫
∫∫
∂
∂
∂
∂
∂
∂
ρ
ρ ρ
ρ
ρ
ρ ρ
ρ ρ
ρ ρ
ρ ρ
56
3
2
55
46
3 45
3
2
44
3 35
3 34
3 33
Ω
Ω
Ω
Ω
Ω
Ω
(2.54)
where
nm mn
K K = for all m and n ,
wp
A is the water plane area,
f
x and
f
y are the
distances from the center of the water plane area to the center of gravity in xdirection
and in ydirection, respectively, ∀ is the buoyancy of the body, ) z y x
cg cg cg
, , ( is the
center of gravity, and ) z y x
b b b
, , ( is the center of buoyancy of the body.
The hydrostatic restoring stiffness will be used for the motion analysis of the
floating body. The radiation potential forces and moments corresponding to the second
term of the equation (2.51) can be rewritten as the form:
25
( ) ( )  
t i
j
a ) ( ) ( a
j
j t i
j R
e i
dS
n
e
B
ω
ω
ς ω ω
φ
φ
ς ρ
−
Ω ∂
−
− = + =
(
(
¸
(
¸
∂
∂
− =
∫∫
C M  Re ς C ς M Re
Re F
2 1 1
) 1 (
& & &
(2.55)
where
a
M is the added mass coefficients, C is the radiation damping coefficients, and
t i
e
ω
ς
−
= ς are the body motions of six degrees of freedom. They can be represented as
follows:
(
(
¸
(
¸
∂
∂
=
∫∫
Ω ∂
B
dS
n
j
j a
φ
φ
ρ Re M (2.56)
(
(
¸
(
¸
∂
∂
=
∫∫
Ω ∂
B
dS
n
j
j
φ
φ
ρ Im C (2.57)
They are symmetric and dependent on the frequency of the body motion.
The last term of the equation (2.51) corresponds to the linear wave exciting force,
and it can be rewritten as the form:
( )
(
(
¸
(
¸
∂
∂
+ − =
∫∫
Ω ∂
−
B
dS
n
Ae
j
D I
t i
E
φ
φ φ ρ
ω
Re F
) 1 (
(2.58)
Therefore, the equation of motion is formed as:
( )
) 1 ( ) 1 ( ) 1 ( ) 1 ( 1
F ς C ς M  Kς F F F ς M
E
a
E R S
) (
+ + − = + + = & & & & & (2.59)
where M is the mass matrix of the body, which is described as:
(
(
(
(
(
(
(
(
¸
(
¸
=
0 
 0
 0
0  0 0
0  0 0
0 0 0
M
33 32 31
23 22 21
13 12 11
I I I mx my
I I I mx mz
I I I my mz
mx my m
mx mz m
my mz m
cg cg
cg cg
cg cg
cg cg
cg cg
cg cg
(2.60)
26
where V represents the body volume,
∫∫∫
∀
= dV m
B
ρ is the body mass,
( )
∫∫∫
∀
− ⋅ = dV x x I
n m mn B mn
δ ρ x x is the moment of inertia,
B
ρ is the density of the body,
and
mn
δ is the Kronecker delta function.
2.3.2 The SecondOrder Hydrodynamic Forces and Moments
The secondorder wave forces and moments on the body can be obtained by direct
integration of the hydrodynamic pressure over the wetted surface of the body at the
instantaneous time step. The secondorder pressure is defined as:
( )
2
) 1 (
) 2 (
) 2 (
2
1
Φ ∇ −
∂
Φ ∂
− = ρ ρ
t
P (2.61)
In consideration of the bichromatic wave, the secondorder pressure is modified as:
 
∑∑
= =
− − − +
− +
+ =
2
1
2
1
* ) 2 (
Re
m n
t i
mn n m
t i
mn n m
e p A A e p A A P
ω ω
(2.62)
where
±
mn
p are defined as the sum and difference frequency quadratic transfer functions
for the secondorder pressure. The secondorder forces and moments are defined as:
) 2 ( ) 2 ( ) 2 ( ) 2 (
F F F F
E R S
+ + = (2.63)
where
) 2 (
F
S
represents the secondorder hydrostatic force,
) 2 ( ) 2 ( ) 2 (
F F F
q p E
+ = is the second
order wave exciting force, and ,
) 2 (
F
R
is the radiation potential force. The components of
) 2 (
F
E
are defined as
) 2 ( ) 2 ( ) 2 (
F F F
D I p
+ = , which denotes the incident and diffraction potential
27
forces, and
) 2 (
F
q
denotes the quadratic product of the firstorder forces. The component
forces are derived in the integration forms of potentials as follows:
( )k x y gA
y f x f z wp S
) 2 ( ) 2 ( ) 2 ( ) 2 (
F α α ξ ρ − + = (2.64)
dS n
t
B
R
R
∫∫
Ω ∂
∂
Φ ∂
=
) 2 (
) 2 (
F ρ (2.65)
dS n
t
B
D I
D I
∫∫
Ω ∂
∂
Φ ∂
=
) 2 (
, ) 2 (
,
F ρ (2.66)
 
∑∑
= =
− − − +
− +
+ =
2
1
2
1
* 2
Re F
m n
t i
mn n m
t i
mn n m
) (
E
e f A A e f A A
ω ω
(2.67)
where
±
mn
f denote the quadratic transfer function (QTF) of the sum and difference
frequency exciting force. QTF is obtained by the addition of
±
mn
h and
±
mn
g , where
±
mn
h are
the contribution of firstorder quadratic transfer function and
±
mn
g are the summation of
the quadratic transfer function of the sum and difference frequency exciting force due to
the incident potential and the diffraction potential. Each component of the QTF is
defined as:
± ± ±
+ =
mn mn mn
g h f (2.68)
( )
n m
L
n m
n m
n m mn
A A dL
g
dS h
B W
/ N
4
n
4
) 1 ( ) 1 ( ) 1 ( ) 1 (
(
(
¸
(
¸
− ∇ ⋅ ∇ − =
∫∫ ∫
Ω ∂
+
φ φ
ω ρω
φ φ
ρ
(2.69)
( )
* * ) 1 ( ) 1 ( * ) 1 ( ) 1 (
/ N
4
n
4
n m
L
n m
n m
n m mn
A A dL
g
dS h
B W
(
(
¸
(
¸
− ∇ ⋅ ∇ − =
∫∫ ∫
Ω ∂
−
φ φ
ω ρω
φ φ
ρ
(2.70)
( ) ( )
*
, / n
n m n m D I mn
A A A A dS i g
B
(
(
¸
(
¸
+ =
∫∫
Ω ∂
± ± ± ±
φ φ ω ρ (2.71)
28
where ( )
2
1 n/ N
z
n − = , and k is the unit vector in the zdirection.
2.4 Multiple Body Interaction of Fluid
The boundary value problem of the multiple body interaction of fluid is explained
that the effects of the incident potential and the scattered potential on the main body and
the adjacent body are investigated. For the single body system, the radiation potential
and the incident potential are obtained as described in the above sections. The diffraction
problem for the isolated body can be defined by the incident potential as follows:
n n
I
I
∂
∂
− =
∂
∂ φ φ
7
on
I
S (2.72)
n n
I
II
∂
∂
− =
∂
∂ φ φ
7
on
II
S (2.73)
where
II I
S S , denotes the wetted surface of the isolated body I and II , respectively,
II I
7 7
, φ φ denotes the scattered potential to the isolated body I and II , respectively, and
I
φ represents the incident wave potential of the isolated body. The radiation potential for
the isolated body can be decomposed in the similar manner to the equation (2.25) as
follows:
∑
=
=
6
1 j
I
j j
I
R
φ ς φ (2.74)
∑
=
=
6
1 j
II
j j
II
R
φ ς φ (2.75)
The radiation problem for the isolated body I and II can be given by:
29
I
j
I
j
n
n
=
∂
∂φ
on
I
S ) 6 ..., 2 , 1 ( = j (2.76)
II
j
II
j
n
n
=
∂
∂φ
on
II
S ) 6 ..., 2 , 1 ( = j (2.77)
where
II
j
I
j
φ φ , denotes the decomposed radiation potential components for the isolated
body I and II , respectively, and
II I
j
n
,
is a unit normal vector for the six degree of
freedom for the isolated body I and II , respectively. In equation (2.76) and (2.77),
II I
j
n
,
is given by:
¹
´
¦
× =
=
n r
~
) 6 , 5 , 4 (
) , , (
n
3 2 1
I,II I,II
I,II
I,II
n n n
n n n
6 , 5 , 4 for
3 , 2 , 1 for
=
=
j
j
(2.78)
where r
~
denotes the relative distance from the origin to each body center.
The boundaryvalue equation and the boundary condition for each body of the
interaction problem is defined in the form of the radiation/scatter potential and the
derivative as follows:
Interaction problem – radiation/scatter from I near II:
n n
I
j
I
j
∂
∂
− =
∂
∂ φ φ
ˆ
on
I
S ) 7 ..., 2 , 1 ( = j (2.79)
0
ˆ
=
∂
∂
n
I
j
φ
on
II
S ) 7 ..., 2 , 1 ( = j (2.80)
Interaction problem – radiation/scatter from II near I:
n n
II
j
II
j
∂
∂
− =
∂
∂ φ φ
ˆ
on
II
S ) 7 ..., 2 , 1 ( = j (2.81)
30
0
ˆ
=
∂
∂
n
II
j
φ
on
I
S ) 7 ..., 2 , 1 ( = j (2.82)
where
II I
j
,
ˆ
φ denotes the interaction potential affected by radiation/scatter potential from
the body I to the body II , and vice versa, respectively. The potential when 7 = j
means the scatter term. If the firstorder radiation/scatter potential is used when the
above BVP is solved, the resultant potential would be the firstorder interaction potential,
while the secondorder radiation/scatter potential leads the secondorder interaction
potential.
2.5 Boundary Element Method
The boundary element method is proper for solving the boundary value problem of
the fluid potential around the floating body since there is no analytic solution except for
some special geometric bodies. BEM is generally called the inverse formulation, since
the solution to satisfy all of the boundary conditions, except the body boundary
condition for the firstorder potential and the body boundary condition and the free
surface condition for the secondorder potential, is used as a weighting function. It is
also based on GreenLagrange’s Identity given by:
( )
∫∫ ∫∫∫
Ω ∂ Ω

.

\

∂
∂
−
∂
∂
= Ω ∇ − ∇ dS
n
G
n
G d G G φ
φ
φ φ
2 2
(2.83)
where G is the Green function to satisfy all of the boundary conditions, Ω denotes the
fluid domain, and Ω ∂ denotes the boundary of the domain. φ is the exact solution of
potential and G satisfies the following equation:
31
) G
2
x ( δ = ∇ (2.84)
where δ is Dirac delta function, and x means the position coordinates. Since φ and G
satisfy all of the boundary conditions except the body or the free surface, the right hand
side of the equation (2.83) becomes:
∫∫ ∫∫
Ω ∂ Ω ∂

.

\

∂
∂
−
∂
∂
+ 
.

\

∂
∂
−
∂
∂
=
F B
dS
n
G
n
G dS
n
G
n
G ) c φ
φ
φ
φ
φ ) x ( x ( (2.85)
where ) x ( c means a shape factor depending on the body geometry,
B
Ω ∂ represents the
body boundary, and
F
Ω ∂ is the free surface boundary. If the body geometry has a
smooth surface, ) x ( c becomes π 2 . The equation (2.85) is a fundamental equation called
the Inverse Formulation.
If the formulation is applied to the firstorder diffraction potential problem for the
smooth surface of body, the equation (2.85) becomes a second kind of Fredholm integral
equation such as:
∫∫ ∫∫
Ω ∂ Ω ∂


.

\

∂
∂
− =
∂
∂
+
B B
dS
n
G dS
n
G
I
D D
) ξ (
) ξ (
) x ; ξ ( ) ξ (
) x ; ξ (
) ξ ( ) x ( 2
) 1 (
) 1 ( ) 1 (
φ
φ πφ (2.86)
where ξ denotes the source point coordinates. If it is applied to the firstorder radiation
potential problem, it becomes as:
( )
¦
¦
¹
¦
¦
´
¦
= ×
=
=
∂
∂
+
∫∫
∫∫
∫∫
Ω ∂
−
Ω ∂
Ω ∂ 6 , 5 , 4 for ) ξ ( n r ) x ; ξ (
3 , 2 , 1 for ) ξ ( ) x ; ξ (
) ξ (
) x ; ξ (
) ξ ( ) x ( 2
3
) 1 ( ) 1 (
k dS G
k dS n G
dS
n
G
B
B
B
k
k
R R
φ πφ (2.87)
If the formulation is applied to the secondorder diffraction potential problem for
the flat surface of body, it becomes as:
32
∫∫ ∫∫ ∫∫
Ω ∂
± ±
Ω ∂
±
± ±
Ω ∂
±
± ±
+


.

\

∂
∂
− =
∂
∂
+
F B B
dS G Q
g
dS
n
B G dS
n
G
I
D D
1
2
φ
φ πφ (2.88)
If it is applied to the secondorder radiation potential problem for a far field, it becomes
as:
( )
( ) ( )
¦
¦
¹
¦
¦
´
¦
= + ×
= +
=
∂
∂
+
∫∫ ∫∫
∫∫ ∫∫
∫∫
Ω ∂
±
∞ →
± ±
Ω ∂
−
±
Ω ∂
±
∞ →
± ±
Ω ∂
±
Ω ∂
±
± ±
6 , 5 , 4 for lim n r
3 , 2 , 1 for lim
2
2
3
2
k dS R ik G dS G
k dS R ik G dS n G
dS
n
G
F B
F B
B
R
R
R k
R
R
R k
R R
φ φ ω
φ φ ω
φ πφ
m
m
(2.89)
In this formulation, it is noted that the integration term for the free surface remains. If
the Constant Panel Method (CPM) of BEM is taken, the simplest form is shown as:
∫∫ ∫∫
Ω ∂ Ω ∂
∂
∂
=
∂
∂
+
B B
dS
n
G dS
n
G
) ξ (
) ξ (
) ξ (
) x , ξ ( ) ξ (
) ξ (
) x , ξ (
) ξ ( ) x ( 2
φ
φ πφ (2.90)
If the equation is applied for the discretized model, it is modified as:
∑
=
=
L
j
j j
x x N
1
2 1
) , ( ) ξ ( φ φ , points) ion Interpolat of No. ( ,..., 2 , 1 = L (2.91)
∑ ∑
= =

.

\

∂
∂
=
M
j
j
ij
M
j
j ij
n
G H
1 1
φ
φ , pannels) of No ( ,..., 2 , 1 = M (2.92)
where
j
N is the shape function, ) , (
2 1
x x is the local coordinate, and
ij
H and
ij
G are as
follows:
∫∫
≠ Ω ∂
∂
∂
+ =
i j
ij ij
B
dS
n
G
H
,
) ξ (
) ξ (
) x , ξ (
4
1
2
1
π
δ (2.93)
∫∫
≠ Ω ∂
=
i j
ij
B
dS G G
,
) ξ ( ) x , ξ (
4
1
π
(2.94)
33
In the equations of (2.92) and (2.94),
n ∂
∂φ
is given by the equation (2.20) and
) ξ (
) x ξ, (
), x ξ, (
n
G
G
∂
∂
are known as the exact forms. Thus, the equation (2.92) can be solved
for the whole panels.
For the BEM program, the WAMIT (Lee et al, 1991) of CPM is well known in this
field. the WAMIT can be applied to the firstorder and secondorder diffraction/radiation
potential problem. In this study, the WAMIT will be taken for solving the fluid
interaction problem of the multiplebody system.
2.6 Motions of the Floating Platform
2.6.1 Wave Loads
The linear wave forces are calculated in the frequency domain, and the second
order sum and difference frequency wave loads are computed by considering the
bichromatic wave interactions. The real sea is made of random waves, so that it is
essential to make the random waves for applying the external wave loads to the floating
body.
The linear and the secondorder hydrodynamic forces can be rewritten as the form
of a twoterm Volterra series in time domain:
∫ ∫ ∫
∞
∞ −
∞
∞ −
∞
∞ −
− − + − = +
2 1 2 1 2 1 2 1
) 2 ( ) 1 (
) ( ) ( ) , ( ) ( ) ( ) ( ) ( τ τ τ η τ η τ τ τ τ η τ d d t t h d t h t F t F (2.95)
where ) (
1
τ h is the linear impulse response function, and ) , (
2 1 2
τ τ h is the quadratic
impulse response function, i.e., the secondorder exciting force at time t for the two
34
different unit amplitude inputs at time
1
τ and
2
τ . ) (t η is the ambient wave free surface
elevation at a reference position. Since ) (t η , ) (
1
τ h and ) , (
2 1 2
τ τ h can be expressed in
the functions of frequency, the unidirectional wave exciting forces induced by the
incident potential and the diffraction potential to have the similar form of the equation
(2.95) can be rewritten in the form of the summation of the frequency components as
follows:
(
¸
(
¸
=
∑
=
N
j
t i
j L j I
e q A t F
1
) 1 (
) ( Re ) (
ω
ω (2.96)
(
¸
(
¸
+ − =
∑∑ ∑∑
= = = =
+ −
N
j
N
k
t i
k j S k j
N
j
N
k
t i
k j D k j I
e q A A e q A A t F
1 1 1 1
* ) 2 (
) , ( ) , ( Re ) (
ω ω
ω ω ω ω (2.97)
where ) (
j L
q ω represents the linear force transfer function (LTF), and ) , (
k j D
q ω ω − and
) , (
k j S
q ω ω are the difference and the sum frequency quadratic transfer functions (QTF),
respectively. Using the Fourier transform, the equation (2.96) and (2.97) can be easily
changed into the energy spectra given by:
2
) 1 (
) ( ) ( ) ( ω ω ω
η L F
q S S = (2.98)
∫
∞
−
− − =
0
2
) ( ) ( ) ( ) , ( 8 ) ( µ µ ω µ µ ω µ ω
η η η
dS S S q S
D F
(2.99)
∫
− + − + =
+
2 /
0
2
) ( )
2
( )
2
( )
2
,
2
( 8 ) (
ω
η η η
µ µ
ω
µ
ω
µ
ω
µ
ω
ω dS S S q S
S F
(2.100)
35
where ) (ω
η
S is the wave spectrum, ) (
) 1 (
ω
F
S is the linear wave force spectrum, and
) (ω
−
F
S and ) (ω
+
F
S are the secondorder sum and differencefrequency wave force
spectrum, respectively.
The first and secondorder radiation potential forces are calculated by the
following formula:
∫ ∫
∞ −
∞
− −


.

\

− =
t
a
R
d (τ ς t R (t) ς tdt t R M t F τ τ ω ω ) ) ( cos ) ( ) ( ) (
0
& & & (2.101)
where ) (ω
a
M is the added mass coefficient as defined in the equation (2.55) at
frequency ω , and ) (t R is called a retardation function as defined below:
∫
∞
=
0
sin
) (
2
) ( ω
ω
ω
ω
π
d
t
C t R (2.102)
where ) (ω C is the radiation damping coefficient in the equation (2.56) at frequency ω .
The total wave forces and moments can be obtained by summation of the equation (2.96),
(2.97) and (2.101) as the same form as the summation of the equation (2.59) and (2.63)
as follows:
R c I T
F F F F
~
+ + = (2.103)
where
) 2 ( ) 1 (
F F F
T
+ = is the total wave exciting force,
) 2 ( ) 1 (
I I I
F F F + = is the sum of the
equation (2.96) and (2.97),
c
F is the last term of the right hand side of the equation
(2.101), and
R
F
~
is the first term of the equation (2.101).
36
2.6.2 Morison’s Equation
For the slender cylindrical floating structure, the inertia and added mass effect and
the damping effect of the drag force on the slow drift motion can be evaluated by using
Morison’s equation. Morison et al. (1950) proposed that the total force is the sum of drag
force and inertia force as follows:
( )
n n n n S D n a n m m
u u D C
2
1
V C u V C F ς ς ρ ς ρ ρ & & & & & − − + − = (2.104)
where
m
F denotes Morison’s force,
4
2
D
V
π
= is the volume per unit length of the
structure, D is the diameter of the slender body,
a m
C C + =1 is the inertia coefficient,
a
C is the added mass coefficient,
D
C is the drag coefficient,
S
D is the breadth or
diameter of the structure,
n
u& and
n
u are the acceleration and the velocity of the fluid
normal to the body, respectively, and
n
ς& & and
n
ς& are the acceleration and the velocity of
the body, respectively. In the above equation, the first term is called FroudeKrylov
force, the second term the added mass effect, and the last term the drag force. The drag
force on the floating structure cannot be neglected, because the slenderness ratio of the
structure (the ratio of breadth or diameter to the length of the structure) is small
compared to the wavelength so that the viscous effect cannot be negligible. The derived
force by the equation (2.104) is added to the wave forces of the equation (2.103) to get
the total force.
37
2.6.3 Single Body Motion
The equilibrium equation using Newton’s second law called the momentum
equation for the floating structure can be given as:
f M =
2
2
x
dt
d
cg
(2.105)
m I I = × + ) ( ϕ ϕ
ϕ
dt
d
(2.106)
where M is the mass of the floating structure,
cg
x is the coordinates of the center of
gravity of the floating body, I is the moment of inertia, and ϕ is the angular velocity, f
and m are the external force and moment. The second term of the lefthand side of the
equation (2.104) and the relative angular motion of the body to the wave motion are
nonlinear. If the rotation is assumed to be small, the equation (2.106) becomes a linear
equation as follows:
) (t F ς M = & & (2.107)
where ς& & is the normal acceleration of body motion, M is the 6 6 × body mass matrix to
be the same as equation (2.59) and (t) F is the external force vector. In the time domain,
the above equation is expanded as:
  ) , ( ) , ( ) ( Kς ς ) ( M t t t
m c I
a
ς ς & & & & F F F M + + = + ∞ + (2.108)
where ) (∞
a
M is a constant, equivalent added mass of the body at the infinite frequency
and can be expressed by :
∫
∞
− = ∞
0
cos ) ( ) ( M ) ( tdt t R
a a
ω ω M (2.109)
38
where ) (ω
a
M is the same as defined in the equation (2.56).
c
F is the same as the second
term of the equation (2.103) and defined as:
∫
∞ −
− − =
t
c
d t t τ τ ς ς ) ( R ) , ( & & F (2.110)
I
F is the same as the equation (2.96) and (2.97), and
m
F is the force by Morison’s
equation such as the equation (2.104). ς& is the normal velocity of the body.
2.6.4 Multiple Body Motion
For the multiple body system, the number of the degrees of freedom of the mass
matrix, the body motion vector and the force vector in the equation (2.106) are changed
to N N 6 6 × , N 6 and N 6 , N of which is the number of bodies. And also in the total
system equation (2.106), the matrix sizes are extended accordingly. For the formulation
of motion, the local coordinate system is used for each body. After forming the equation
of motion for each body, the coordinate transformation is needed. Finally, the total
equation of motion in the global coordinate system is assembled for the combined
system. The hydrodynamic coefficients are premade in consideration of the fluid
interaction terms influenced on each body by using WAMIT. The hydrodynamic
coefficients are solved in the sequence as follows:
1) The radiation/diffraction problem for each body in isolation
2) The interaction problem resulting from radiation/scatter from body I in the
presence of body II, and radiation/scatter from body II in the presence of body I.
39
Where body I and II represent one pair of bodies which interact hydrodynamically. If
there are several bodies, the twobody problem should be addressed for each unique pair
of bodies. The boundaryvalue problem is formed differently due to the different
kinematic boundary condition on the immersed surface of bodies, but other boundary
conditions for the bodies are the same as those in the isolated body.
The boundary–value problem of fluid interaction is solved using the equation
(2.81) and (2.82) in the section 2.4 in the form of an excitation force coefficient as
follows:
∫
− =
I
S
j
I I I
j
dS n a C
7
,
ˆ
φ , ( 6 , , 2 , 1 L = j ) (2.111)
∫
− =
I
S
j
II II II
j
dS n a C
7
,
ˆ
φ , ( 6 , , 2 , 1 L = j ) (2.112)
∫
+ − =
I
S
j
II II II I
j
dS n a C )
ˆ
(
7 7
,
φ φ , ( 6 , , 2 , 1 L = j ) (2.113)
∫
+ − =
II
S
j
I I I II
j
dS n a C )
ˆ
(
7 7
,
φ φ , ( 6 , , 2 , 1 L = j ) (2.114)
where the superscript I and II represent the body I and II. If the coefficients are written
in the form of equation (2.109), the hydrodynamic coefficients are obtained by:
6 , , 2 , 1 , ,
ˆ
) (
,
L = − = ∞
∫
j i dS n M
I
S
i
I
j
I I
a
φ (2.115)
6 , , 2 , 1 , ,
ˆ
) (
,
L = − = ∞
∫
j i dS n M
II
S
i
II
j
II II
a
φ (2.116)
6 , , 2 , 1 , , )
ˆ
( ) (
,
L = + − = ∞
∫
j i dS n M
I
S
i
II
j
II
j
II I
a
φ φ (2.117)
6 , , 2 , 1 , , )
ˆ
( ) (
,
L = + − = ∞
∫
j i dS n M
II
S
i
I
j
I
j
I II
a
φ φ (2.118)
40
Then, for the twobody problem, the equation (2.113) to equation (2.116) are replaced
for the equation (2.107), and the replaced equations mean the matrix ) (∞
a
M in the
equation (2.106). In the equation (2.106), the other matrices contain the terms for two
bodies. Thus,
(
¸
(
¸
=
II
I
M
M
M
0
0
, (2.119)
(
(
¸
(
¸
=
II II
I I
, I II,
II I, ,
K K
K K
K , (2.120)
(
(
¸
(
¸
=
II
I
I
I
I
F
F
F , (2.121)
(
(
¸
(
¸
=
II
C
I
C
C
F
F
F , (2.122)
(
(
¸
(
¸
=
II
m
I
m
m
F
F
F , (2.223)
where the superscript I and II represent the body I and II. The total equation of
motion of the system has the same form of equation (2.106), but for the Nbody with 6
DOF for each body, the matrices are of the size of N N 6 6 × .
2.6.5 Time Domain Solution of the Platform Motions
Since the system contains the nonlinear effect, the numerical scheme of the
iterative procedure in the time domain is commonly used. The equation of motion in
time domain for a singlebody system and/or a twobody system is expressed as the
41
equation (2.108) with the equation (2.109) and (2.110). For the numerical integration in
the time domain, there are several kinds of implicit methods developed, such as the
NewmarkBeta method, RungeKuta method and the AdamsMoulton method (or mid
point method). The last is used for the purpose of the guarantee of the secondorder
accuracy. Another reason to use it is that the method has the merit to solve together the
coupled equations of the platform motion and mooring line motions at each time step.
Furthermore, the AdamsBashforth method is also used for the time integration of the
nonlinear force.
In the first step, the equation (2.108) is derated to the first order differential
equation:
ς ς ς η K F F F M − + + = ) , ( ) , ( ) (
~
t t t
m c I
& (2.124)
ς η & = (2.125)
where ) (
~
∞ + =
a
M M M denotes the virtual mass matrix. If the integration from time step
) (n
t to
) 1 ( + n
t is performed, the following equation is obtained:
∫ ∫
+ +
− + + + =
+
) 1 (
) (
) 1 (
) (
) (
~ ~
) ( ) 1 (
n
n
n
n
t
t
t
t
m c I
n n
dt dt ς η η K F F F M M (2.126)
∫
+
+ =
+
) 1 (
) (
) ( ) 1 (
n
n
t
t
n n
dt η ς ς (2.127)
If the AdamMoulton method is applied to the equation (2.126) and (2.127), the
following equation is obtained after the resultant equation rearranged:
) (
2
) (
2
~ ~
) ( ) 1 ( ) ( ) ( ) ( ) 1 ( ) 1 ( ) 1 ( ) ( ) 1 ( n n n
m
n
c
n
I
n
m
n
c
n
I
n n
t t
ς ς η η +
∆
− + + + + +
∆
+ =
+ + + + +
K F F F F F F M M
(2.228)
42
) ( ) ( ) 1 ( ) 1 (
) (
2
n n n n
t
η ς ς η − −
∆
=
+ +
(2.229)
The equations (2.228) and (2.229) are the combination of two linear algebraic equations
with the unknowns of
) 1 ( + n
η and
) 1 ( + n
ς . To solve the above equations, the assumption of
the first terms is needed. It means that the time integration may have an error term due
to the arbitrary adoption of the first term. For the evaluation of the first terms of time
varying unknowns to avoid the abovementioned problem, the AdamsBashforth scheme
is used. Thus, the time integration of the nonlinear term of radiation damping force is as
follows:
) 3 (
2
) 1 ( ) (
) 1 (
) (
−
−
∆
=
∫
+
n
c
n
c
t
t
c
t
dt
n
n
F F F (2.230)
0 for
) 0 (
) 1 (
) (
= ∆ =
∫
+
n t dt
c
t
t
c
n
n
F F (2.231)
In the same sense, the time integration of the nonlinear term of drag force in Morison’s
formulation is as follows:
) 3 (
2
) 1 ( ) (
) 1 (
) (
−
−
∆
=
∫
+
n
m
n
m
t
t
m
t
dt
n
n
F F F (2.232)
0 for
) 0 (
) 1 (
) (
= ∆ =
∫
+
n t dt
m
t
t
m
n
n
F F (2.233)
Eventually, the equation (2.124) and (2.125) are derived as follows:
0
) (
) 1 ( ) (
) 1 ( ) ( ) ( ) 1 (
) (
2
2 2 ) 3 (
) 3 ( ) (
~ 4 ~ 4
F K F F
F F F F M K M
+ − − +
− + + +
∆
= ∆
(
¸
(
¸
+
∆
−
− +
n
n
m
n
m
n
c
n
c
n
I
n
I
n
t t
ς
η ς
(2.234)
) ( ) 1 ( n n
ς ς ς − = ∆
+
(2.235)
43
where
0
F represents the net buoyancy force for balancing the system. Firstly, the
equation (2.234) is solved for the unknown of ς ∆ . Then,
) 1 ( + n
η and
) 1 ( + n
ς can be
obtained from the equation (2.229) and (2.235). To obtain the stability and the accuracy
of the solution, the time interval of t ∆ may be small enough to solve the mooring line
dynamics, since the mooring line shows a stronger nonlinear behavior than the platform
movement.
44
CHAPTER III
DYNAMICS OF MOORING LINES AND RISERS
3.1 Introduction
In this chapter, the theory and the numerical method for the dynamic analysis of
the mooring lines and risers are explained.
The platform is considered as a singlepoint floating system when the behavior of
the mooring line is taken into account. To maintain the sea keeping, several types and
different materials of mooring lines have been installed. A steel wire rope with chains at
both ends has been used for SPAR platform in deep water. There are also taut vertical
mooring lines and tethers made of several vertical steel pipes, usually intended to be
installed in the TLP. Synthetic mooring lines made of polyester are now considered as a
more efficient solution. For the sea keeping for FPSOs, the attempt is to use synthetic
mooring lines for fixing those structures in very deep water(over 6,000 ft). Sometimes
FPSOs are needed to construct the mooring lines and risers, and to be connected to the
TLP, the Single Point Mooring System (SPM) and the shuttle tankers with hawsers or
fluid transfer lines(FTLs). The multiple body interaction problems are caused by those
kinds of system arrangements. The interaction problem between the floating platforms is
the matter to be presolved before planning the deep water installation of FPSOs.
For exporting and importing gas and water, and for the production of gas, risers
are taken into account. The main purpose of risers is not to fix the floating structure in
45
the station keeping position, but to act the roles. It tends that the steel catenary risers are
used more and more because they are inexpensive. Both mooring lines and risers are the
same from the viewpoint of the installation, in that they don’t have bending stiffness and
are the slender members. The restoring forces of both lines result from gravitational
forces, geometries and line tensions. But, the bending stiffness of the tendon and the
riser in a TLP has a restoring effect. In the mooring lines and risers, the geometric
nonlinearity is dominant on the line behavior.
The analysis of line dynamics is developed on the basis of the theories of behaviors
of slender structures. The static position and the line tension are obtained by using the
catenary equation. In the catenary equation, no hydrodynamic force on the line is
considered. For the consideration of the hydrodynamic force on the line, the tensioned
string theory is used, but in the theory the structural strain and stress contribution are
missing. The strain and the stress of a structure with geometric nonlinearity can be
solved with the beam theory using the updated Lagrangian approach. Therefore, in the
program, the tensioned string theory using the string modeled as the beam elements is
adopted for its rigorous analysis. It is called the elastic rod theory, and the formula was
derived by Nordgen(1974) and Garret(1982). The finite difference method was applied
to this problem by the former. Here the FEM technique suggested by the latter is taken.
Garret proved line dynamics could be solved more accurately by the FEM.
In this study, a threedimensional elastic rod theory containing line stretching and
bending behavior is adopted. The advantage of the elastic rod theory is that the
governing equation, including the geometric nonlinearity, can be treated in the global
46
coordinate system without transforming the coordinate system. In this chapter, the
governing equation of the static equilibrium and the dynamic problem of the body and
lines is constructed in a form of weak formulation based on the Galerkin method to
apply the Finite Element Method.
3.2 Theory of the Rod
The behavior of a slender rod can be expressed in terms of the variation of the
position of the rod centerline. A position vector ) , ( t s r is the function of the arc length s
of the rod and time t . The space curve can be defined by the position vector r . The unit
tangential vector of the space curve is expressed as r′ , the principal normal vector as r ′ ′ ,
and the binormal as r r ′ ′ × ′ , where the prime means the derivative with respect to the
arclength s . Figure 3.1 shows the coordinate system of the rod.
Figure 3.1 Coordinate system of the rod
X
s
Z
Y
F
M
r (s, t)
47
r q F & & ρ = + ′ (3.1)
0 = + × ′ + ′ m F r M (3.2)
where
centerline the along acting force resultant = F
centerline the along acting moment resultant = M
length unit per force applied = q
rod the of length unit per mass = ρ
length unit per moment applied = m
The dot denotes the time derivative. For the moment equilibrium, the bending moment
and the curvature has the relationship as:
r H r EI r M ′ + ′ ′ × ′ = (3.3)
where EI is the bending stiffness, and H is the torque. Equation (3.2) and (3.3) can be
combined as follows:
( ) 0 = + ′ ′ + ′ ′ +
(
¸
(
¸
+
′
′ ′ × ′ m r H r H F r EI r (3.4)
The scalar product with r′ for the equation (3.4) yields
0 = ′ ⋅ + ′ r m H (3.5)
where r m ′ ⋅ is the distributed torsional moment. Since there is no distributed torsional
moment, 0 = ′ ⋅ r m and 0 = ′ H . This means that the torque is independent on the
arclength s. Furthermore, the torque in the line is usually small enough to be negligible.
48
Here the torque H and the applied moment m on the line are assumed to be zero. Thus,
the equation (3.4) can be rewritten in the reduced form:
( ) 0 =
(
¸
(
¸
+
′
′ ′ × ′ F r EI r (3.6)
If a scalar function, ) , ( t s λ , which is also called Lagrangian multiplier, is introduced to
the equation (3.6) and the product with r′ is taken, then the following formula is
obtained:
( ) r r EI F ′ +
′
′ ′ − = λ (3.7)
where λ is the Lagrangian multiplier. r′ should satisfy the inextensibility condition:
1 = ′ ⋅ ′ r r (3.8)
Applying dot product with r′ to (3.7) using the relation of (3.8),
( ) r r EI r F ′ ⋅
′
′ ′ + ′ ⋅ = λ (3.9)
or
2
κ λ EI T − = (3.10)
If the equation (3.7) is substituted into (3.1), the following equation of motion is
obtained:
( ) ( ) r q r r EI & & ρ λ = +
′
′ +
″
′ ′ − (3.11)
If the stretch of rod is assumed to be linear and small, the inextensibility condition (3.8)
can be approximated as:
AE AE
T
r r
λ
≈ = − ′ ⋅ ′ ) 1 (
2
1
(3.12)
49
In the floating platforms, the applied force on the rod comes from hydrostatic and
hydrodynamic forces caused by the environmental excitation by the surrounding fluid,
and the gravitational force on the rod. Thus, q may be rewritten by:
d s
F F w q + + = (3.13)
where w is the weight of the rod per unit length,
s
F is the hydrostatic force on the rod
per unit length, and
d
F is the hydrodynamic force on the rod per unit length. The
hydrostatic force can be defined by:
( )
′
′ − = r P B F
s
(3.14)
where B is the buoyancy force on the rod per unit length, and P is the hydrostatic
pressure at the point r on the rod. The hydrodynamic force on the rod can be derived
from the Morison formula as:
d n
A
n n n n
D
n
M
n
A
d
F r C
r V r V C V C r C F
+ − =

.

\

− − + + − =
& &
&
&
&
& &
& &
(3.15)
where
A
C is the added mass coefficient (added mass per unit length ),
M
C is the inertia
coefficient (inertia force per unit length per unit normal acceleration of rod),
D
C is the
drag coefficient (drag force per unit length per unit normal velocity),
n
V is the normal
velocity to the rod centerline,
n
V
&
is the normal acceleration to the rod centerline,
n
r& is
the component of the rod velocity normal to the rod centerline, and
n
r& & is the component
of the rod acceleration normal to the rod centerline. The velocity and acceleration of the
50
rod can be derived from the fluid velocity vector, the line tangential vector, and their
derivatives.
( ) ( )  r r r V r V V
n
′ ′ ⋅ − − − = & & (3.16)
( )r r V V V
n
′ ′ ⋅ − =
& &
(3.17)
r r r r r
n
′ ′ ⋅ − = ) ( & & & (3.18)
r r r r r
n
′ ′ ⋅ − = ) ( & & & & & & (3.19)
When the above equation (3.13), (3.14) and (3.15) are used, then the equation (3.11) can
be rewritten as:
d
n
w a
F w r r EI r C r
~
~
)
~
( ) ( + = ′ ′ − ′ ′ ′ ′ + + λ ρ ρ & & & & (3.20)
where
2 2
~ ~
κ κ λ EI T EI P T − = − + = (3.21)
B w w + =
~
(3.22)
P T T + =
~
(3.23)
T
~
is the effective tension in the rod, and w
~
is the effective weight or the wet weight of
the rod. The equation (3.20) with the equation (3.12) is the fundamental equation of
motion for the elastic rod to be applied to the FEM formulation.
3.3 Finite Element Modeling
The governing equation (3.20) is nonlinear, and can be solved except for special
cases with particular conditions. Nordgren (1974) applied the finite difference method
51
to the governing equation and the inextensibility condition. His analysis results showed
satisfactorily the dynamic behavior of the pipe on the sea floor. In this study, the FEM
technique is taken due to its various merits. The application of the FEM starts from
describing the equation in the form of tensor such as:
0
~
~
)
~
( ) ( = + + ′ ′ + ′ ′ ′ ′ − − −
d
i i i i
n
i A i
F w r r EI r C r λ ρ & & & & (3.24)
and
0 ) 1 (
2
1
= − − ′ ′
AE
r r
r r
λ
(3.25)
Here the unknown variable λ , r can be approximated as:
) ( ) ( ) , ( t U s A t s r
il l i
= (3.26)
) ( ) ( ) , ( t s P t s
m m
λ λ = (3.27)
where, L s ≤ ≤ 0 ,
l
A ,
m
P are the interpolation(shape) functions, and
m
, λ
il
U are the
unknown coefficients. By introducing shape functions for the solution, the weak
formulations for applying the FEM technique are written by multiplying the weighting
function of
i
r δ as follows:
  0
~
~
)
~
( ) (
0
= + + ′ ′ + ′ ′ ′ ′ − − −
∫
ds F w r r EI r C r r
L
d
i i i i
n
i A i i
λ ρ δ & & & & (3.28)
0 ) 1 (
2
1
0
=
(
¸
(
¸
− − ′ ′
∫
ds
AE
r r
L
r r
λ
δλ (3.29)
The following cubic shape functions for
l
A and quadratic shape functions for
m
P are
used on the basis of the relation of ) (t U A r
il l i
δ δ = and
m m
P λδ δλ = such as equation
(3.26) and (3.27):
52
) (
2 3
) 2 (
2 3 1
3 2
4
2
3
3 2
2
3 2
1
ξ ξ
ξ ξ
ξ ξ ξ
ξ ξ
+ − ⋅ =
− =
+ − ⋅ =
+ − =
L A
A
L A
A
(3.30)
) 1 2 (
) 1 ( 4
2 3 1
3
2
2
1
− =
− =
+ − =
ξ ξ
ξ ξ
ξ ξ
P
P
P
(3.31)
where
L
s
= ξ .
) , ( ), , (
), , 0 ( ), , 0 (
4 3
2 1
t L r U t L r U
t r U t r U
i i i i
i i i i
′ = =
′ = =
(3.32)
) , ( ), ,
2
( ), , 0 (
3 2 1
t L t
L
t λ λ λ λ λ λ = = = (3.33)
Thus, the equation (3.30) and (3.31) can be extended in term by term as follows:
∫ ∫
+ = +
L
il l
n
i A i
L
n
i A i i
ds U A r C r ds r C r r
0 0
) ( ) ( δ ρ ρ δ & & & & & & & & (3.34)
il
L
i l
L
l i
L
l i
il l
L
i
L
i i
U ds r A EI A r EI A r EI
ds U A r EI ds r EI r
δ
δ δ
(
¸
(
¸
′ ′ ′ ′ + ′ ′ ′ − ′ ′ ′ =
′ ′ ′ ′ = ′ ′ ′ ′
∫
∫ ∫
0
0 0
0
0
) (
) ( ) (
(3.35)
il
L
l i
L
l i
il l i
L
i i
U ds A r A r
ds U A r ds r r
δ λ λ
δ λ λ δ
(
¸
(
¸
′ ′ − ′ =
′ ′ = ′ ′
∫
∫ ∫
0
0
0
~
)
~
(
)
~
( )
~
(
(3.36)
53
 
il
L
l
d
i i
L
d
i i i
U ds A F w ds F w r δ δ )
~
~
(
~
~
0 0



.

\

+ = +
∫ ∫
(3.37)
( ) ( )
∫ ∫
(
¸
(
¸
− − ′ ′ =
(
¸
(
¸
− − ′ ′
L
m r r m
L
r r
ds
AE
r r P ds
AE
r r
0
0
1
2
1
1
2
1
δλ
λ λ
δλ (3.38)
If the equation (3.34) to (3.37) are assembled and the term of
il
U δ is canceled out in
both sides of the above equations, the following equation is obtained:
( ) ( ) { }
( )  
L
l i i
L
l i
L
d
i i l l l
n
i A i l
A r EI r A r EI
ds F w A r A r A EI r C r A
0
0
0
~
~
~
~
′
′ ′ + ′ + ′ ′ ′ − =
+ − ′ ′ + ′ ′ ′ ′ + +
∫
λ
λ ρ & & & &
(3.39)
If the same operation is done for the equation (3.38), and
m
δλ is removed from both
sides of the equation (3.38), the equation (3.38) becomes as:
0 ) 1 (
2
1
0
=
)
`
¹
¹
´
¦
− − ′ ′
∫
ds
AE
r r P
L
r r m
λ
(3.40)
If the partial integrations are applied twice term by term for the equation (3.39) and
(3.40), and the boundary conditions satisfy the equation (3.39), then the following
equations are obtained:
∫ ∫
=
L
jk ij k l
L
i l
U ds A A ds r A
0 0
& &
& & δ ρ ρ (3.41)
( ) ( )
jk
L
ij js it t s k l
L
ij k l A
L
n
i A l
U ds U U A A A A ds A A C ds r C A
& &
& &
(
¸
(
¸
′ ′ − =
∫ ∫ ∫
0 0 0
δ δ (3.42)
∫ ∫
′ ′ ′ ′ = ′ ′ ′ ′
jk ij k l
L
i l
dsU A A EI ds r A EI δ
0
(3.43)
54
∫ ∫
′ ′ = ′ ′
L
ij k l n n
L
l
ds A A P ds r A
0 0
~
δ λ λ (3.44)
jk jl
L
k l m
L
r r m
U U ds A A P ds r r P
∫ ∫
′ ′ = ′ ′
0 0
2
1
2
1
(3.45)
∫ ∫
=
L
n m n
L
m
ds P P
AE
ds
AE
P
0
0
1
λ
λ
(3.46)
Using the equation (3.41) to (3.46), the equation (3.39) and (3.40) can be rewritten in a
matrix form as follows:
0 ) ( ) (
2 1
= − + + +
il jk nijlk n ijlk jk
a
ijlk ijlk
F U K K U M M λ
& &
(3.47)
0 = − − =
n mn m ki kl mil m
C B U U A G λ (3.48)
where,
∫
= ds A A M
ij k l ijlk
δ ρ (3.49)
(
(
¸
(
¸


.

\

′ ′ − =
∫ ∫
ij js it
L
t s k l
L
ij k l A
a
ijlk
U U ds A A A A ds A A C M δ δ
0 0
(3.50)
∫
′ ′ ′ ′ =
L
ij k l ijlk
ds A A EI K
0
1
δ (3.51)
∫
′ ′ =
L
ij k l n nijlk
ds A A P K
0
2
δ (3.52)
∫
+ =
L
l
d
i i il
ds A F w F
0
)
~
~
( (3.53)
55
and
ds A A P A
L
l i m mil
∫
′ ′ =
0
2
1
(3.54)
∫
=
L
m m
ds P B
0
2
1
(3.55)
∫
=
L
n m mn
ds P P
AE
C
0
1
(3.56)
and
ij
δ is the Kronecker Delta function. The equation (3.47) and (3.48) are used for
solving the rod dynamics. The program is implemented for calculating the equation
(3.49) to (3.56), using the system parameters and the integration of the shape functions.
Since the force vector,
il
F , contains nonlinear terms, the total equations are nonlinear.
So, in addition to the above manipulation, some numerical approaches for solving the
nonlinear timedomain problem in time domain are needed. In the following sections,
these schemes are introduced and explained.
3.4 Formation of Static Problem
The equations (3.47) and (3.48) can be called the equilibrium equation of the
system energy and the equation of the extensible conditions in the FEM. If the residuals
are taken from the system energy equation and the inextensibility equation, they should
be zero. Thus, the total force and the stretching force are described as
il
R and
m
G as:
56
0 =
il
R (3.57)
0 =
m
G (3.58)
In the static problem, the dynamic term is removed in the equation (3.36). It becomes as:
il jk nijlk n ijlk il
F U K K R − + = ) (
2 1
λ (3.59)
where
il
F is a static forcing term formed by gravity force, drag force and uniform current
and the other applied static force on the line. It is a nonlinear force vector. For solving
the equation, NewtonRaphson’s iterative method is used. Using the Taylor series
expansion, the equation (3.57) and (3.58), with neglecting the higher order terms, can be
expressed by:
0 ) ( ) (
) ( ) 1 (
= ∆
∂
∂
+ ∆
∂
∂
+ =
+
n
n
il
jk
jk
il n
il
n
il
R
U
U
R
R R λ
λ
(3.60)
0 ) ( ) (
) ( ) 1 (
= ∆
∂
∂
+ ∆
∂
∂
+ =
+
n
n
m
jk
jk
m n
m
n
m
G
U
U
G
G G λ
λ
(3.61)
And,
2 1
nijlk n ijlk
jk
il
K K
U
R
λ + =
∂
∂
(3.62)
jk nijlk
n
il
U K
R
2
=
∂
∂
λ
(3.63)
jk mkl
jk
m
U A
U
G
2 =
∂
∂
(3.64)
mn
n
m
C
G
− =
∂
∂
λ
(3.65)
57
If the equation (3.60) and (3.61) is rearranged by replacing the equation (3.62) to (3.65)
and is rewritten, they are given by:
) ( 2 2 1
) )( ( ) )( (
n
il
n jl nijlk jk nijlk n ijlk
R U K U K K − = ∆ + ∆ + λ λ (3.66)
) (
) ( ) ( 2
n
m n mn jk jl mkl
G C U U A − = ∆ − ∆ λ (3.67)
They can be rewritten in matrix form as follows:
¦
)
¦
`
¹
¦
¹
¦
´
¦
−
−
=
)
`
¹
¹
´
¦
∆
∆
(
(
¸
(
¸
) (
) (
) ( 1 ) ( 0
) ( 1
ln
) ( 0
n
m
n
il
n
jk
n t
mn
n t
mjk
n t
i
n t
ijlk
G
R
λ
U
D D
K K
(3.68)
where,
2 ) ( 1 ) ( 0
nijlk
n
n ijlk
n t
ijlk
K K K λ + = (3.69)
) (
0
) ( 2 ) ( 1
ln
n
jk
L
k l n
n
jk
nijlk
n t
i
U ds A A P U K K



.

\

′ ′ = =
∫
(3.70)
) (
0
) ( ) ( 0 n
jp
L
p k m
n
jp mkp
n t
mjk
U ds A A P U A D



.

\

′ ′ = =
∫
(3.71)
∫
− = − =
L
n m mn
n t
mn
ds P P
AE
C D
0
) ( 1
1
(3.72)
il
n
jk
nijlk n ijlk
n
il
F U K K R − + =
) ( 2 1 ) (
) ( λ (3.73)
0
) ( ) ( ) ( ) (
= − − =
n
n mn m
n
kl
n
ki
mil
n
m
C B U U A G λ (3.74)
After renumbering, the assembly equation in matrix form is given by:
) ( ) (
) (
n n
F y K = ∆ (3.75)
where,
58
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
¸
(
¸
−
−
−
−
−
−
=
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
¸
(
¸
′ ′ ′ −
′ ′ ′ + ′
′ ′ ′ −
′ ′ ′ + ′
′ ′ ′ −
′ ′ ′ + ′
′ ′ ′
′ ′ ′ + ′ −
′ ′ ′
′ ′ ′ + ′ −
′ ′ ′
′ ′ ′ + ′ −
=
=
=
=
=
=
=
=
=
=
=
=
=
0
0
0
0
] [
] ) ( [
] [
] ) ( [
] [
] ) ( [
0
0
] [
] ) ( [
] [
] ) ( [
] [
] ) ( [
] 2 [
3
] 2 [
3
] 2 [
2
] 2 [
2
] 2 [
1
] 2 [
1
] 1 [
3
] 1 [
3
] 1 [
2
] 1 [
2
] 1 [
1
] 1 [
1
3
3 3
2
2 2
1
1 1
0
3
0
3 3
0
2
0
2 2
0
1
0
1 1
L
N
L
N
L
N
L
N
L
N
L
N
A r EI
A r B r
A r EI
A r B r
A r EI
A r B r
A r EI
A r B r
A r EI
A r B r
A r EI
A r B r
L s
L s
l
L s
L s
l
L s
L s
l
s
s
l
s
s
l
s
l
s
l
r
λ
λ
λ
λ
λ
λ
F
(3.76)
 
3 34 33 24 23 14 13 2 1 32 31 22 21 12 11
λ λ λ U U U U U U U U U U U U
T
= y (3.77)
 
3 34 33 24 23 14 13 2 1 32 31 22 21 12 11
G R R R R R R G G R R R R R R
T
= F (3.78)
y y y ∆ + =
+ ) ( ) 1 ( n n
(3.79)
where [1] denotes the first end of element, and [2] the second end of element,
} {
T
N N N N
3 2 1
= is the nodal resultant force, } {
T
L L L L
3 2 1
= is the force relating to the
nodal resultant moment, and r L M ′ × = is the nodal resultant moment.
In every time step, the stiffness K and the force vector F are recalculated to
solve y ∆ . The bandwidth of the assembled stiffness matrix is 15, and the total number
of equations is 1 8 1 − × + ) (N , where N is the number of elements for a line. The stiffness
matrix is the symmetric and banded matrix. The Gauss elimination method for solving
59
the equation (3.75) conforming the symmetry and band is used. In addition, the iterative
solution scheme is used to get y ∆ until it becomes smaller than a given tolerance. The
resultant force can be obtained from force vector
r
F .
) 1 ( +
− =
n r
F F (3.80)
3.5 Formulation for Dynamic ProblemTime Domain Integration
The equation of motion, (3.47) and the stretch condition (3.48) can be rearranged.
il
il jk nijlk n ijlk jk ijlk
F
F U K K U M
ˆ
) (
ˆ
2 1
=
+ + − = λ
& &
(3.81)
0 = − − =
n mn m ki kl mil m
C B U U A G λ (3.82)
where,
jk nijlk n il
jk ijlk il
il il il il
a
ijlk ijlk ijlk
U K F
U K F
F F F F
M M M
2 2
1 1
2 1
ˆ
ˆ
λ =
=
+ − − =
+ =
(3.83)
The equation (3.81) is the second order differential equation, and the equation (3.82) is
an algebraic equation. The order of the equation (3.81) is derated using the first
derivative of the displacement of the rod, so that the equation results in two first order
differential equations as follows:
il jk ijlk
F V M
ˆ ˆ
=
&
(3.84)
jk jk
V U =
&
(3.85)
60
If the two equations are integrated, then they are given by:
∫ ∫
+ +
=
) 1 (
) (
) 1 (
) (
ˆ ˆ
n
n
n
n
t
t
jl
t
t
jk ijlk
dt F dt V M
&
(3.86)
∫ ∫
+ +
=
) 1 (
) (
) 1 (
) (
n
n
n
n
t
t
jk
t
t
jk
dt V dt U
&
(3.87)
In the equation (3.86),
ijlk
M
ˆ
is not a constant with respect to the time, since it includes
the added mass term. In order that the time integration is possible, a constant mass is
newly introduced.
)
2
1
(
ˆ
+ n
ijlk
M is the mass at time
2
) (
)
2
1
(
t
t t
n
n
∆
+ =
+
and a constant mass.
When the time step is ) (n 1 + ,
)
2
1
(
ˆ
+ n
ijlk
M can be used for the integration of the equation
(3.86). Then the integration is achieved with the 2
nd
order accuracy:
∫
+
= −
+
+
+
) 1 (
) (
ˆ ˆ ˆ
) (
)
2
1
(
) 1 (
)
2
1
(
n
n
t
t
jl
n
jk
n
ijlk
n
jk
n
ijlk
dt F V M V M (3.88)
The
) 1 ( + n
jk
V of the equation (3.87) is obtained from the following sequential calculations:
( )
) ( ) 1 ( ) ( ) 1 (
2
n
jk
n
jk
n
jk
n
jk
V V
t
U U +
∆
+ =
+ +
(3.89)
( )
) ( ) 1 ( ) ( ) 1 (
2
n
jk
n
jk
n
jk
n
jk jk
V V
t
U U U +
∆
= − = ∆
+ +
(3.90)
) ( ) 1 (
) (
2
n
jk jk
n
jk
V U
t
V − ∆
∆
=
+
(3.91)
61
Using the equation (3.91) and multiplying
t ∆
2
to both sides, the equation (3.88) can be
rewritten as:
∫
+
∆
+
∆
= ∆
∆
+ +
) 1 (
) (
ˆ
2
ˆ
4
) (
ˆ
4
) (
)
2
1
( )
2
1
(
2
n
n
t
t
jl
n
jk
n
ijlk
jk
n
ijlk
dt F
t
V M
t
U M
t
(3.92)
The integration of the right hand side of the equation (3.92) consists of three parts of
integration:
∫ ∫ ∫ ∫
+ + + +
+ − − =
) 1 (
) (
) 1 (
) (
) 1 (
) (
) 1 (
) (
2 1
ˆ
n
n
n
n
n
n
n
n
t
t
jl
t
t
il
t
t
il
t
t
jl
dt F dt F dt F dt F (3.93)
If the trapezoidal integration rule is applied, each term of the equation (3.93) is given by:
( )
 
) ( 1 1
) ( 1 ) 1 ( 1 1
2 ) (
2
2
) 1 (
) (
n
jk
ijlk jk ijlk
n
il
n
il
t
t
il
U K U K
t
F F
t
dt F
n
n
+ ∆
∆
=
+
∆
=
+
∫
+
(3.94)
( )
 
) ( 2 ) ( ) 1 ( 2 ) 1 (
) ( 2 ) 1 ( 2 2
2
2
) 1 (
) (
n
jk nijlk
n
n
n
jk nijlk
n
n
n
il
n
il
t
t
il
U K U K
t
F F
t
dt F
n
n
λ λ +
∆
=
+
∆
=
+ +
+
∫
+
(
¸
(
¸
∆ + ∆ +
∆
=
(
¸
(
¸
+
∆
≈
− −
+
+
+
) ( ) ( 2 2
2
2
2
)
2
1
(
) ( 2 ) ( 2
)
2
1
(
) ( 2
)
2
1
(
) 1 ( 2
)
2
1
(
jk nijlk
n
n n
n
jk nijlk
n
jk nijlk
n
n
n
jk nijlk
n
n
n
jk nijlk
n
n
U K U K U K
t
U K U K
t
λ λ λ
λ λ
(3.95)
where,
)
2
1
( )
2
1
( − +
− = ∆
n
n
n
n n
λ λ λ . The third term of the right hand side of the equation
(3.93) is the gravitational force and the hydrodynamic force. The gravitational force is a
62
constant with time. The hydrodynamic force can be calculated by applying Morison’s
formula and the AdamBashforth explicit integration scheme:
( )
¦
¹
¦
´
¦
−
∆
∆
=
− ∫
+
steps other for , 3
2
1 step for ,
) 1 ( ) (
) 0 (
) 1 (
) ( n
il
n
il
il t
t
il
F F
t
tF
dt F
n
n
(3.96)
The integration of force can be obtained by replacing the equations from (3.94) to (3.96)
into the equation (3.93). The time integration of the equation (3.92) is represented by:
( )
) ( 2
)
2
1
(
) ( 1 ) 1 ( ) ( ) (
)
2
1
(
) ( 2 2
)
2
1
(
1
)
2
1
(
2
2 2 3
ˆ
4
) ( 2 ) (
ˆ
4
n
jk
nijlk
n
n
n
jk
ijlk
n
il
n
il
n
jk
n
ijlk
n
n
jk
nijlk jk nijlk
n
n ijlk
n
ijlk
U K U K F F V M
t
U K U K K M
t
−
−
+
− +
− − − +
∆
=
∆ + ∆
(
(
¸
(
¸
+ +
∆
λ
λ λ
(3.97)
The mass at time
2
) (
)
2
1
(
t
t t
n
n
∆
+ =
+
is approximated using the AdamBashforth method
by:
( )
) 1 ( ) (
)
2
1
(
ˆ ˆ
3
2
1
ˆ
−
+
− =
n
ijlk
n
ijlk
n
ijlk
M M M (3.98)
By applying Taylor expansion to the stretching condition of the equation (3.82):
) ( 2 ) (
ˆ
2
) ( 2 ) ( 2 2
) ( 2 ) ( 2 2 2 0
) ( 1 ) ( 0 ) (
2 ) (
) ( ) (
) ( ) 1 (
n
n t
mn jk
n t
mjk
n
m
n mn jk il mijlk
n
m
n
n
n
m
jk
jk
n
m n
m
n
m
D U D G
C U U K G
G
U
U
G
G G
λ
λ
λ
λ
∆ + ∆ + =
∆ − ∆ + =
∆
∂
∂
+ ∆
∂
∂
+ ≈ =
+
(3.99)
Using the equation (3.97) and (3.99), the equation of motion and the stretching condition
can be written as follows,
) ( ) ( 1 ) ( 0
ˆ
) (
ˆ
) (
ˆ
n
il
n
n t
lin
jk
n t
ijlk
R K U K − = ∆ + ∆ λ (3.100)
63
) ( ) ( 1 ) ( 0
ˆ
) (
ˆ
) (
ˆ
n
m n
n t
mn jk
n t
mjk
G D U D − = ∆ + ∆ λ (3.101)
If the equation is written in matrix form, it gives:
¦
)
¦
`
¹
¦
¹
¦
´
¦
−
−
=
)
`
¹
¹
´
¦
∆
∆
(
(
¸
(
¸
) (
) (
) ( 1 ) ( 0
) ( 1 ) ( 0
ˆ
ˆ
ˆ
ˆ
ˆ
ˆ
n
m
n
il
n
jk
n t
mn
n t
mjk
n t
lin
n t
ijlk
G
R
λ
U
D D
K K
(3.102)
where,
( )
2
)
2
1
(
1 ) 1 ( ) (
2
) ( 0
ˆ ˆ
3
2
ˆ
nijlk
n
n ijlk
n
ijlk
n
ijlk
n t
ijlk
K K M M
t
K
−
−
+ + −
∆
= λ (3.103)
) ( 2 ) ( 1
2
ˆ
n
jk nijlk
n t
lin
U K K = (3.104)
) ( 0 ) ( 2 ) ( 0
2 2
ˆ
n t
mjk
n
il nijlk
n t
mjk
D U K D = = (3.105)
) ( 1 ) ( 1
2 2
ˆ
n t
mn mn
n t
mn
D C D = − = (3.106)
( ) ( )
) ( 2
)
2
1
(
) ( 1
) 1 ( ) ( ) ( ) 1 ( ) ( ) (
2 2
3
ˆ ˆ
3
2
ˆ
n
jk nijlk
n
n
n
jk ijlk
n
il
n
il
n
jk
n
ijlk
n
ijlk
n
il
U K U K
F F V M M
t
R
−
− −
− −
− + −
∆
=
λ
(3.107)
) ( ) (
2
ˆ
n
m
n
m
G G = (3.108)
The total equation in matrix form is written by:
ˆ
) (
ˆ
F y K = ∆ at time step n (3.109)
) 1 (
ˆ
+
− =
n r
F F (3.110)
3.6 Modeling of the Seafloor
The anchors are used for fixing the mooring lines and risers on the sea floor. The
interaction effect between the line and seafloor acts the important role on the line
64
movement. Thus, in the program, the seafloor is modeled as an elastic foundation, and
the friction force is not considered. With the origin of the coordinate system located on
the mean water surface and zaxis pointing upwards, the interaction force f on the line
from the sea floor can be expressed as;
0
1
= f , 0
2
= f ,
¦
¹
¦
´
¦
≥ −
< − −
=
0 for , 0
0 for , ) (
3
3
2
3
3
D r
D r D r c
f (3.111)
where D is the water depth or vertical distance between the sea floor and the origin of
the coordinate, and
3
r is the zcomponent of the line position vector r .
When the force from the sea floor is added, the equation of motion is rewritten by;
il
f
il
jk nijlk n ijlk jk
a
ijlk ijlk
F F U K K U M M + = + + + ) ( ) (
2 1
λ
& &
(3.112)
where
¦
¹
¦
´
¦
≥ −
< − −
=
=
∫
∫
0 for , 0
0 for , ) (
3
3
0
2
3 3
0
D r
D r ds D r c A
ds f A F
L
i l
L
i l
f
il
δ
(3.113)
¦
¹
¦
´
¦
≥ −
< − −
=
∫
0 for , 0
0 for , ) (
3
3
0
2
3 3
D r
D r ds D U A c A
L
jk k i i l
δ δ
and,
¹
´
¦ =
=
otherwise 0,
3 i for , 1
3 i
δ (Kronecker Delta) (3.114)
In the static analysis using Newton’s method, the dynamic stiffness matrix is modified
as:
65
¦
¹
¦
´
¦
≥ −
< − −
=
∂
∂
=
∫
0 for , 0
0 for , ) ( 2
) (
3
) (
3
0
) (
3 3 3
3
D U A
D U A ds D U A A c A
U
F
K
n
mn n m
n
mn n m
L
n
mn n m k j i l
jk
f
il
ijlk
δ
δ δ δ δ
(3.115)
This
3
ijlk
K is added to
0 t
ijlk
K in order to form the tangential stiffness matrix in the
equation (3.69). In time domain analysis using the trapezoidal rule, the dynamic stiffness
matrix is modified as:
( )
 
) ( 3
) ( ) 1 (
2 ) (
2
2
) 1 (
) (
n f
il
jk ijlk
n f
il
n f
il
t
t
f
il
F U K
t
F F
t
F
n
n
+ ∆
∆
≈
+
∆
=
+
∫
+
(3.116)
The first term in the RHS of the above equation is added to the LHS of the equation
(3.97), and it is finally combined into
0
~
t
ijlk
K . The second term in RHS of the equation
(3.116) is added to the RHS of the equation (3.97). Thus,
( )
) ( 2
)
2
1
(
1 ) ( ) 1 ( ) ( ) (
)
2
1
(
) ( 2 3 2
)
2
1
(
1
)
2
1
(
2
2 2 3
ˆ
4
) ( 2 ) (
ˆ
4
n
jk
nijlk
n
n ijlk
n f
il
n
il
n
il
n
jk
n
ijlk
n
n
jk
nijlk jk ijlk nijlk
n
n ijlk
n
ijlk
U K K F F F V M
t
U K U K K K M
t


.

\

+ − + − +
∆
=
∆ + ∆
(
(
¸
(
¸
− + +
∆
−
−
+
− +
λ
λ λ
(3.117)
66
CHAPTER IV
COUPLED ANALYSIS OF INTEGRATED PLATFORM AND MOORING
SYSTEM
4.1 Introduction
The statics and dynamics of the mooring lines and risers can be solved with the
given data and the boundary conditions. At both ends of the lines, different boundary
conditions are applied. The upper ends or the upper/lower ends, if the cable is installed
for the connection of the vessel to vessel (for the multiple body interaction problem), of
the lines are connected to the platform with strong springs. Thus, the end nodes are
moved with almost the same displacements as the floating platform. The other ends of
the lines are connected to the anchors on the seafloor and constrained with the fixed
conditions in six degrees of freedom. The platform is concentrated as a single point on
the center of the global coordinate and moved as a rigid body. It has six degrees of
freedom. The body behavior is greatly influenced by the movement of the mooring lines
and risers.
In the quasistatic analysis, the mooring lines and risers are treated separately to the
body motion. The motion of the body is solved first, and then, in the postprocessing, the
dynamics of the mooring lines and risers are analyzed with the motions of the end nodes
that are assumed to be the same amount as the body motion. The coupling effect of the
body and the lines can be considered, since the system matrices of body and lines are
assembled and solved together. But, the preobtained body motion cannot be evaluated
67
properly to consider the inertia effects and the hydrodynamic loads on the lines, because
the body motion is analyzed separately without considering the line dynamics.
On the contrary, in the coupled analysis, the body and lines are analyzed at the
same time. All dynamic effects of body and lines are included in system matrices, and
solved together. As the water depth gets deeper and deeper, the inertia effect increases.
So, the interaction effect greatly influences body and line motions. The coupled analysis
is to be an essential tool for solving the floating platform motion and line dynamics in
ultra deep water over 8,000 ft. in depth. The coupling effects were studied by Ran(2000).
He developed the mathematical formulation to be applied to solving the coupled system.
In his study, for static analysis, NewtonRaphson’s iterative scheme was used. But, for
the timedomain analysis, the AdamBashforth method was adopted as an explicit
numerical scheme. In this study, the above numerical methods are also adopted as a
numerical tool of the main solver, and the scheme is extended to the interaction problem
of multiple body systems of floating platforms.
4.2 The Spring to Connect the Platform and the Mooring System
The end connection is modeled numerically by the translational and rotational
springs between the body and lines. The stiffness should be considered strong enough so
that the body reacts with the same amount of motion as the lines’ in six DOFs (degrees
of freedom). If the spring is strong enough, the applied force and moment to come from
lines directly affects the body. If the angular motion is assumed small, the formulations
of the forces and moments to be transferred to the body from the lines is given by:
68
( )
i k j i i
L
i
S
i
r p p X K N − × + + = θ (4.1)
( ) ( )


.

\

′ ′
′ ′
−
′ ′
′
− × + =
2 / 3 2 / 1
n n
j i
m m
i
k j i
S
i
r r
r r
r r
r
e e K L θ
θ
(4.2)
where
¸ ¸
T
S S S S
i
N N N N
3 2 1
= and
¸ ¸
S S S S
i
L L L L
3 2 1
= are the nodal resultant forces and moments
on the end node of the line,
¸ ¸
L L L L
i
K K K K
3 2 1
= and
¸ ¸
θ θ θ θ
3 2 1
K K K K
i
= are the translational
and the rotational spring constants in the z y x , , direction and in the
z y x
θ θ θ , , direction,
i
X and
j
θ are the translational and rotational motions of the body,
i
p is the position
vector of the node of the body connected to the spring,
i
r is the position vector of the
ending or the starting node of the line attached by the spring to the body,
i
r′ is the space
derivative of the position vector
i
r , and
i
e is a unit vector of the reference direction of
the rotational spring. The
i
r vector at the end node of the line is defined as:
When the connection point is the starting point of the line:
11 1
U r = ,
21 2
U r = ,
31 3
U r = (4.3)
12 1
U r = ′ ,
22 2
U r = ′ ,
32 3
U r = ′ (4.4)
When the connection point is the ending point of the line:
13 1
U r = ,
23 2
U r = ,
33 3
U r = (4.5)
14 1
U r = ′ ,
24 2
U r = ′ ,
34 3
U r = ′ (4.6)
ji
C and
ji
D are defined to make easy the numerical manipulation of the vector product
with the position vector
i
p and the unit vector
i
e as:
69
 
(
(
(
¸
(
¸
−
− −
−
=
0
0
0
1 2
1 3
2 3
p p
p p
p p
C (4.7)
 
(
(
(
¸
(
¸
−
− −
−
=
0
0
0
1 2
1 3
2 3
e e
e e
e e
D (4.8)
If the equations (4.7) and (4.8) are used in equations (4.1) and (4.2), the equations are
rewritten as:
( )
i ji j i i
L
i
S
i
r C p X K N − + + = θ (4.1’)
( ) ( )


.

\

′ ′
′ ′
−
′ ′
′
− + =
2 / 3 2 / 1
n n
j i
m m
i
ji j i
S
i
r r
r r
r r
r
D e K L θ
θ
(4.2’)
The resultant force
S
i
F and moment
S
i
M transferred to the body are defined as follows:
S
i
S
i
N F − = (4.9)
ki
S
k ki
S
k
i
L
i
S
i
D L C N
M M M
+ =
+ =
θ
(4.10)
where
j
S
k
L
i
p N M × = is the moment resulting from the linear spring, and
j
S
k i
e L M × =
θ
is the moment resulting from the rotational spring. The force
S
i
F and the moment
S
i
M
act on the body.
4.2.1 Static Analysis
The connector force and moment on the end node of the line are included in the
equation of motion of the integrated system as external forces. In the static analysis, the
70
NewtonRaphson method is applied, so that the force and moment in (n+1) iteration are
approximated as follows:
For
i
r :
j ij j
rX
ij j
rr
ij
n
S
i
n
S
i
K X K r K N N θ
θθ
∆ + ∆ + ∆ + =
+ ) ( ) 1 (
(4.11)
For
i
r′ :
j
r
ij j
r r
ij
n
S
i
n
S
i
K r K L L θ
θ
∆ + ′ ∆ + =
′ ′ ′
+ ) ( ) 1 (
(4.12)
Where,
ij
L
i
j
S
i rr
ij
K
r
N
K δ =
∂
∂
− =
ij
L
i
j
S
i rX
ij
K
X
N
K δ − =
∂
∂
− =
ij
L
i
j
S
i r
ij
C K
N
K − =
∂
∂
− =
θ
θ
(4.13)
(
¸
(
¸
′ ′
′ ′
−
′ ′
=
′ ∂
∂
− =
′ ′
2 / 3 2 / 1
) ( ) (
n n
j i
m m
ij
i
j
S
i r r
ij
r r
r r
r r
K
r
L
K
δ
θ
ij i
j
S
i
ij
D K
L
K
θ θθ
θ
− =
∂
∂
− =
These equations that shows forces and moments will be expressed with the coupled
terms between body and line motions.
Similarly, the connector force and moment on the rigid body at iteration (n+1) are
approximated as follows using Newton’s method:
For
i
X :
j
X
ij j
XX
ij j
Xr
ij
n
i
n
i
K X K r K F F θ
θ
∆ + ∆ + ∆ + =
+ ) ( ) 1 (
(4.14)
For
i
θ :
j ij j
r
ij j
r
ij
n
i
n
i
K r K r K M M θ
θθ θ θ
∆ + ′ ∆ + ∆ + =
′ + ) ( ) 1 (
(4.15)
Where,
71
ij
L
i
j
i Xr
ij
K
r
F
K δ =
∂
∂
− =
ij
L
i
j
i XX
ij
K
X
F
K δ − =
∂
∂
− =
ij
L
i
j
i X
ij
C K
F
K − =
∂
∂
− =
θ
θ
(4.16)
ji j
j
i r
ij
C K
r
M
K
θ θ
=
∂
∂
− =
ji
n n
j i
m m
ij
j
j
i r
ij
D
r r
r r
r r
K
r
M
K
(
¸
(
¸
′ ′
′ ′
−
′ ′
=
′ ∂
∂
− =
′
2 / 3 2 / 1
) ( ) (
δ
θ θ
 
kj ki j kj ki
L
j
j
i
ij
D D K C C K
M
K
θ θθ
θ
+ − =
∂
∂
− =
The stiffness coefficients
rr
ij
K and
r r
ij
K
′ ′
are added the stiffness matrix of elements.
XX
ij
K ,
θ X
ij
K and
θθ
ij
K are included in the stiffness matrix of the platform. The other terms,
rX
ij
K ,
θ r
ij
K ,
θ r
ij
K
′
,
r
ij
K
θ
, and
r
ij
K
′ θ
, form the coupling terms in the assembled system matrix as
the symmetric matrices. At each iteration step, the coupled assembly system equations
are solved to obtain the behaviors for the body and lines simultaneously, and the
iteration continues until the norms of the solutions reach a specified tolerance.
4.2.2 TimeDomain Analysis
The integrations from time
) (n
t to
) 1 ( + n
t of the connector forces and moments on the
end node of the lines are expressed by applying Newton’s method as:
72
For
i
r :
( )
( )
) (
) ( ) 1 (
2
2
2
) 1 (
) (
n
S
i j
r
ij j
rX
ij j
rr
ij
n
S
i
n
S
i
t
t
S
i
N K X K r K
t
N N
t
dt N
n
n
+ ∆ − ∆ − ∆ −
∆
=
+
∆
=
+
∫
+
θ
θ
(4.17)
For
i
r′ :
( )
( )
) (
) ( ) 1 (
2
2
2
) 1 (
) (
n
S
i j
r
ij j
r r
ij
n
S
i
n
S
i
t
t
S
i
L K r K
t
L L
t
dt L
n
n
+ ∆ − ′ ∆ −
∆
=
+
∆
=
′ ′ ′
+
∫
+
θ
θ
(4.18)
The integrations from time
) (n
t to
) 1 ( + n
t of the connector forces and moments on the rigid
body are expressed as:
For
i
X :
( )
( )
) (
) ( ) 1 (
2
2
2
) 1 (
) (
n
i j
X
ij j
XX
ij j
Xr
ij
n
i
n
i
t
t
i
F K X K r K
t
F F
t
dt F
n
n
+ ∆ − ∆ − ∆ −
∆
=
+
∆
=
+
∫
+
θ
θ
(4.19)
For
i
θ :
( )
( )
) (
) ( ) 1 (
2
2
2
) 1 (
) (
n
i j ij j
r
ij j
r
ij
n
i
n
i
t
t
i
M K r K r K
t
M M
t
dt M
n
n
+ ∆ − ′ ∆ − ∆ −
∆
=
+
∆
=
′
+
∫
+
θ
θθ θ θ
(4.20)
Where the notations and the expressions for the K matrices follow the same convention
as the equations (4.13) and (4.16) in the static analysis.
4.3 Modeling of the Damper on the Connection
The damper on the connector is used for controlling the excessive resonance of the
high frequency vibration of the tensioned line like the tether or the riser in the TLP. The
damper is modeled as a linear damping force proportional to the vibratory velocity of the
73
line on the top connection node of the body and the line. The damping force,
D
i
N , on the
connection node of the line is given by:
( )
i k j i d
D
i
r p X C N &
& &
− × + = θ (4.21)
where
d
C is the damping coefficient, X
&
and θ
&
are the translational and rotational
velocity of the rigid body, r& is the velocity of the attached node of the line to the body.
k
p is the position vector of the attached node of the line at the connection point, and the
vector product of the
j
θ
&
and
k
p can be rewritten in the tensor form as
ji j k j
C p θ θ
& &
= × ,
as shown in the equation (4.1’). So, the equation (4.21) becomes:
( )
i ji j i d
D
i
r C X C N &
& &
− + = θ (4.21’)
It acts on the rigid body as reaction force by:
D
i
D
i
N F − = (4.22)
In the time domain analysis, the integration from time
) 1 ( + n
t to
) (n
t is obtained as:
For
i
r :
( )
i d j ji d i d
t
t
i ji j i d
t
t
D
i
r C C C X C
dt r C X C dt N
n
n
n
n
∆ − ∆ + ∆ =
− + =
∫ ∫
+ +
θ
θ
) 1 (
) (
) 1 (
) (
&
& &
(4.23)
For
i
X :
( )
i d j ji d i d
t
t
i ji j i d
t
t
D
i
r C C C X C
dt r C X C dt F
n
n
n
n
∆ + ∆ − ∆ − =
+ − − =
∫ ∫
+ +
θ
θ
) 1 (
) (
) 1 (
) (
&
& &
(4.24)
The equations of (4.23) and (4.24) show the terms of the geometric stiffness matrix of
the system. There are coupled terms with the body and the lines on the connection point.
The coupled terms can be solved together for body and line motions in the assembled
system matrix equations.
74
4.4 Modeling the Connection between Lines and Seafloor
The lower ends of the mooring lines and risers are normally connected to the
seafloor. The formulation for the connection part of the lines and the seafloor are very
similar to the modeling of the connection part of the body and the line. If the end
connection of the line consists of the anchor, the clamped or hinged boundary condition
is needed, and then it can be obviously replaced by considering a proper spring so that
the spring constant in the corresponding direction is to be large enough to hold the
rigidity of the anchor or the hinged boundary sufficiently. The connector force
F
i
N and
moment
F
i
L are defined by:
( )
i
F
i
L
i
F
i
r p K N − = (4.25)
( ) ( )


.

\

′ ′
′ ′
−
′ ′
′
− =
2 / 3 2 / 1
n n
j i
m m
i F
i
F
i
r r
r r
r r
r
e K L
θ
(4.26)
The damping force is defined as:
i
L
i
Fd
i
r K N & − = (4.27)
where
F
i
p is the position vector of the attached point of the seafloor,
F
i
e is the reference
direction vector of the rotational spring fixed on the seafloor, and
i
r and r′ are the
position vector and the tangential vector of the attached node to the seafloor. Since the
numbering of the lines starts from the seafloor when the line is attached to the seafloor,
the position vector is assigned as:
11 1
U r = ,
21 2
U r = ,
31 3
U r = (4.28)
12 1
U r = ′ ,
22 2
U r = ′ ,
32 3
U r = ′ (4.29)
75
4.5 Formulation for the Multiple Body System
The equation of motion and the equation of the stretching condition for the
multiple body system combined with any types of vessels can be derived in the same
way as the equation (3.47) and (3.48) for a single body system.
0 ) ( ) (
2 1
= − + + +
il jk nijlk n ijlk jk
a
ijlk ijlk
F U K K U M M λ
& &
(3.48)
0 = − − =
n mn m ki kl mil m
C B U U A G λ (3.49)
The two equations for a multiplebody system has the same form, and they can be
simplified as follows:
F KU U M = +
& &
(4.30)
0 Cλ B AU
2
= − − (4.31)
The   M ,   K ,   A and   C have the size of rows   1 ) 1 ( 8 − + × ×
E L
N N and the
bandwidth of 15, and   B , { } U
& &
, { } U
& &
, { } U , { }
2
U , { } F and { } λ are the vectors of the size
of   1 ) 1 ( 8 − + × ×
E L
N N , where
L
N is the total number of lines and
E
N is the number of
elements per each line. The global coordinate is used for composing each matrix,
regardless of the body to which the line is connected. In the next step, the matrix of
equations for the lines is combined with the matrix for the body motion including the
coupled terms in the stiffness matrix, and the assembled matrix and system equations are
dealt with in the next section.
After applying the Taylor expansion, the AdamsMoulton method, and the Adams
Bashforth method, and the Newton method of static and dynamic analysis, the equations
can be expressed in the matrix form as:
76
In static analysis:
¦
)
¦
`
¹
¦
¹
¦
´
¦
−
−
=
)
`
¹
¹
´
¦
∆
∆
(
(
¸
(
¸
) (
) (
) ( 1 ) ( 0
) ( 1
ln
) ( 0
n
m
n
il
n
jk
n t
mn
n t
mjk
n t
i
n t
ijlk
G
R
λ
U
D D
K K
(4.32)
where,
0
) (
) (
) ( 2 1 ) (
) ( 1
) ( ) ( 0
) ( 2 ) ( 1
ln
2 ) ( 1 ) ( 0
=
− + =
− =
=
=
+ =
n
m
il
n
jk nijlk n ijlk
n
il
mn
n t
mn
n
jp mkp
n t
mjk
n
jk nijlk
n t
i
nijlk
n
n ijlk
n t
ijlk
G
F U K K R
C D
U A D
U K K
K K K
λ
λ
(4.33)
In the dynamic analysis in time domain:
¦
)
¦
`
¹
¦
¹
¦
´
¦
−
−
=
)
`
¹
¹
´
¦
∆
∆
(
(
¸
(
¸
) (
) (
) ( 1 ) ( 0
) ( 1
ln
) ( 0
ˆ
ˆ
ˆ
ˆ
ˆ
ˆ
n
m
n
il
n
jk
n t
mn
n t
mjk
n t
i
n t
ijlk
G
R
λ
U
D D
K K
(4.34)
where,
( )
( ) ( )
) ( ) (
) ( 2
)
2
1
(
) ( 1
) 1 ( ) ( ) ( ) 1 ( ) ( ) (
) ( 1 ) ( 1
) ( 0 ) ( 2 ) ( 0
) ( 2 ) ( 1
2
)
2
1
(
1 ) 1 ( ) (
2
) ( 0
2
ˆ
2 2
3
ˆ ˆ
3
2
ˆ
2 2
ˆ
2 2
ˆ
2
ˆ
ˆ ˆ
3
2
ˆ
n
m
n
m
n
jk nijlk
n
n
n
jk ijlk
n
il
n
il
n
jk
n
ijlk
n
ijlk
n
il
n t
mn mn
n t
mn
n t
mjk
n
il nijlk
n t
mjk
n
jk nijlk
n t
lin
nijlk
n
n ijlk
n
ijlk
n
ijlk
n t
ijlk
G G
U K U K
F F V M M
t
R
D C D
D U K D
U K K
K K M M
t
K
=
− −
− + − =
= − =
= =
=
+ + − =
−
− −
−
−
λ
λ
∆
∆
(4.35)
77
The assembled equation of the coupled system of the rigid body and the lines can
be expressed as:
   
( )    
¦
)
¦
`
¹
¦
¹
¦
´
¦
=
¦
)
¦
`
¹
¦
¹
¦
´
¦
(
(
(
¸
(
¸
B
L
B
L
B C
C L
F
F
U
U
K K
K K
                 
T
(4.36)
where  
L
K is composed with the stiffness matrix of the lines and the connector springs,
 
B
K is the stiffness matrix of the rigid body,  
C
K and ( )  
T
C
K are the coupled stiffness
matrices and its transpose matrix including the coupling terms of the rigid body and the
lines.  
L
U and  
B
U denote the displacement matrices of the lines and the body, and
 
L
F and  
B
F are the force and moment terms acting on the lines and the body. The size
of  
B
K is 6 6× for a single body system, but for the multiplebody system N N 6 6 × ,
where N is the number of the multiple bodies. For a singlebody system,  
C
K has the
size of   1 ) 1 ( 8 − + ×
E
n rows and 6 columns per line. It has nontrivial terms of the size of
6 7 × at the last end rows of the matrix, and the remaining terms subtracting the
nontrivial terms from  
C
K are filled with zeros. The matrix ( )  
T
C
K is the transpose
matrix of  
C
K . When the multiplebody system is considered, and the hawser or the
fluid transfer line (FTL) between one body and another body is connected to body, the
total number of rows of the matrix  
C
K becomes   1 ) 1 ( 8 − + ×
E
n rows and
N × 6 columns per connecting line, where
E
n is the number of elements per line. It
makes two coupled terms on the starting node and the ending node of the connecting line.
78
Thus, it has the nontrivial terms twice of N 6 7 × in size, and the remaining terms except
the nontrivial terms are filled with zeros like those in a single body. The displacement
vector  
B
U and the force vector  
B
F for the rigid body have the size of 1 6 × N . The
stiffness matrix,  
L
K , of the lines has   1 ) 1 ( 8 − + × ×
E L
n n rows and the bandwidth of
15, where
L
n is the total number of lines. The matrix equation of total system explicitly
has the sparse matrix form. It means that a special consideration should be required to
solve it. Nowadays, some updated sparse matrix solvers are developed and announced
by many mathematical researchers. For this study, the forward and backward Gauss
elimination method as the rigorous and traditional solver is used, and modified slightly
for the purpose of treating the sparseness of the system matrix effectively. After the
forward elimination process is performed in the first step for solving the system matrix,
the backward substitution follows it next.
79
CHAPTER V
CASE STUDY 1:
DYNAMIC ANALYSIS OF A TANKER BASED FPSO
5.1 Introduction
As mentioned in the previous chapter, the hull/mooring line/riser coupled analysis
program for solving the twobody interaction problem was developed. Using this
program, the following case studies were performed for verification of the program. For
the first case, a tankerbased FPSO is taken. The tankerbased FPSO is designed for the
purpose of installation in the sea at the water depth of 6,000 ft. The environmental
conditions of the GoM (Gulf of Mexico) are used for the design.
The FPSO has a large, rotational movement during operation in the sea. In general,
due to this kind of specific large yaw rotation, the current and the wind force coefficients
are specially considered, and the experimental data of many years, based on many
VLCCs investigated and developed by Oil Company International Marine Forum
(OCIMF) is used. The wave loads induced by potential velocities are calculated by using
WAMIT that is a program to solve the potential problem of the fluid interaction.
The test model is selected as a turret moored FPSO in 6,000 ft. of water depth,
where the environmental conditions are the extreme hurricane conditions in the Gulf of
Mexico. The mooring system is a semitaut steel wire system. The results of the analysis
are compared with MARIN’s experimental results.
80
5.2 Design Premise Data of FPSO and Mooring Systems
The design premise data is described in this section. The vessel for this study is an
FPSO in 6,000 ft of the water depth. The capacity of the vessel storage is 1,440,000 bbls,
and the production level is 120,000 bpd. The dead weight of this vessel is 200 kDWT.
This vessel has an LBP of 310 meters, a molded breadth of 47.17 meters, and a depth of
28.04 meters as the main dimensions. In the full load condition, the draft is 18.9 meters
and the displacement is 240,869 MT. The turret is located at 63.55 meters aft of the
forward perpendicular of the vessel. The details of the design premise data are shown in
Table 5.1. The body plan and the isotropic view of the vessel are shown in Figure 5.1. In
the figure, the bow of the vessel is heading toward the east.
The mooring lines and risers are spread from the turret. There are 12 combined
mooring lines with chain, wire and chain, and 13 steel wire risers. Table 5.2 shows the
main particulars of mooring lines. Table 5.3 gives the hydrodynamic coefficients for
mooring lines. The main particulars of risers are shown in Table 5.4, and the
hydrodynamic coefficients are depicted in Table 5.5. The schematic plot of the
arrangement for mooring lines is shown in Figure 5.2. There are 4 groups of mooring
lines, each of which is normal to the other group. Each group is composed of 3 mooring
lines 5 degree apart from each mooring line in the group. The center of the first group is
heading the true East, and so the second group is toward the true North. Each mooring
line has a studless chain anchor of grade K4.
On the contrary, for the riser system, 19 lines are used in the prototype FPSO, but
for the simulation, only 13 risers among them are modeled equivalently as to what
81
MARIN did in their experimental tests. The risers are arranged nonsymmetrically with
respect to the xaxis (the axis toward the East). With respect to the yaxis (the axis
toward the North), the arrangement is also not symmetrical. But the risers are almost
balanced in the viewpoint of top tension with respect to both axes. The top view of the
arrangement of risers is shown in Table 5.6 and Figure 5.3 on the horizontal plane based
on the earth. In this study, the riser bending stiffness is not considered.
Table 5.1 Main particulars of the turret moored FPSO 6,000 ft
Description Symbol Unit Quantity
Production level bpd 120,000
Storage bbls 1,440,000
Vessel size kDWT 200
Length between perpendicular Lpp m 310.0
Breadth B m 47.17
Depth H m 28.04
Draft (in full load) T m 18.09
Diaplacement (in full load) MT 240,869
Lengthbeam ratio L/B 6.57
Beamdraft ratio B/T 2.5
Block coefficient Cb 0.85
Center of buoyancy forward section 10 FB m 6.6
Water plane area A m
2
13,400
Water plane coefficient Cw 0.9164
Center of water plane area forward section 10 FA m 1.0
Center of gravity above keel KG m 13.32
Transverse metacentric height MGt m 5.78
Longitudinal metacentric height MGl m 403.83
Roll raius of gyration in air R
xx
m 14.77
Pitch raius of gyration in air R
yy
m 77.47
Yaw radius of gyration in air R
ζζ
m 79.30
Frontal wind area Af m
2
1,012
Transverse wind area Ab m
2
3,772
Turret in center line behind Fpp (20.5 % Lpp) Xtur m 63.55
Turret elevation below tanker base Ztur m 1.52
Turret diameter m 15.85
82
Figure 5.1 The body plan and the isotropic view of FPSO 6,000 ft
83
Table 5.2 Main particulars of mooring systems
Table 5.3 Hydrodynamic coefficients of the chain, rope and polyester
Description Unit Quantity
Pretension kN 1,201
Number of lines 4*3
Degrees between 3 lines deg 5
Length of mooring line m 2,087.9
Radius of location of chain stoppers on turn table m 7.0
Length at anchor point m 914.4
Diameter mm 88.9
Weight in air kg/m 164.9
Weight in water kg/m 143.4
Stiffness, AE kN 794,841
Mean breaking load, MBL kN 6,515
Length m 1127.8
Diameter mm 107.9
Weight in air kg/m 42.0
Weight in water kg/m 35.7
Stiffness, AE kN 690,168
Mean breaking load, MBL kN 6,421
Length m 45.7
Diameter mm 88.9
Weight in air kg/m 164.9
Weight in water kg/m 143.4
Stiffness, AE kN 794,841
Mean breaking load, MBL kN 6,515
Segment 1 (ground position): chain
Segment 2: Polyester
Segment 3 (hangoff position): chain
Hydrodynamic Coefficients Symbol Chain Rope/Poly
Normal drag C
dn
2.45 1.2
Tangential drag C
dt
0.65 0.3
Normal added inertia coefficient C
in
2.00 1.15
Tangential added inertia coefficient C
it
0.50 0.2
Coulomb friction over seabed F 1.0 0.6
84
Figure 5.2 Arrangement of the mooring lines for FPSO 6,000 ft
Table 5.4 Main particulars of risers
Table 5.5 Hydrodynamic coefficients of risers
Points on
turnable
Connection
level
Total length
kN mm kN kg/m N/m m m m
Liquid production 4 1112.5 444.5 1.83E+07 196.4 1927/1037 1.0 4.88 1.52 1829
Gas production 4 609.7 386.1 1.08E+07 174.1 1708/526 1.0 4.88 1.52 1829
Water injection 2 2020.0 530.9 1.86E+07 285.7 2803/1898 1.414 4.88 1.52 1829
Gas injection 2 1352.8 287.0 3.14E+07 184.5 1810/1168 1.414 4.88 1.52 1829
Gas export 1 453.9 342.9 8.60E+06 138.4 1358/423 1.0 4.88 1.52 1829
Description No.
Radius of riser connection
Top
tension
Out
diameter
Stiffness,
AE
Mass
Dry weight/
wet weight
Cdn
Description Symbol Coefficients
Normal drag C
dn
1.0
Tangential drag C
dt
0.4
Normal added inertia coefficient C
in
1.0
Coulomb friction over seabed F 0.6
#3
#2
#1
#7
#8
#9
#10#11 #12
#4 #5 #6
NORTH
EAST
85
Table 5.6 Azimuth angles of risers bounded on the earth
Figure 5.3 Arrangement of the risers for FPSO 6,000 ft
5.3 Environmental Data
For the loading condition for the analysis, the 100year extreme hurricane
condition at the GoM is used, which is one of the severest in the world. The wave
condition is composed of the significant wave height of 12 m, the peak period of 14 sec,
and the overshooting parameter of 2.5. The wind spectrum of API formulae is taken as
(North)
(East)
X
1
X
2
LP#15
LP#14
LP#13
LP#16
GP#17 GP#18
GP#19
GP#20
WI#21
WI#22
GI#23
GI#24
GE#25
#1 #2 #3 #4
Liquid production (LP) 0 90 180 270
Gas production (GP) 45 135 225 315
Water injection (WI) 165 337.5
Gas injection (GI) 30 210
Gas export (GE) 300
Description
Azimuth angle of riser
86
the design condition. The mean wind velocity at the reference height of 10 m for one
hour sustained is 41.12 m/s. The current is mainly induced by the storm. The velocity of
current at the sea surface is 1.0668 m/s, and it keeps until 60.96 m under the sea surface.
From 60.96 m to 91.44 m under the sea surface, the current speed is varied from 1.0668
m/s to 0.05 m/s. For the intermediate region between 60.96 m to 91.44 m, the current
profile is determined by the linear interpolation. The current speed is uniformly kept
0.05 m/s from 91.44 m under the surface to the sea bottom.
While the storm wave and wind arise, the current is assumed as a onedirectional
current. But, when the GoM environmental condition is applied to the platform design,
the loop current in the GoM should be considered as a design loading condition. In this
study, however, the loopcurrent condition will not be applied, since the hurricane
condition is more severe than the loop current case. The summary of the environmental
condition for this study is shown in Table 5.7.
Table 5.7 Environmental loading condition
Description Unit Quantity
Significant wave height, Hs m 12.19
Peak period, Tp sec 14
Wave spectrum
Direction deg 180
1)
Velocity m/s 41.12 m/s @ 10m
Spectrum
Direction deg 210
1)
Profile
at free surface (0 m) m/s 1.0668
at 60.96 m m/s 1.0668
at 91.44 m m/s 0.0914
on the sea bottom m/s 0.0914
Direction deg 150
1)
Remark: 1) The angle is measured counterclockwise from the xaxis (the East).
Wind
Current
Wave
JONSWAP ( γ =2.5)
API RP 2AWSD
87
5.3.1 Wave Force
The JONSWAP spectrum was developed to define the wave by Hasselman, et al.
(1973) for the Joint North Sea Wave Project. The formula is to be derived from the
modified PiersonMoskowitz spectrum formula. The formula is given by:
(
(
¸
(
¸
−
−
−
−
(
(
¸
(
¸


.

\

− =
2
0
2
2
0
2
) (
exp
4
0
5 2
25 . 1 exp ) (
ω τ
ω ω
γ
ω
ω
ω α ω g S
(5.1)
where α is a parameter related to the prevailing wind field with the wind velocity of
w
U
and a fetch length of X , g is the gravitational acceleration, γ is the overshooting or
peakness parameter, and τ is the shape parameter. The α , γ and
0
ω are determined by
the following formulae:
( )
22 . 0
0
076 . 0
−
= X α (5.2)
¹
´
¦
>
≤
=
0
0
for 09 . 0
for 07 . 0
ω ω
ω ω
τ (5.3)
( )
33 . 0
0 0
2
−


.

\

= X
U
g
w
π ω (5.4)
where,
2
0
w
U
X
g X = . When X is unknown, α is taken as 0.0081. In this study, the wave
frequencies are considered to be between in 0.2 rad/s and 1.5 rad/s. Figure 5.4 shows the
wave spectrum with the given data.
88
Figure 5.4 JONSWAP wave spectrum
5.3.2 Wind Force
The formulae of API wind spectrum is as follows:
 
2
3 / 5
) (
/ 5 . 1 1
/
) ( z
f f
f f
f S
p
p
uu
σ
+
= (5.5)
where:
) ( f S
uu
= the spectral energy density at elevation z.
f = the frequency in hertz.
z V f
z p
/ 025 . 0 = = the average value of the frequencies of the measured wind
spectra
) (z σ = the standard deviation of wind speed, i.e.
JONSWAP Spectrum (Hs=12.19 m/s, Tp=14 s)
0
10
20
30
40
50
60
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
Frequency (rad/s)
P
o
w
e
r
S
p
e
c
t
r
a
l
D
e
n
s
i
t
y
(
m
^
2

s
e
c
)
89
z
V
z I
z
) (
) ( = σ (5.6)
125 . 0
) / ( H z V V
H z
= = the mean wind speed at elevation z for one hour
H
V = the mean wind speed at elevation 10 m for one hour
¦
¹
¦
´
¦
>
≤
= =
−
−
s s
s
z
z z z z
z z
V
z
z I
for ) / ( 15 . 0
z z for ) / ( 15 . 0
) (
) (
275 . 0
s
125 . 0
σ
(5.7)
= turbulence intensity over one hour
where
s
z = 20 m is the thickness of the surface layer.
Figure 5.5 shows the API wind spectrum of the given wind speed at the reference
elevation. After the normal wind force is calculated using the above wind spectrum, the
actual wind force varying with the weathervaning angle (yaw) of the vessel should be re
estimated by considering the force coefficients of the wind and the current in the OCIMF
booklet.
Figure 5.5 API wind spectrum
API Wind Spectrum (V
z
=41.12 m/s at 10 m)
0
2
4
6
8
10
12
0 5 10 15 20 25 30 35 40 45 50
Frequency (Hz)
P
S
D
o
f
W
i
n
d
S
p
e
e
d
(
m
/
s
)
^
2

s
e
c
90
5.3.3 Wind and Current Forces by OCIMF
The FPSO is a kind of tankerbased vessel. The OCIMF is the international
research committee that has been investigated the wind and current foresee subjected on
VLCC. In this study, the OCIMF booklet published in 1998 is referred to for calculating
the wind and current force coefficients. They suggest the following formula of the wind
and current force coefficients:
T w w xw xw
A V C F
2
2
1
ρ = (5.8)
L w w yw yw
A V C F
2
2
1
ρ = (5.9)
PP L w w xyw xyw
L A V C M
2
2
1
ρ = (5.10)
T L V C F
PP c c xc xc
2
2
1
ρ = (5.11)
T L V C F
PP c c yc yc
2
2
1
ρ = (5.12)
T L V C M
PP c c xyc xyc
2 2
2
1
ρ = (5.13)
where
xw
F and
yw
F are the surge and sway wind forces,
xyw
M is the yaw wind moment,
xc
F and
yc
F are the surge and sway current forces, and
xyc
M is the current yaw moment.
xw
C ,
yw
C and
xyw
C are the wind force and moment coefficients, and
xc
C ,
yc
C and
xyc
C
are the current force and moment coefficients.
w
ρ and
c
ρ are the densities of air and
fluid, and
w
V and
c
V are the wind velocity and current speed at the free surface.
T
A ,
L
A ,
91
T and
PP
L are the transverse area, the longitudinal area, the draft and the length
between perpendiculars of the vessel, respectively. They surveyed the force and moment
coefficients on the varying attack angle, for the two loading conditions, and for two
kinds of bow shapes. The attack angle is measured from 180 degree on the bow to 0
degree on the stern. The considered loading conditions are ballast and full load
conditions. For the bow shape, the cylindrical bow and the conventional bulbous bow are
taken. In the OCIMF booklet, the force and moment coefficients are shown in the
variation of the attack angle with parameters of the loading condition and the bow
configuration. For the current force coefficients, the water depth to draft ratio is also
taken as a parameter.
In this study, the tanker area and drag coefficients are assumed unchanged during
the time simulation. But, the coefficient for every 5 degree of attack angle is prepared in
advance, and at every time step during analyzing the yaw angle is swept. Whenever the
angle exceeds 5 degree, the wind and current force coefficients are recalculated using
the premade coefficient data files. The OCIMF formula for the wind and current forces
are to be expressed with respect to the center of the vessel, which is located near the
midship. Thus, the forces and moments give the localized components acting on the
vesselwise coordinate. The subject vessel is a turretmoored tanker, so the center of the
vessel movement should be the center of turret position, not the center of the vessel.
Therefore, to calculate the global motions of the vessel, the forces and moments are
transferred to the global coordinate components according to the yaw angle at every time
92
step during simulation. The force and moment are transferred by the inverse of rotation
matrix as follows:
Rotational matrix:
(
(
(
¸
(
¸
− =
1 0 0
0 cos sin
0 sin cos
θ θ
θ θ
T (5.14)
Inverse of rotational matrix:
(
(
(
¸
(
¸
−
=
−
1 0 0
0 cos sin
0 sin cos
1
θ θ
θ θ
T (5.15)
Coordinate transformation of force vector:
 Global force vectors:
¦
)
¦
`
¹
¦
¹
¦
´
¦
=
XY
Y
X
M
F
F
F (5.16)
 Local force vectors:
¦
)
¦
`
¹
¦
¹
¦
´
¦
=
¦
)
¦
`
¹
¦
¹
¦
´
¦
+
+
+
=
xy
y
x
xyc xyw
yc yw
xc xw
M
F
F
M M
F F
F F
f (5.17)
¦
)
¦
`
¹
¦
¹
¦
´
¦
(
(
(
¸
(
¸
−
= =
−
xy
y
x
M
F
F
1 0 0
0 cos sin
0 sin cos
θ θ
θ θ
f T F
1
(5.18)
Considering the translation of turret position:
xy turret y XY
M x F M + = (5.19)
Resultant force vectors:
¦
)
¦
`
¹
¦
¹
¦
´
¦
(
(
(
¸
(
¸
−
= =
−
xy
y
x
turret
M
F
F
x 1 0
0 cos sin
0 sin cos
θ θ
θ θ
f T F
1
(5.20)
93
where θ is the yaw rotation angle of the vessel and
turret
x is the xcoordinate of the turret
position in the body (local) coordinate system.
5.4 Hydrodynamic Coefficients
The hydrodynamic coefficients are calculated by using WAMIT, which can solve
the diffraction/radiation and the interaction problem of fluid and the platform structure.
The WAMIT is the program to solve the velocity potential on the wetted surface around
the floating structure based on the potential theory by means of the Boundary Element
Method (BEM) using the 3dimensional panel elements. BEM is the numerical
technique for considering only the wetted body surface and/or the water free surface
instead of considering the whole fluid domain. Taking Green’s function to satisfy all
other boundary conditions in the fluid domain as the weighting function in the integral
equation of motion makes it possible to solve the potential in the fluid domain.
In the linear theory, the added mass and linear damping coefficients, exciting
forces by diffraction potential, and mean drift forces can be obtained from the WAMIT.
By using the second order WAMIT, the quadratic transfer functions corresponding to the
secondorder difference frequency forces and the secondorder sum frequency forces can
be withdrawn. The modeling of the subject vessel is shown in Figures 5.6 and 5.7. Only
the port side of the vessel is modeled, and the symmetric condition is used for the
potential calculation in WAMIT. In the numerical model, the number of elements on the
body is 1870. Several models with other sized numberings are selected for convergence
94
study. Through the convergence study, the determined model was proved to be proper
for the analysis.
For the hydrodynamic coefficients, Newman’s (1974) approximation method is
used. In this method, the different frequency components are replaced by the mean part
of the linear transfer function (LTF). It is well known that the difference frequency
component of the quadratic transfer function is not sensitive to the frequency when two
frequencies are close. When two frequencies are quite large, the different frequency is
also large. Then, the frequency is far away from the natural frequency of the body or
mooring system. So, it also does not have much influence on the body or on the mooring
system.
X
Y
Z
Figure 5.6 Modeling of body surface of FPSO
95
Figure 5.7 Modeling of body surface and free surface of the water
5.5 Coupled Analysis of FPSO
In this study, the analysis case is explained for the turretmoored FPSO mentioned
in the previous section. The water depth is 6,000 ft (about 1828.8 m). The hydrodynamic
coefficients are calculated at every 5 degree of yaw angle by WAMIT, and WIMPOST
FPSO is used for the coupled analysis. The results are compared with MARIN’s. The
mooring lines and risers are modeled for preparing the input data of WINPOSTFPSO.
The mooring lines consisted of three parts, i.e., a chain anchor part, a wire part of
mid and a hangoff chain part. The first part is divided into 5 elements, the midpart
(wire) into eight elements, and the last chain part for the connection to the turret into 1
element. The connection boundary to the turret is modeled as a hinged joint. So, the
X
Y
Z
96
rotations are free, but no translation movement is allowed on that point. At the first node
of mooring line on the sea bed, the Dirichlet boundary condition is applied.
All Risers are treated as Steel Catenary Risers (SCRs). The risers are divided
uniformly into 12 elements. The boundary conditions for risers are the same as those for
mooring lines. The input data for wind, current force and wave loading are described in
Table 5.7.
Before the coupling dynamic analysis is performed, a static and dynamic balancing
test should be provided. Through these tests, the stiffness and system parameters such as
natural frequencies and damping factors of the numerical model can be judged whether
they are equivalent to the real system or not.
Firstly, the static offset test is carried out for the surge motion. During this test, the
FPSO is kept heading to 0 degree. From this test, the static weight balance with the top
tension of mooring lines and risers, the vessel weight and the buoyancy are checked.
Until a wellbalanced state is obtained, the footprints of mooring lines and risers are
adjusted back and forth. The stiffness of the combined system with the body and
mooring system is reviewed as well. To review the surge stiffness is a measure to judge
whether the vessel combined with mooring system is properly modeled or not.
Secondly, the free decay test is conducted for the surge, sway, heave, roll, pitch
and yaw motion in the calm water and in the 0 degree heading angle of the vessel. The
initial external force in the direction of the surge motion is set as 2.0E+07 N. The time
interval is defined as 0.02 sec. The surge external force is increased up to the initial force
97
level during four time steps, and then is released for 2,000 seconds. This test gives the
critical damping coefficients in the still water.
Finally, the coupled analysis in the time domain is carried out in irregular waves.
51 wave components are combined to generate the time series wave data with random
phases. The firstorder and also the secondorder wave forces are calculated using the
concept of a twoterm Volterra series model. The frequency range for this combination
is 0.15 rad/s to 1.2 rad/s. These are corresponding to 42 sec and to 5.2 sec, respectively.
Additional hull drag damping forces in the irregular state due to the current and waves
are evaluated with reference to the paper produced by Wichers(1996). The damping
coefficients for the hull drag forces are depicted in Figure 5.8. For the time simulation,
the time interval is set to 0.02 sec, and the total time to 3 hours. In the beginning part of
time duration, the ramping function is adopted to smoothly increase for 200 sec in order
to avoid the peculiar transient state.
Figure 5.8 Hull drag damping coefficients (Wichers, 1996)
1.46
2.64
1.36
1.00
#0 #2 #4 #18 #20
98
5.6 Results and Discussion
The added mass and radiation damping, firstorder wavefrequency forces, and
secondorder mean and differencefrequency forces are calculated from the secondorder
diffraction/radiation program WAMIT (Lee et al, 1991). Figure 5.9 shows the
distribution of panels on the body surface and free surface. Taking advantage of
symmetry, only half domain is discretized (1684 panels for hull and 480 panels for free
surface). All the hydrodynamic coefficients were calculated in the frequency domain,
and then the corresponding forces were converted to the time domain using twoterm
Volterra series expansion (Ran and Kim, 1997). The frequencydependent radiation
damping was included in the form of convolution integral to the time domain equation.
The wave drift damping was expected to be small and thus not included in the ensuing
analysis.
The methodology for hull/mooring/riser coupled statics/dynamics is similar to that
of Ran and Kim, 1997 and Kim et al., 1999. The mooring lines are assumed hinged at
the turret and anchor position. The nearvertical riser is also hinged at the turret, and
therefore, riser tension is included in the vertical static equilibrium of the hull. The
calculated platform mass for the given condition is
8
10 3686 . 2 × kg at 62ft draft. The
empirical coefficients for the viscous damping of the same FPSO hull in normal
direction were obtained from the model test by Wichers(2000a).
The wave force quadratic transfer functions are computed for 9 wave frequencies,
ranging from 0.24 to 1.8 rad/sec and the intermediate values for other frequencies are
interpolated. The hydrodynamic coefficients and wave forces are expected to vary
99
appreciably with large yaw angles and the effects should be taken into consideration for
the reliable prediction of FPSO global motions. Therefore, they are calculated in
advance for various yaw angles with a 5degree interval and the data are then tabulated
as inputs. The secondorder diffraction/radiation computation for a 3D body is
computationally very intensive especially when it has to be run for various yaw angles.
Therefore, many researchers avoided such a complex procedure and have instead used
simpler approach called Newman’s approximation(Faltinsen, 1998) i.e. the offdiagonal
components of the secondorder differencefrequency QTFs are approximated by their
diagonal values (mean drift forces and moments). This approximation can be justified
only when the relevant natural frequency is very small and the slope of QTFs near the
diagonal is not large. In this paper, the full QTFs are calculated and the validity of
Newman’s approximation is tested against more accurate results with complete QTFs.
The wind and current force coefficients on the vessel are read from OCIMF data. The
dynamic wind loading was generated from the wind velocities obtained from the API
wind spectrum. The yaw wind moments are increased by 15% considering the effects of
superstructures.
5.6.1 Static Offset Test (in Calm Water without Current)
The surge static offset test was conducted by pulling the VCG (Vertical Center of
Gravity) in the horizontal direction in calm water. Typical results for surge offsets are
shown in Figure 5.9. The surge staticoffset test shows a weakly softening trend, which
is contrary to the typical hardening behavior of catenary lines. The surge static offset
100
curves with risers are in general greater than those without risers due to the contribution
of riser tension. On the other hand, the effects of risers on individual mooring tension are
less appreciable. The results are shown in Figure 5.9.
(a) Static offset test results for surge motion
(b) Static offset test results of #2 mooring line in the surge direction
Figure 5.9 Static offset test results for surge motion
0.0E+00
2.0E+06
4.0E+06
6.0E+06
8.0E+06
1.0E+07
1.2E+07
1.4E+07
1.6E+07
1.8E+07
0 10 20 30 40 50 60 70 80 90 100
Offset [m]
S
u
r
g
e
f
o
r
c
e
[
N
]
Full Load(w. risers)
Full Load(w/o risers)
0.0E+00
1.0E+06
2.0E+06
3.0E+06
4.0E+06
5.0E+06
6.0E+06
7.0E+06
8.0E+06
0 10 20 30 40 50 60 70 80 90 100
Offset [m]
M
o
o
r
i
n
g
l
i
n
e
#
2
t
e
n
s
i
o
n
[
N
] Full Load (w. risers)
Full Load (w/o risers)
101
(c) Static offset test results of #8 mooring line in the surge direction
Figure 5.9 Continued
5.6.2 Freedecay Tests (in Calm Water without Current)
To see the effects of risers (mostly the amount of damping from risers) in the free
decay tests more clearly, a simpler riser model was developed i.e. all the 13 risers are
replaced by a single equivalent massless riser at the center with the same total tension.
The resulting surge/sway stiffness at the turret is then approximately calculated and
added to the hydrostatic matrix. Figure 5.10 shows typical freedecay test results for
surge, heave, roll, and pitch modes. The natural frequency and the damping coefficients
obtained from the free decay test are summarized in Table 5.8 and Table 5.9.
0.0E+00
2.0E+05
4.0E+05
6.0E+05
8.0E+05
1.0E+06
1.2E+06
1.4E+06
1.6E+06
0 10 20 30 40 50 60 70 80 90 100
Offset [m]
M
o
o
r
i
n
g
l
i
n
e
#
8
t
e
n
s
i
o
n
[
N
] Full Load (w. risers)
Full Load (w/o risers)
102
(a) Free decay test for surge motion
(b) Free decay test for heave motion
(c) Free decay test for roll motion
Figure 5.10 Freedecay test results for surge, heave and roll motions
120
100
80
60
40
20
0
20
40
60
80
100
0 200 400 600 800 1000 1200 1400 1600 1800
Time [sec]
S
u
r
g
e
[
m
]
Full Load (w. risers)
Full Load (w/o risers)
10
5
0
5
10
15
0 20 40 60 80 100 120 140 160 180 200
Time [sec]
H
e
a
v
e
[
m
]
Full Load (w. risers)
Full Load (w/o risers)
6
4
2
0
2
4
6
0 20 40 60 80 100 120 140 160 180 200
Time [sec]
R
o
l
l
[
d
e
g
]
Full Load (w. risers)
Full Load (w/o risers)
103
Table 5.8 Natural periods from freedecay tests
Table 5.9 Damping from freedecay tests estimated from the first 4 peaks
assuming linear damping
5.6.3 Timedomain Simulation for Hurricane Condition
The current is assumed to be steady and the irregular wave unidirectional. A
JONSWAP spectrum of significant wave height
s
H = 12.192 m, peak period
p
T =14s,
and overshoot parameter γ =2.5 was selected to represent a typical 100yr storm in the
Gulf of Mexico. The storm induced current flows from 30deg. right of wave direction.
The current velocity is assumed to be 3.5ft/s between 0200ft and reduced to 0.3ft/s at
300ft3000ft. The wind speed used is 92mph@10m and its direction is 30deg. left of
waves. The API wind spectrum is used for the generation of timevarying wind forces.
The drag coefficients for wave forces are 1.0 for mooring lines, 1.0 to 1.414 for risers.
The low and wavefrequency regions are defined as 00.2 rad/s and 0.21.3 rad/s,
respectively. The timedomain simulation results are summarized in Table 5.10.
Surge Heave Roll Pitch
Full draft
(with risers)
209.8 s 18.7 s 13.0 s 18.6 s
Full draft
(w/o risers)
225.9 s 18.7 s 13.4 s 18.6 s
Surge Heave Roll Pitch
Full draft (with risers)
11.0 %
(97.5 ~ 12.2 m)
6.5 %
(10.9 ~3.2 m)
0.86 %
(5 ~ 4.2 deg)
6.7 %
(5 ~ 1.4 deg)
Full draft (w/o risers)
5.8 %
(96.7 ~32.7 m)
6.1 %
(10.4 ~3.3 m)
0.68 %
(5 ~ 4.4 deg)
6.0 %
(5 ~ 1.6 deg)
104
Table 5.10 Timedomain simulation results (unit: m , deg.)
From this result, it is clearly seen that slowly varying components are dominant in
horizontalplane motions (surge, sway, yaw), while wavefrequency responses are more
important in verticalplane motions (heave, roll, pitch). It is also found that the effect of
riser damping is very important in the surge, particularly its slowly varying component.
When riser damping is absent, the surge rms and maximum values are
overestimated by about 47% and 35%, respectively. For the other modes, the effect of
riser damping is less significant. If riser damping is not accounted for, the total rms
Condition Mean Lowfreq.
RMS
Wave
freq. RMS
Total
RMS
Max
Newman’s Approx.
(with risers)
13.9 6.98 0.49 7.0 34.6
Newman’s Approx.
(w/o risers)
13.9 10.32 0.44 10.3 46.7
Surge
(m)
Full QTF (with risers) 14.7 8.42 0.44 8.4 39.5
Newman’s Approx.
(with risers)
4.7 2.50 0.49 2.5 13.4
Newman’s Approx.
(w/o risers)
4.6 2.84 0.45 2.8 13.8
Sway (m)
Full QTF (with risers) 4.8 3.04 0.46 3.1 16.9
Newman’s Approx.
(with risers)
0 0.04 3.36 3.4 10.9
Newman’s Approx.
(w/o risers)
0 0.03 3.46 3.5 12.1
Heave
(m)
Full QTF (with risers) 0.1 0.07 3.37 3.4 11.1
Newman’s Approx.
(with risers)
0.2 0.16 0.98 1.0 3.5
Newman’s Approx.
(w/o risers)
0.2 0.15 1.26 1.3 4.3
Roll
(deg.)
Full QTF (with risers) 0.1 0.38 1.22 1.3 5.5
Newman’s Approx.
(with risers)
0.0 0.02 1.33 1.3 4.3
Newman’s Approx.
(w/o risers)
0.0 0.02 1.39 1.4 4.7
Pitch
(deg.)
Full QTF (with risers) 0.0 0.04 1.34 1.3 4.5
Newman’s Approx.
(with risers)
15.3 2.74 0.28 2.6 22.7
Newman’s Approx.
(w/o risers)
13.7 2.57 0.31 2.7 22.3
Yaw
(deg.)
Full QTF (with risers) 15.1 3.86 0.28 3.9 24.3
105
tension values on taut(#2) and slack(#8) mooring lines are overestimated by 38% and
40%, respectively. The simulation results for mooring lines and risers are summarized in
Table 5.11. There also exist significant differences in rms and maximum tension of
individual risers, which indirectly shows the importance of fully coupled analysis.
Table 5.11 The results of tensions on the mooring lines and risers (unit: kN)
Condition Mean
Total
RMS
Max
Newman’s Approx.
(with risers)
2160 424 3529
Newman’s Approx.
(w/o risers)
2157 583 4252 Mooring Line #2
Full QTF
(with risers)
2201 479 3639
Newman’s Approx.
(with risers)
903 249 1860
Newman’s Approx.
(w/o risers)
943 349 2319 Mooring Line #8
Full QTF
(with risers)
901 296 2077
Newman’s Approx.
(with risers)
2345 272 4941
Liquid production
riser #13
Full QTF
(with risers)
2343 262 5393
Newman’s Approx.
(with risers)
1253 278 3509
Gas production riser
#20 Full QTF
(with risers)
1254 265 3213
Newman’s Approx.
(with risers)
4284 403 7629
Water injection riser
#22 Full QTF
(with risers)
4383 391 6923
Newman’s Approx.
(with risers)
2744 234 4082
Gas injection riser
#23 Full QTF
(with risers)
2746 227 4054
Newman’s Approx.
(with risers)
960 166 1804
Gas export riser #25
Full QTF
(with risers)
961 166 1781
106
In Table 3 and 4, the comparison between Newman’s approximation and the full
QTF is also shown. As expected, only horizontalplane motions are appreciably affected.
In general, the horizontalplane motion amplitudes (slowly varying parts) are under
estimated by using Newman’s approximation, but the differences are not large. The
error caused by massless riser modeling appears to be much more serious than that
caused by Newman’s approximation in this example.
5.7 Summary and Conclusions
The global motions of a turretmoored FPSO with 12 chainpolyesterchain
mooring lines and 13 steel catenary risers in a nonparallel windwavecurrent
environment are investigated in the time domain using a fully coupled hull/mooring/riser
dynamic analysis program. This case is similar to the relevant study in DEEPSTAR
Offshore Industry Consortium and the overall comparison looks reasonable.
In horizontalplane motions, slowly varying components are dominant, and
therefore, the reliable estimation of the secondorder mean and slowly varying wave
forces and the magnitude of total system damping is very important. For verticalplane
motions, wavefrequency responses are dominant and even the firstorder potential
based theory can do a good job in heave and pitch. The coupling effects are also minimal
in verticalplane motions.
In the present study, we particularly addressed two points, the effects of riser
coupling/damping and the validity of Newman’s approximation. The riser damping is
found to be important in surge/sway modes, particularly in surge. The use of Newman’s
107
approximation slightly underestimates the actual horizontalplane motions but seems to
be adequate in practical applications. However, when an input wave spectrum is not
narrowbanded or doublepeaked, care should be taken.
In a fully coupled simulation in the time domain, the behaviors of vessel, risers,
and mooring lines can be directly seen on the screen through graphicsanimation
software, which will greatly enhance the understanding of the relevant physics and the
overallperformance assessment of the system.
108
CHAPTER VI
CASE STUDY 2:
DYNAMIC ANALYSIS OF A TANKER BASED FPSO
COMPARED WITH THE OTRC EXPERIMENT
6.1 Introduction
In this study, the tanker based FPSO designed for the water depth of 6,000 ft and
tested in the OTRC basin is adopted for the verification of the WINPOSTFPSO
program. This FPSO is also a tanker–based and turretmoored vessel. The GoM
environmental conditions for wave, wind and current force are used in the analysis as
what the OTRC used in the experiment. The numerical model is made based on the
experimental model conducted in the OTRC basin. The principle data is the same as the
FPSO introduced in the previous chapter, but the loading condition is different, and the
turret position is moved forward to the bow. So, the draft is changed to 15.121 m, which
corresponds to 80 % loading of full load. The x coordinate of the turret position is
116.27 m along the ship’s center line, which is positioned at 38.734 meters aft of the
forward perpendicular of the vessel.
For the wind and current forces, the OCIMF data is used. The force coefficients are
taken for the full load and ballast loading. The force coefficients for 80 % loading are
interpolated automatically in the program using both data. The wave loads in the
consideration of the different loading with the previous vessel are calculated by using
WAMIT.
109
6.2 OTRC Experimental Results and Design Premise Data
Here the OTRC experimental results in the published paper in ISOPE 2001 will be
used for comparison with the analysis results by WINPOSTFPSO. The paper contains
the experimental results of the static offset test, the free decay test and some time
simulation. Due to the change of draft for the different loading conditions, many design
premise data should be changed. With the given draft, the principle data of vessel and
mooring line are estimated by some hand calculations and rechecked by some numerical
calculations.
The design premise data is basically the same as this in the previous chapter, except
for the draft and turret position. Using this basic design data and the OTRC experimental
results, the attempt to find the model data and the experimental condition data is tried.
The top tension of mooring lines is assumed to be the same as that of the original FPSO.
On the basis of this starting point, the weight balance is checked. The displacement can
be evaluated with the different loading condition data and corresponding draft. In this
loading condition, the draft is given as 15.121 meters. The displacement can be expected
to be 80 % of that of full load, so it will be 192,625 MT.
The details of the design premise data are shown in Table 6.1. The general
arrangement and body plan of the vessel are shown in Figure 6.1. As shown in the above
Figure, the vessel is toward the East (the bow is heading the East).
The mooring lines and risers are spread from the turret. In the original design data
there are 12 combined mooring lines with chain, wire and chain, and 13 steel wire risers.
There are 4 groups of mooring lines, each of which is normal to other group. Each group
110
is composed of 3 mooring lines 5 degrees apart from each mooring line in the group. The
center of the first group is heading the true East, and so the second group is toward the
true North. Each mooring line has a studless chain anchor of Grade K4.
Figure 6.1 General arrangement and body plan of FPSO 6,000 ft
Station#0 Station#20 Station#10
A.P. F.P. C.L.
0
1
2
3
4
5
610
1115
16
17
18
19
20
111
Table 6.1 Main particulars of the turret moored for the OTRC FPSO
Description Symbol Unit Quantity
Production level bpd 120,000
Storage bbls 1,440,000
Vessel size kDWT 200
Length between perpendicular Lpp m 310.0
Breadth B m 47.17
Depth H m 28.04
Draft (in full load) T m 15.121
Diaplacement (in full load) MT 240,869
Lengthbeam ratio L/B 6.57
Beamdraft ratio B/T 3.12
Block coefficient Cb 0.85
Center of buoyancy forward section 10 FB m 6.6
Water plane area A m
2
12,878
Water plane coefficient Cw 0.9164
Center of water plane area forward section 10 FA m 1.0
Center of gravity above keel KG m 13.32
Transverse metacentric height MGt m 5.78
Longitudinal metacentric height MGl m 403.83
Roll raius of gyration in air R
xx
m 
Pitch raius of gyration in air R
yy
m 
Yaw radius of gyration in air R
ζζ
m 
Frontal wind area Af m
2

Transverse wind area Ab m
2

Turret in center line behind Fpp (12.5 % Lpp) Xtur m 38.73
Turret elevation below tanker base Ztur m 1.52
Turret diameter m 15.85
112
Table 6.2 Main particulars of mooring systems for the OTRC FPSO
Table 6.3 Hydrodynamic coefficients of the chain, rope and wire for the OTRC
FPSO
Description Unit Quantity
Pretension kN 1,201
Number of lines 4*3
Degrees between 3 lines deg 5
Length of mooring line m 2,087.9
Radius of location of chain stoppers on turn table m 7.0
Length at anchor point m 914.4
Diameter mm 88.9
Weight in air kg/m 164.9
Weight in water kg/m 143.4
Stiffness, AE kN 794,841
Mean breaking load, MBL kN 6,515
Length m 1127.8
Diameter mm 107.9
Weight in air kg/m 42.0
Weight in water kg/m 35.7
Stiffness, AE kN 690,168
Mean breaking load, MBL kN 6,421
Length m 45.7
Diameter mm 88.9
Weight in air kg/m 164.9
Weight in water kg/m 143.4
Stiffness, AE kN 794,841
Mean breaking load, MBL kN 6,515
Segment 1 (ground position): chain
Segment 2: Polyester
Segment 3 (hangoff position): chain
Hydrodynamic Coefficients Symbol Chain Rope/Poly
Normal drag C
dn
2.45 1.2
Tangential drag C
dt
0.65 0.3
Normal added inertia coefficient C
in
2.00 1.15
Tangential added inertia coefficient C
it
0.50 0.2
Coulomb friction over seabed F 1.0 0.6
113
However, in ORTC model, only four equivalent mooring lines were used without
risers. One equivalent mooring line is combined with 3 mooring lines. Table 6.2 shows
the main particulars of equivalent mooring lines. Table 6.3 gives the hydrodynamic
coefficients for mooring lines. The equivalent mooring lines are spread 90 degrees apart
from the adjacent mooring lines. #1 equivalent mooring line goes to 45 degrees apart
from the true East. So, #2 equivalent mooring line is spread toward 135 degrees apart
from the true East. The schematic plot of the arrangement for mooring lines is shown in
Figure 6.2. With respect to the x and yaxis (the xaxis toward the East and the yaxis
toward the North), the mooring lines are arranged symmetrically. In the numerical model
for this study, the equivalent mooring system is used.
(a) Mooring system of the original FPSO (b) Mooring system of the OTRC experiment
Figure 6.2 Arrangement of mooring lines for turretmoored FPSO
Mooring Line #1
Mooring Line #3
Mooring Line #2
Mooring Line #4
Incident Wave
NORTH
EAST
4
50
#3
#2
#1
#7
#8
#9
#10#11 #12
#4 #5 #6
NORTH
EAST
114
6.3 Environmental Data
For the loading condition for the analysis, the 100year extreme hurricane condition
at the Gulf of Mexico (GoM) is used as the same as in the previous case. The wave
condition is composed of the significant wave height of 12 m, the peak period of 14 sec,
and the overshooting parameter of 2.5. The wind spectrum of NPD formulae is taken as
the design condition, which spectrum is shown in Figure 6.3. The mean wind velocity at
the reference height of 10 m for one hour sustained is 41.12 m/s. The current is mainly
induced by the storm. The wind direction is applied differently with the original FPSO
case in Chapter V. The velocity of current at the sea surface is 0.9144 m/s, and it keeps
until 60.96 m under the sea surface. From 60.96 m to 91.44 m under the sea surface, the
current speed is varied from 0.9144 m/s to 0.09144 m/s.
Table 6.4 Environmental loading condition for the OTRC FPSO
Description Unit Quantity
Significant wave height, Hs m 12.19
Peak period, Tp sec 14
Wave spectrum
Direction deg
180
1)
Velocity m/s 41.12 m/s @ 10m
Spectrum
Direction deg
150
1)
Profile
at free surface (0 m) m/s 0.9144
at 60.96 m m/s 0.9144
at 91.44 m m/s 0.0914
on the sea bottom m/s 0.0914
Direction deg
210
1)
Remark: 1) The angle is measured counterclockwise from the xaxis (the East).
Wind
Current
Wave
JONSWAP ( γ =2.5)
API RP 2AWSD
115
Figure 6.3 NPD wind spectrum curve
For the intermediate region between 60.96 m to 91.44 m, the current profile is
determined by the linear interpolation. The current speed is uniformly kept 0.09144 m/s
from 91.44 m under the surface to the sea bottom. While the storm wave and wind arise,
the current is assumed as one directional current. But, when the GoM environmental
condition is applied to the platform design, the loop current in the GoM should be
considered as a design loading condition. In this study, however, the loopcurrent
condition will not be applied, since the hurricane condition is severer than the loop
current case. The summary of the environmental conditions for this study is shown in
Table 6.4.
The current speed and direction in the OTRC experiment were set up differently
with the original FPSO case. In the original data, the current speed at the free surface is
1.07 m/s, and the direction is
o
150 from the xaxis (true East). But, in the OTRC
NPD Wind Spectrum, S(F)
0
500
1000
1500
2000
2500
3000
3500
4000
0 0.02 0.04 0.06 0.08 0.1
F (Hz)
S
(
F
)
116
experiment, the current speed was applied at the free surface of 0.9144 m/s, and the
direction of
o
210 .
6.4 Regeneration of the Experimental Model
The design data are reestimated to match the experimental model condition. The
natural frequencies obtained from the free decay test in the OTRC experiment are known
in a published paper (2001). The given data are D B L × × , T , KG, the turret position,
and the top tension of mooring lines as shown in Table 6.1. Using the experimental
model data and results, the required data should be newly estimated.
First, the hydrodynamic coefficients can be calculated by making the
hydrodynamic modeling and by using WAMIT (the fluid interaction software to get the
hydrodynamic coefficients), since the data of D B L × × , T and the body plan are given.
The numerical modeling for WAMIT is very similar to the FPSO model in the previous
chapter except the draft. From the WAMIT output, the displacement volume, the center
of buoyancy and the restoring coefficients can be obtained. The obtained data from the
WAMIT output is summarized in Table 6.5. Based on these data the weight of the model
can be derived from the static equilibrium condition that the sum of the line top tensions
and the weight is to be equal to the buoyancy. That’s the reason why the top tension is
called the net buoyancy:
Static equilibrium:
g
T
W B + = (6.1)
117
where B is the buoyancy, W denotes the weight of the body in mass unit,
g
T
is the
mass tension or the net buoyancy, and so T and g mean the top tension of mooring
lines and the gravitational constant, respectively.
Table 6.5 WAMIT output and handcalculation
The relations between the natural frequency, and the restoring coefficients and the
masses are defined as follows:
ij V
ij
M
C
f
π 2
1
= (1/sec or Hz) ) 6 , , 2 , 1 , ( L = j i (6.2)
Description Symbol Unit Quantity Reference
Displaced volumn m
3
182,499 WAMIT
Buoyancy m.ton 187,060
Total top tension kN 11,649 Given data
Weight in mass m.ton 185,870 Static equilibrium
Center of gravity m 109.670 Given data
m 1.801
Center of buoyancy m 89.086 WAMIT
m 7.401
Restoring coefficients 56.3226 WAMIT
22.3251
4688.27
Added mass/moment m.ton 1.9566E+05 WAMIT
m.tonm
2
1.1018E+07
m.tonm2 3.5189E+09
B
w
ρ × ∀
33 C
33 a
M
44 C
55 C
44 a
M
55 a
M
∀
T
W
b
z
g
x
g
z
b
x
118
where f is the natural frequency,
ij
C is the restoring coefficients in which i and j can
be any combination of six DOF, and ) (
ij ij a ij V
m M M + = is the virtual mass in which
ij a
M
is the added mass and
ij
m is the mass of the body in the i and j direction. The
relationship between
ij
m and W are as follows:
W m =
33
(6.3)
) (
2 2 2
44 g g xx
y z R W m + + = (6.4)
) (
2 2 2
55 g g yy
x z R W m + + = (6.5)
where ) , , (
g g g
z y x is the center of the gravity, and
xx
R ,
yy
R are the radii of gyrations for
roll and pitch motions. From the WAMIT output,
ij V
M can be obtained. These data are
also summarized in Table 6.5. The restoring coefficients are defined by:
w w
gA C ρ =
33
,
2
33
33
R w
gL
C
C
ρ
= (6.6)
Gt w g b w
A
w
M g mgz z g ds n y g C
w
∀ = − ∀ + =
∫∫
ρ ρ ρ
3
2
44
,
4
44
44
R w
gL
C
C
ρ
= (6.7)
Gl w g b w
A
w
M g mgz z g ds n x g C
w
∀ = − ∀ + =
∫∫
ρ ρ ρ
3
2
55
,
4
55
55
R w
gL
C
C
ρ
= (6.8)
where 33 C , 44 C and 55 C are the nondimensionalized restoring coefficients,
w
ρ and
w
A
are the water density and the water plane area, ∀ is the displaced volume,
b
z is the z
coordinate of the center of buoyancy, m is the mass of the body to be the same as W ,
and
R
L is the referenced length that is taken as the depth or the breadth of the vessel.
119
Here,
Gt
M and
Gl
M denotes the transverse and longitudinal metacentric heights and
3
n
represents the directional cosine in zdirection. Therefore, if the data in Table 6.6 and the
equation (6.3) to (6.8) are taken advantage of, the radii of gyrations, restoring
coefficients, and metacentric heights can be derived. The acquired data will be used as
the analysis model data, and are summarized in Table 6.6.
Next, using the equation (6.2) and the experimental results in Table 6.7, the data
are verified. It is the process to clarify whether the data obtained from the above
equations are acceptable for the numerical calculation on behalf of the experimental
model.
Table 6.6 Reestimated data from WAMIT output and handcalculation
6.5 Results and Discussion
6.5.1 Static Offset Test with Regenerated Model Data
The static offset tests are performed with the data obtained above by WINPOST
FPSO. The test results are depicted in Figure 6.4. They show the stiffness of the re
Description Symbol Unit Quantity
Water plane area A
w
m
2
12,878
Radius of roll gyration m 14.036
Radius of pitch gyration m 79.674
Radius of yaw gyration
m 81.400
Transverse metacentric height
m 11.950
Longitudinal metacentric height
m 1349.0
xx
R
yy
R
zz
R
Gt
M
Gl
M
120
estimated model is well matched with that of the OTRC model. Only a small difference
is shown in the initial point. It results from the fact that the OTRC experiment started
with the initial setting of the experimental instruments after a standing position in the
calm water at a certain moment was set as the static equilibrium state. But, it is hard to
say that moment is the same instant as the time when the model reached static
equilibrium position. The line tensions at #1 mooring line and #3 mooring line show a
slight difference from the experiments. It can make the difference in surge motion.
6.5.2 Free Decay Test with Regenerated Model Data
The proportional hull damping coefficients can be obtained from the free decay
tests and the results are compared with the OTRC experiments. With the regenerated
data, it is impossible and cannot be expected to get the same results once in the
numerical calculation. Fortunately, very similar results were obtained. After small
modification of the restoring coefficients, the compatible results for the natural periods
are obtained as in Table 6.7. The reason to adjust the restoring coefficients for matching
with the experimental is why the mooring line stiffness may contribute to the restoring
forces of the system.
121
(a) Static offset curves for surge motion obtained by experiments and WINPOSTFPSO
(b) Static offset test result of #2 mooring line in the surge direction
(c) Static offset test result of #1 mooring line in the surge direction
Figure 6.4 Comparison of the static offset test results
Static Offset Curve of FPSO 6000 ft Polyester  Surge Motion
0.0E+00
2.0E+06
4.0E+06
6.0E+06
8.0E+06
1.0E+07
1.2E+07
1.4E+07
1.6E+07
1.8E+07
0 10 20 30 40 50 60 70 80 90 100
Offset [m]
S
u
r
g
e
f
o
r
c
e
[
N
]
WINPOST(Full Load) WINPOST(OTRC)
MARIN(Experiment) OTRC(Experiment)
Static Offset Curve of FPSO 6000 ft Polyester  Mooring Line #2
0.0E+00
1.0E+06
2.0E+06
3.0E+06
4.0E+06
5.0E+06
6.0E+06
7.0E+06
8.0E+06
0 10 20 30 40 50 60 70 80 90 100
Offset [m]
M
o
o
r
i
n
g
l
i
n
e
t
e
n
s
i
o
n
[
N
]
WINPOST(Full Load) WINPOST(OTRC) MARIN(Experiment)
Static Offset Curve of FPSO 6000 ft Polyester  Mooring Line #1
0.0E+00
2.0E+05
4.0E+05
6.0E+05
8.0E+05
1.0E+06
1.2E+06
1.4E+06
1.6E+06
0 10 20 30 40 50 60 70 80 90 100
Offset [m]
M
o
o
r
i
n
g
l
i
n
e
t
e
n
s
i
o
n
[
N
]
WINPOST(Full Load) WINPOST(OTRC) MARIN(Experiment)
122
(a) Hull drag coefficients not in consideration of the current effect
(b) Hull drag coefficients in consideration of the current effect
Figure 6.5 Hull drag coefficients proposed by Wichers (1998 & 2001)
Table 6.7 Comparison of the free decay test results
1.46
2.64
1.36
1.00
#0 #2 #4 #18 #20
2.40
0.48
1.32
0.38
1.72
#0 #2 #4
#18 #20
0.23 0.19
1.13
Full Load
Ballast
period(sec) damping(%) period(sec) damping(%) period(sec) damping(%) period(sec) damping(%)
surge (m) 206.8 3.0 182.5 5.8 181.5 5.5 193.8 4.9
heave (m) 10.7 13.9 8.2 6.0 10.4 5.1 10.9 5.1
roll (deg) 12.7 4.4 13.4 0.9 12.7 1.1 12.6 0.8
pitch (deg) 10.5 16.5 13.9 6.0 10.8 8.5 10.9 8.5
OTRC Experiment
(4 equiv. Mooring
lines)
4 equiv. moorings
+ 1 riser
4 equiv. mooring lines
w/o riser
WINPOST
12 mooring lines
+13 risers
123
6.5.3 Time Simulation Results
The comparison of the OTRC experiment and the WINPOSTFPSO analysis is
shown in Table 6.8. In the table, the hull drag coefficients proposed by Wichers (1998,
2001) are used in this study as shown in Figure 6.5. The first column in the table is the
case to use the hull drag coefficients without considering the current. In cases illustrated
in the second and third column of the table, the hull drag coefficients considering the
current in sway and/or surge direction are used. When the drag coefficients considering
the current effect are used, the analysis results have the trend to follow the experiment in
sway and roll. But, in surge and yaw motion, there are still rather big differences
between the experiment and the numerical simulation results. The frontal wind area is
20 % larger, and the lateral area is 30 % larger than that of the full load case. The
difference in the projected wind areas can results in the difference of statistically
calculated values of motions. It can be caused by taking the mooring line truncation in
the experiment due to the depth limitation of the OTRC basin and the difference of the
mooring lines between the experimental model and the real vessel. Normally, the linear
steel springs are used for the implementation of the steel wiring mooring lines in the
experiments. As is well known, the spring has no static and dynamic mass. For the last
test among four different cases, the frontal areas in surge and sway direction are used as
the same as those in full load condition, and the drag coefficients in surge are multiplied
by 2.5 for reviewing the drag force effect.
124
Table 6.8 Comparison of time simulation results
In addition, it has no lateral stiffness, so it can react only in line. They can make
the difference in the surge and the yaw motions. The difference in the line tension as
Old Sway Cd
(1.5 hrs)
New Sway Cd
(1.5 hrs)
New Sway Cd
and Surge Cd
(1.5 hrs)
New Sway and
Surge Cd*2.5+old
wind area (3 hrs)
mean 22.92 25.22 20.26 19.39 20.89
min. 61.26 83.10 83.33 78.64 88.72
max. 2.29 21.31 22.67 18.94 24.49
rms. 9.72 24.13 23.18 21.02 18.84
mean 0.09 4.76 2.99 2.90 3.66
min. 21.43 8.17 8.21 7.15 12.14
max. 13.08 22.96 21.67 21.18 31.75
rms. 4.57 6.48 5.44 5.16 5.96
mean 0.14 0.39 0.38 0.38 0.38
min. 11.31 5.05 3.91 4.11 5.58
max. 10.91 4.28 3.28 3.26 5.15
rms. 3.08 1.51 1.32 1.31 1.42
mean 0.10 0.72 0.59 0.54 0.38
min. 3.60 11.41 11.91 11.70 14.95
max. 3.50 8.89 9.20 8.47 9.58
rms. 0.90 3.52 3.73 3.27 3.68
mean 0.01 0.06 0.04 0.03 0.05
min. 4.99 2.09 2.01 2.02 2.29
max. 4.45 1.35 1.35 1.46 1.64
rms. 1.31 0.59 0.53 0.53 0.56
mean 16.00 10.25 14.81 16.16 11.02
min. 24.60 20.23 22.95 22.61 24.07
max. 3.40 1.49 6.67 7.79 5.55
rms. 3.80 4.18 3.11 2.84 5.48
mean 5,907 6,403 6,487 6,440 7,757
min. 3,679 1,230 1,218 1,566 2,447
max. 10,360 14,600 14,893 14,173 16,783
rms. 827 2,688 2,735 2,565 2,359
mean 2,400 2,379 2,333 3,457
min. 197 202 204 511
max. 7,883 7,853 7,537 9,537
rms. 2,046 2,036 1,931 1,506
mean 2,644 2,593 2,562 3,657
min. 630 530 782 1,163
max. 7,540 7,543 7,067 9,233
rms. 1,893 1,898 1,796 1,346
mean 5,600 7,597 7,643 7,590 8,803
min. 2,927 802 827 1,041 2,511
max. 8,127 13,333 13,600 12,800 23,697
rms. 801 2,020 2,047 1,870 3,560
WINPOST (with 4equiv. line model)
Motion
roll (deg)
Mooring Tension
OTRC
Experiment
pitch (deg)
yaw (deg)
surge (m)
sway (m)
heave (m)
Mooring line
#1 (kN)
Mooring line
#2 (kN)
Mooring line
#3 (kN)
Mooring line
#4 (kN)
125
shown in the static offset tests in Table 6.3 (b) and (c) may be the reason for the
discrepancy. The new sway hull drag coefficients are used as shown in Table 6.5.
Furthermore, the surge drag force is newly considered (Cd=1.0). The analysis results are
rather close to the experiments in viewpoint of overall trend. But, the yaw and surge
motion still has a little large difference compared to the experiment.
For the consistency, Newman’s approximation scheme is used for evaluating the
wave forces applied to the single body model and also to the twobody model.
6.6 Summary and Conclusions
In this study, some efforts are exerted to regenerate the experimental results by the
OTRC. To find the model parameters, the experimental static offset curve and the free
decay test results are used. With the numerical model to be matched to the experimental
model, some analyses are conducted with the WINPOST program. When the hull drag
coefficients are applied in consideration of the current effect, the trends in sway and roll
motion may well follow the experimental results, but those in surge and yaw motion
show no good agreement. Some reasons for these differences can be imagined, such as
the wind force generation, the current profile control, the mooring line truncation and the
usage of springs for the steel wiring mooring lines. There are still many uncertainties for
the reasons for the differences between the experiment and the numerical analysis results.
For example, the investigation of the wind and current generated in the basin might give
some clues.
126
CHAPTER VII
CASE STUDY 3:
CALCULATION OF HYDRODYNAMIC COEFFICIENTS FOR TWO BODY
SYSTEM OF FPSO AND SHUTTLE TANKER
7.1 Introduction
In this study, the hydrodynamic coefficients for the twobody system are
performed and compared with the experimental results of other institutes (KRISO, 2002).
The multiple body system is composed of an LNG FPSO and a shuttle tanker. In many
cases of the conventional tandem mooring of the FPSO and shuttle tanker, the
hydrodynamic interaction between the two bodies has been ignored since the interaction
is not considered large enough to be taken account of. It has resulted in conservative
estimates for the behaviors of two bodies.
In this study, the interaction characteristics for the tandem and sidebyside moored
vessels are investigated and compared with the experiments carried out for a twobody
tanker model with different arrangements in regular waves. Motions and drift forces are
mainly reviewed with the numerical calculations by the WAMIT (Wave Analysis
program, developed by MIT using Boundary Element Method) program and
experiments. This program has the module to solve the interaction problem based on the
multiple body interaction theory. The changes of the distances between two vessels and
the mooring types are used as parameters for investigation of the interaction
characteristics.
127
There are several research works on this matter. Garrison (2000) developed the
numerical tool for the timedomain analysis of the hydrodynamic loads and motions for
a very large multibody floating structure(VLFS) using the panel method based on the
timedependent Green’s function. Inuoue and Islam(2001) investigated the roll motion
effect on wave drift force for the sidebyside moored vessels. Huijsmans, Pinkster and
Wilde(2001) tried to obtain the numerical approach to solve the diffraction and the
radiation potential problem for a very close multibody system. For the same topic,
Buchner, Dijk and Wilde(2001) developed the numerical time simulation solver to
predict the hydrodynamic response of alongside moored vessels.
Here, as the conventional mooring pattern, the tandem mooring is taken into
account since this type of mooring system has been used for the offloading operation in
the way that the shuttle tanker is located behind FPSO. On the situation, the distances are
kept between
4
1 to
3
1 of the ship’s length. As another mooring system, sidebyside
mooring is being considered since the offloading operations are sometimes preferred
under the parallel position in relatively calm seas. In such a case, the distance between
the two is very close, and so the hydrodynamic interaction and mooring design are very
important. For the test models, an LNG FPSO and a shuttle tanker are taken. For two
types of moorings and two different distances between the LNG FPSO and the shuttle
tanker, parametric studies of the interaction effects on the drift forces and vessel
behaviors are being performed in this study.
128
7.2 Particulars of Models and Arrangements for the Tests
Both models are tanker type vessels, of which the FPSO is fully loaded and the
other is ballast loaded. The main particulars, including the principle data of the vessels,
are listed in Table 7.1. The arrangements of tandem and sidebyside mooring are shown
in Figure 7.1. The distances between the two vessels in tandem mooring are taken as 30
m and 50 m. On the other hand, the distances for sidebyside mooring are determined as
4 m and 10 m. Steel springs for the mooring systems are used, and the stiffness of the
springs is set to 320 kN/m. The mooring lines modeled as springs are posted at the posts
located at the end of the mooring lines. For the calculation of the hydrodynamic
coefficients, the springs are not considered since the stiffness of the spring is too small
and so their hydrodynamic effects can be negligible. For the validity of the numerical
modeling for the two vessels, the natural frequencies are compared with each other.
According to the experiment by KRISO (2002), the roll natural period of the LNG FPSO
is 15.7 sec, and that of the shuttle tanker 9.97 sec. The free decay tests are conducted
with the numerical models, and according to the test results, the roll natural period of
15.8 for LNG FPSO, and of 10.1 sec for shuttle tanker. Table 7.2 shows the free decay
test results. The test reveals that the numerical model is good enough to use for the
numerical calculation. In Figure 7.2, the numerical models are shown. In Figure 7.3, the
finemeshed numerical models are shown, which is made for a sensitivity study. It has 4
times number of elements of the roughmeshed model. Consequently, it was proved that
the model size, i.e., the number of elements was not very sensitive to the results. In
129
Table 7.3, the comparison of the hydrodynamic coefficients obtained from the rough
model and the fine model is shown.
Table 7.1 Main particulars of two vessels
Description Symbol Unit LNG FPSO
Shuttle
Tanker
Length b/w perpendiculars Lpp m 239 223
Bredth B m 45.82 42
Draft at FP T
FP
m 15.82 6.8
Draft at midship T
MID
m 15.82 7.65
Draft at AP T
AP
m 15.82 8.5
Displacement m
3
139,585 53,743.20
Longitudinal center of gravity LCG m 9.636 8.152
Vertical center of gravity KG m 14.54 9.577
Metacentric height GM m 6.028 12.888
Radius of roll gyration K
xx
m 16.04 14.7
Radius of pitch gyration K
yy
m 59.75 55.75
Radius of yaw gyration K
zz
m 59.75 55.75
∀
130
Table 7.2 Freedecay test results for a LNG FPSO and a shuttle tanker
(heave and roll)
Time(s) Period(s) Heave(m) ln(x1/x2) Damp. Ratio 1st 3 Ave. Time(s) Period(s) Heave(m) ln(x1/x2) Damp. Ratio 1st 3 Ave.
0.0 0.0 2.399 0.0 0.0 3.118
12.0 12.0 1.459 0.50 7.91% 10.4 10.4 2.608 0.18 2.84%
23.6 11.6 0.801 0.60 9.54% 20.8 10.4 2.122 0.21 3.28%
35.4 11.8 0.46 0.55 8.83% 8.76% 31.2 10.4 1.739 0.20 3.17% 3.10%
47.0 11.6 0.258 0.58 9.20% 41.8 10.6 1.434 0.19 3.07%
58.4 11.4 0.152 0.53 8.42% 52.2 10.4 1.183 0.19 3.06%
69.8 11.4 0.092 0.50 7.99% 62.6 10.4 0.976 0.19 3.06%
81.2 11.4 0.06 0.43 6.80% 73.0 10.4 0.802 0.20 3.13%
92.4 11.2 0.042 0.36 5.68% 83.4 10.4 0.657 0.20 3.17%
Average 11.55 0.42 0.51 8.07% Average 10.43 1.44 0.19 3.13%
LNG FPSO SHUTTLE TANKER
4
3
2
1
0
1
2
3
4
0 20 40 60 80 100 120
Time [sec]
H
e
a
v
e
M
o
t
i
o
n
[
m
]
LNG FPSO Shuttle Tanker
Time(s) Period(s) Roll(deg) ln(x1/x2) Damp. Ratio 1st 3 Ave. Time(s) Period(s) Roll(deg) ln(x1/x2) Damp. Ratio 1st 3 Ave.
0.0 0.0 1.808 0.0 0.0 1.808
15.8 15.8 1.792 0.01 0.14% 10.0 10.0 1.798 0.01 0.09%
31.6 15.8 1.777 0.01 0.13% 20.0 10.0 1.786 0.01 0.11%
47.4 15.8 1.762 0.01 0.13% 0.14% 30.2 10.2 1.784 0.00 0.02% 0.07%
63.2 15.8 1.747 0.01 0.14% 40.2 10.0 1.779 0.00 0.04%
79.0 15.8 1.732 0.01 0.14% 50.2 10.0 1.765 0.01 0.13%
94.8 15.8 1.717 0.01 0.14% 60.2 10.0 1.756 0.01 0.08%
110.6 15.8 1.703 0.01 0.13% 70.4 10.2 1.75 0.00 0.05%
126.6 16.0 1.689 0.01 0.13% 80.4 10.0 1.738 0.01 0.11%
Average 15.8 1.74 0.01 0.13% Average 10.1 1.77 0.00 0.08%
LNG FPSO SHUTTLE TANKER
3.0
2.0
1.0
0.0
1.0
2.0
3.0
0 10 20 30 40 50 60 70 80 90 100
Time [sec]
R
o
l
l
M
o
t
i
o
n
[
d
e
g
]
LNG FPSO Shuttle Tanker
131
Table 7.3 Comparison of the hydrodynamic coefficients obtained from the rough
model and the fine models
(a) Tandem arrangement (b) Sidebyside arrangement
Figure 7.1 Configuration of the mooring system
Simple model
Extended
model
Simple model
Extended
model
Ma
11
2.9242E+06 2.9253E+06 7.0029E+05 7.1139E+05
Ma
22
3.7754E+07 3.7570E+07 9.1748E+06 9.1447E+06
Ma
33
1.2637E+08 1.2623E+08 9.4537E+07 9.4468E+07
Ma
44
1.2937E+09 1.2794E+09 4.3208E+06 4.3145E+06
Ma
55
1.2265E+11 1.2282E+11 2.8218E+08 2.8210E+08
Ma
66
5.2205E+10 5.2050E+10 4.4606E+07 4.4853E+07
Fd
11
1.8793E+02 1.8811E+02 2.7447E+01 2.7347E+01
Fd
22
5.1766E+02 5.1563E+02 4.2727E+01 4.2678E+01
Fd
33
1.4782E+06 1.4776E+06 8.9999E+05 8.9963E+05
Fd
44
1.0976E+03 1.1462E+03 1.4176E+01 1.4379E+01
Fd
55
8.9999E+06 9.0734E+06 1.3147E+05 1.3179E+05
Fd
66
1.9093E+04 1.8971E+04 2.8058E+00 2.6910E+00
Shuttle tanker
Added mass
Radiation damping
Symbol
Hydrodynamic
coefficients
1.6%
Max. difference in added
masses
LNG FPSO
1.0%
4.2% 4.3%
Max. difference in radiation
dampings
SHUTTLE TANKER
LNG FPSO
LNG FPSO SHUTTLE TANKER
132
(a) the sidebyside mooring arrangement (b) the tandem mooring arrangement
Figure 7.2 Roughmeshed numerical modeling for a LNG FPSO and a shuttle
tanker
(a) the sidebyside mooring (b) the tandem mooring
Figure 7.3 Finemeshed numerical modeling for a LNG FPSO and a shuttle tanker
7.3 Environmental Conditions
Regular waves are taken for the calculation of the beam sea and head sea
conditions. Only head sea conditions are considered for the tandem moored case. On the
contrary, for the sidebyside moored vessels, both beam sea and head sea conditions are
133
considered. The range of the wave frequencies is from 0.4 rad/s to 1.2 rad/s with 50
intermediate intervals.
7.4 Results and Discussion
The analysis results and the experiments can now be compared. The distances for
the sidebyside mooring are taken as 4 and 10 meter as the parameters, and on the
contrary, those for the tandem mooring are selected as 30 m and 50 m. Motion RAOs as
varying the distance apart from each other for the sidebyside mooring are compared as
shown in Figures 7.4 to 7.5 for heave and roll motions in beam sea state. For the
different mooring systems, the longitudinal drift forces are compared as shown in
Figures 7.6 and 7.7 for the head sea condition.
The distance effect on the longitudinal drift force is shown in Figure 7.8 for the
head sea condition.
The drift forces in the lateral direction for the sidebyside moored vessels are
shown in Figures 7.9 and 7.10 in different heading condition. For more clear comparison,
the calculated RAOs and drift forces for a single body of the FPSO and a single body of
the shuttle tanker in the same condition are depicted in the above figures. The whole
trends show good agreement to the experiments.
The shielding effects on heave and roll motion RAO are well investigated in the lee
side vessel of the sidebyside mooring vessels as shown in Figures 7.4 and 7.5. They are
very clear over the whole frequency range. As is well known, the effects are large
enough to pay attention to the matter for solving the interaction problem more accurately.
134
Figure 7.4 Heave response operators of sidebyside moored vessels in the beam
Sea
Heave RAO for a sidebyside mooring, Head=90 deg, Distance=10m
0.0
0.5
1.0
1.5
2.0
2.5
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
Frequency (rad/s)
H
e
a
v
e
R
A
O
(
Z
/
A
)
FPSOTwo Body
ShuttleTwo Body
FPSOSingle Body
ShuttleSingle Body
FPSOTwo Body (Exp.)
ShuttleTwo Body (Exp.)
Heave RAO for a sidebyside mooring, Head=90 deg, Distance=4 m
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
Frequency (rad/s)
H
e
a
v
e
R
A
O
(
Z
/
A
)
FPSOTwo Body
ShuttleTwo Body
FPSOSingle Body
ShuttleSingle Body
FPSOTwo Body (Exp.)
ShuttleTwo Body (Exp.)
135
Figure 7.5 Roll response operators of sidebyside moored vessels in the beam sea
Roll RAO for a sidebyside mooring, Head=90 deg, Distance=10m
0
2
4
6
8
10
12
14
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
Frequency (rad/s)
R
o
l
l
R
A
O
(
p
h
i
/
k
A
)
FPSOTwo Body
ShuttleTwo Body
FPSOSingle Body
ShuttleSingle Body
FPSOTwo Body (Exp.)
ShuttleTwo Body (Exp.)
Roll RAO for a sidebyside mooring, Head=90 deg, Distance=4 m
0
2
4
6
8
10
12
14
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
Frequency (rad/s)
R
o
l
l
R
A
O
(
p
h
i
/
k
A
)
FPSOTwo Body
ShuttleTwo Body
FPSOSingle Body
ShuttleSingle Body
FPSOTwo Body (Exp.)
ShuttleTwo Body (Exp.)
136
Figure 7.6 Longitudinal wave drift force of tandem moored vessels in the head sea
Drift force: Tandem mooring, Head=180 deg, Distance=50m
160
120
80
40
0
40
80
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
Frequency (rad/s)
X

D
I
R
.
D
r
i
f
t
F
o
r
c
e
(
k
N
/
m
2
)
FPSOTwo Body
ShuttleTwo Body
FPSOSingle Body
ShuttleSingle Body
FPSOTwo Body (Exp.)
ShuttleTwo Body (Exp.)
Drift force: Tandem mooring, Head=180 deg, Distance=30m
160
120
80
40
0
40
80
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
Frequency (rad/s)
X

D
I
R
.
D
r
i
f
t
F
o
r
c
e
(
k
N
/
m
2
)
FPSOTwo Body
ShuttleTwo Body
FPSOSingle Body
ShuttleSingle Body
FPSOTwo Body (Exp.)
ShuttleTwo Body (Exp.)
137
Figure 7.7 Longitudinal wave drift force of sidebyside moored vessels in the head
sea
Drift force: SideBySide, Head=180 deg, Distance=10m
320
240
160
80
0
80
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
Frequency (rad/s)
X

D
I
R
.
D
r
i
f
t
F
o
r
c
e
(
k
N
/
m
2
)
FPSOTwo Body
ShuttleTwo Body
FPSOSingle Body
ShuttleSingle Body
FPSOTwo Body (Exp.)
ShuttleTwo Body (Exp.)
Drift force: SideBySide, Head=180 deg, Distance=4m
320
240
160
80
0
80
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
Frequency (rad/s)
X

D
I
R
.
D
r
i
f
t
F
o
r
c
e
(
k
N
/
m
2
)
FPSOTwo Body
ShuttleTwo Body
FPSOSingle Body
ShuttleSingle Body
Exp(FPSO)KRISO
EXP(Shuttle)KRISO
138
Figure 7.8 The distance effect on the longitudinal wave drift force for a twobody
and a single body model in the head sea
Longitudinal Drift force: FPSO, Tandem mooring, Head=180 deg
200
160
120
80
40
0
40
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
Frequency (rad/s)
X

D
I
R
.
D
r
i
f
t
F
o
r
c
e
(
k
N
/
m
2
)
Two Body (50m)
Two Body (30m)
Single Body
Two Body (Exp.) (50m)
Two Body (Exp.) (30m)
Longitudinal Drift force: FPSO, SideBySide mooring, Head=180 deg
240
200
160
120
80
40
0
40
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
Frequency (rad/s)
X

D
I
R
.
D
r
i
f
t
F
o
r
c
e
(
k
N
/
m
2
)
Two Body (10m)
Two Body (4m)
Single Body
Two Body (Exp.) (10m)
Two Body (Exp.) (4m)
139
Figure 7.9 Lateral wave drift force of sidebyside moored vessels in the head sea
Drift force: SideBySide, Head=180 deg, Distance=10m
800
600
400
200
0
200
400
600
800
1000
1200
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
Frequency (rad/s)
Y

D
I
R
.
D
r
i
f
t
F
o
r
c
e
(
k
N
/
m
2
)
FPSOTwo Body
ShuttleTwo Body
FPSOSingle Body
ShuttleSingle Body
FPSOTwo Body (exp.)
ShuttleTwo Body (Exp.)
Drift force: SideBySide, Head=180 deg, Distance=4m
3000
2500
2000
1500
1000
500
0
500
1000
1500
2000
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
Frequency (rad/s)
Y

D
I
R
.
D
r
i
f
t
F
o
r
c
e
(
k
N
/
m
2
)
FPSOTwo Body
ShuttleTwo Body
FPSOSingle Body
ShuttleSingle Body
FPSOTwo Body (Exp.)
ShuttleTwo Body (Exp.)
140
Figure 7.10 Lateral wave drift force of sidebyside moored vessels in the beam sea
Drift force: SideBySide, Head=90 deg, Distance=10m
2000
1500
1000
500
0
500
1000
1500
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
Frequency (rad/s)
Y

D
I
R
.
D
r
i
f
t
F
o
r
c
e
(
k
N
/
m
2
)
FPSOTwo Body
ShuttleTwo Body
FPSOSingle Body
ShuttleSingle Body
FPSOTwo Body (Exp.)
ShuttleTwo Body (Exp.)
Drift force: SideBySide, Head=90 deg, Distance=4m
8000
6000
4000
2000
0
2000
4000
6000
8000
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
Frequency (rad/s)
Y

D
I
R
.
D
r
i
f
t
F
o
r
c
e
(
k
N
/
m
2
)
FPSOTwo Body
ShuttleTwo Body
FPSOSingle Body
ShuttleSingle Body
FPSOTwo Body (Exp.)
ShuttleTwo Body (Exp.)
141
As shown in Figures 7.6 and 7.7, the shielding effects on the longitudinal drift
forces for the head sea conditions are investigated, and are also remarkable in the tandem
moored vessel, but are not clear in the sidebyside moored vessels. The distance effect
on the drift force is not significant. The lateral drift force of sidebyside moored vessels
in head sea and in beam sea are quite different. As the distance gets closer, the blockage
effect on the lateral drift force increases. It causes the force to be magnified as the lee
side vessel approaches the weather side vessel, as shown in Figure 7.9 and 7.10.
7.5 Summary and Conclusions
The hydrodynamic interaction effects for the multibody system are investigated by
a comparative study for the numerical calculations and experiments. The LNG FPSO
and a shuttle tanker are taken as the multibody system, and the sidebyside and tandem
mooring are considered. The distance effects on the motions and drift forces of the two
vessels are also reviewed.
In tandem mooring, the shielding effect is noticeable on the drift force. The
distance has no great effect on the longitudinal force. In sidebyside mooring, the
shielding effect of the lee side vessel is significant on the drift force and motion RAO.
In lateral, the lee side ship acts as a block to disturb the flow pattern of the wave.
Furthermore, when the distance between both vessels gets closer, the magnitude of the
lateral drift seems to be reciprocally amplified against the distance. With comparing the
experiment, the WAMIT gives the fairly reasonable results, so that the conclusion is
drawn that the program can be applied to that kind of interaction problem.
142
CHAPTER VIII
CASE STUDY 4:
DYNAMIC COUPLED ANALYSIS FOR A TWOBODY SYSTEM COMPOSED
OF SPAR AND SPAR
8.1 Introduction
In this study, the dynamic coupled analysis for twobody structures is performed to
verify the program (WINPOSTMULT) for the dynamic coupled analysis of the
multiplebody floating platforms and the results are compared with the analysis results
using the idealized model of a twomassspring model. The multiple body system is
composed of two identity spars. The conventional tandem moorings have been taken for
the multiplebody connection in many cases. For the multiplebody model of spar
structures, the sidebyside mooring and the tandem mooring have no difference, since
the structure is symmetric about the x and yaxis. The simplified massspring model
will give a compatible result to judge the validity of the multiplebody program.
In this study, the body motions and line tensions are mainly reviewed with the
numerical calculations performed by WINPOSTMULT, the dynamic coupled analysis
program for multiplebody platforms. The hydrodynamic coefficients in consideration of
the multiplebody interaction are calculated by the WAMIT. The twobody interaction
problem of the fluid was studied in the previous chapter. The WAMIT program has the
module to solve the fluid interaction problem based on multiple body interaction theory,
as explained before. The analysis results by the program are compared with the analysis
143
results of the twobody spar model connected by a hawser with and without the
hydrodynamic interaction effect, and also compared with the results by the linear spring
model replaced for the hawser. Especially, for the linear spring modeling, the program is
modified slightly. From this study, the effect of the hawser to connect the two structures
can also be clarified. For this verification, the models with a hawser and without a
hawser are made and analyzed.
For the mooring system, the tandem mooring is taken into account since this type
of mooring system has been used for many years for offloading operations to transfer the
oil from one platform to other structures. The distance is kept as close as possible. Thus,
the distance is determined to be 30 meter to allow the maximum surge or sway motion,
since the expected maximum surge motion is about 30 meters and the maximum sway
motion about 10 meters. It can be said that the sidebyside mooring should be identical
to the tandem mooring due to the symmetry of the structure.
8.2 Particulars of Models and Arrangements for the Analyses
The main particulars including the principle data of spar are listed in Table 8.1.
The arrangement of the tandem is shown in Figure 8.1. The distance between the two
spars in tandem mooring is taken as 30 m. The mooring lines are fixed at the sea floor.
For the calculation of the hydrodynamic coefficients, the WAMIT program is. For the
validity of the numerical modeling, Static offset test and free decay tests are performed
and compared with the target values, which are given from experiments conducted by
other institute.
144
Table 8.1 Main particulars of moored spar
Figure 8.1 Configuration of the mooring system and the environmental loads
(Tandem arrangement, d=30m)
Description Symbol Unit Quantity
Water depth m 914.4
Production level of oil bpd 55,000
Production level of gas mmscfd 72
Length m 214.88
Draft T m 198.12
Hard tank depth H m 67.06
Well bay dimension (25 slots) m 17.68 x 17.68
Center of buoyancy center above base line KB m 164.59
Center of gravity above base line KG

m 129.84
KG (based on total displacement) KG m 95.71
Displacement

mT 53,600
Total displacement mT 220,740
Pitch radius of gyration in air R
xx
m 67.36
Yaw radius of gyration in air R
ζζ
m 8.69
Drag force coefficient Cd 1.15
Wind force coefficient Cw N/(m/s)
2
2671.6
Center of pressure above base line m 220.07
∀
∀
Dia.=37.1856 m d=30 m
SPAR #2 SPAR #1
Hawser
Wave
Wind
Current
145
In Figures 8.2 and 8.3, the numerical models are shown. In Table 8.2, the particulars of
the mooring systems are tabulated.
Table 8.2 Particulars of the mooring systems
Description Unit Quantity
Pretension kN 2,357
Number of lines 14
Scope ratio 1.41
Length of mooring line m 1,402.08
Firlead location above base line m 91.44
Length at anchor point m 121.92
Diameter mm 24.5
Weight in air kg/m 287.8
Weight in water kg/m 250.3
Stiffness, AE kN 1.03E+06
Minimum breaking load, MBL kN 1.18E+04
Added mass kg/m 37.4
Current force coefficient 2.45
Length m 2347.44
Diameter mm 21.0
Weight in air kg/m 36.52
Weight in water kg/m 7.77
Stiffness, AE kN 3.18E+05
Mean breaking load, MBL kN 1.28E+04
Added mass kg/m 28.8
Current force coefficient 1.20
Length m 91.44
Segment 1 (ground position): chain
Segment 2: wire
Segment 3 (hangoff position): chain
Other parameters are the same as those of segment 1.
146
8.3 Environmental Conditions
The environmental conditions to be used in this analysis correspond to the 100
year storm conditions in Gulf of Mexico. The wind velocity is 41.12 m/s at 10 m of
reference height for 1 minute sustained. For wind force calculation, API RP2T is used.
For wave, irregular waves are taken for the calculation of the head sea condition. The
range of the wave frequencies is from 0.5 rad/s to 1.2 rad/s with 50 intermediate
intervals. The wave spectrum used here is the JONSWAP spectrum, as shown in Figure
8.3, which has the significant wave height of 12.192 meters, the peak period of 14
seconds, and the overshooting parameter of 2.5. The current velocity is 1.0668 m/s at the
free surface, and it is kept 60.96 m under the water surface. After that, it varies from
1.0668 m/s to 0.0914 m/s from 60.96 m to 91.44 m under the water surface. Under the
water depth of 91.44 m, the current speed becomes uniform as 0.0914 m/s. In Table 8.3,
the environmental conditions are summarized.
Table 8.3 Environmental conditions
Description Unit Quantity
Significant wave height, Hs m 12.19
Peak period, Tp sec 14
Wave spectrum
Direction deg 180
1)
Velocity m/s 41.12 m/s @ 10m
Spectrum
Direction deg 210
1)
Profile
at free surface (0 m) m/s 1.0668
at 60.96 m m/s 1.0668
at 91.44 m m/s 0.0914
on the sea bottom m/s 0.0914
Direction deg 150
1)
Remarks: 1) The angle is measured from xaxis (the East)
in the counterclockwise.
Wind
Current
Wave
JONSWAP ( γ =2.5)
API RP 2AWSD
147
8.4 Calculation of Hydrodynamic Coefficients Using WAMIT 1
st
and 2
nd
Order
In Figures 8.2 and 8.3, the numerical models are shown. The hydrodynamic
coefficients are calculated by WAMIT. For the single body analysis, the 2
nd
order wave
force coefficients are calculated with free surface modeling. For the twobody analysis,
the 1
st
order wave force coefficients and wave drift force coefficients are calculated. The
hydrodynamic coefficients of added mass, wave damping, linear transfer function (LTF)
of diffraction potential force and the sum and differencefrequency quadratic transfer
function (QTF) of diffraction potential force are calculated by the WAMIT 1
st
order
module and the 2
nd
order module. In Figure 8.2, the model for the 2
nd
order wave force
coefficients is shown. The body has 1024 elements, and the free surface has 576 panel
elements. In Figure 8.3, the twobody model for the 1
st
order wave force coefficients is
shown. Here, for the purpose of comparison, the 1
st
order model is used for the single
body analysis and also for the twobody analysis, so that for both analyses Newman’s
Approximation Method is adopted for conforming the full QFT when the wave force
coefficients are considered. The hawser connecting each spars to the other is taken to
have 1/100 of the mooring stiffness and 1/10 of the mooring pretension.
The hydrodynamic interaction effect is calculated with the 1
st
order model. All
coupling terms are considered for the twobody analysis. The program WINPOST
MULT can treat the numerical calculation with the fully coupled system matrices
composed by multiple bodies. The added mass, the linear wave damping, the system
stiffness and the resorting coefficient matrix are fully coupled with each other due to the
interaction effects of both structures. Especially, if the hawser or the fluid transfer lines
148
are connected, they will cause to make the stiffness matrix coupled so that the whole
system stiffness matrix composed by the body and line stiffness and restoring
coefficients comes to a huge sparse matrix.
As mentioned above, the analysis of the twobody system is performed using the
1
st
order model with and without interaction effects. In the case of no interaction effects,
the coupling terms of the hydrodynamic coefficients are set as zero.
Figure 8.2 Configuration of the modeling of a single spar
Figure 8.3 Configuration of the modeling of a twobody spar
149
8.5 Linear Spring Modeling
The hawser for connecting the two spars can be replaced by a linear spring. For
verifying the numerical analysis results by the full numerical model, a linear spring for
the hawser is considered by putting the linear spring constant as a restoring coefficient in
surge direction into the body system matrix of the restoring force coefficients inside the
program. Furthermore, the WINPOSTMULT program is modified slightly since the
replaced spring can work only when two bodies move in the opposite direction against
each other out of phase. At every time step, the distance between both spars is checked
in the modified program, and then the spring works only when spars are moving over 30
m in surge direction.
8.6 Results and Discussion
The analysis results using the twobody spar model with a hawser connection and a
linear spring model between two spars are compared with the results of a single spar as
shown in Table 8.4. In the table, the sparspringspar model is considered an ideal case
so that the responses of both spars are identical. The corresponding case to this is the
sparhawserspar model with no interaction effect. These models show a good agreement
to each other. The results of the interaction case and the nointeraction case with no
cable reveal that the fluid interaction effect makes the rear side structure move a little
less in all directional motion except the sway motion. However, the effect makes the
sway motion of the lee side structure amplified a little. It means that the weather side
structure acts as a protector for the leeside structure.
150
When one hawser is used for the connection, it also forces the second body to
move in a more restricted way and less than the first body in the front side of the wave,
wind and current. The cable can be imagined to limit the motion of the second body,
since the hawser has the rigidity in the surge direction and so it will go to the opposite
direction against the second body movement when they are in an outofphase state. The
magnitude of the compensating reaction will vary according to the stiffness of the
hawser. To get some clues for the reason of the sudden increases in surge and yaw
motion RMS in the case of interaction effect with one hawser, the surge motion RAO is
illustrated in Figure 8.4.a. The heave motion RAO and the roll motion RAO are shown
in Figures 8.4.b and 8.4.c. As shown in Figure 8.4.a, the surge motion RAO for the two
body model has a similar trend to that for the singlebody model. As shown in Figures
8.4.b and 8.4.c, the heave and roll motion RAOs for the twobody model have similar
trends to those for the singlebody model. But, the surge drift force for the twobody
model has twice large than that for a single body model. It can make the differences
between the analysis results for the singlebody model and the twobody model in surge,
heave and roll dynamic motions. In Figure 8.5, the surge mean drift forces for a single
body and those for twobody by the pressure integration method are shown for
comparison purpose. In the figure, the twobody interaction effect can be seen.
151
Figure 8.4.a Comparison of the surge motion RAOs
Figure 8.4.b Comparison of the heave motion RAOs
Heave Motion RAOs
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
Freqiency (rad/s)
H
e
a
v
e
R
A
O
(
Z
/
A
)
Single SPAR
TwoBody (SPAR #1)
TwoBody (SPAR #2)
Surge Motion RAOs
0
1
2
3
4
5
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
Frequency (rad/s)
S
u
r
g
e
R
A
O
(
X
/
A
)
Single SPAR
TwoBody (SPAR #1)
TwoBody (SPAR #2)
152
Figure 8.4.c Comparison of the roll motion RAOs
Figure 8.5 Comparison of the surge drift force
Roll Motion RAOs
0.00
0.02
0.04
0.06
0.08
0.10
0.12
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
Freqiency (rad/s)
R
o
l
l
R
A
O
(
t
h
e
t
a
/
k
A
)
Single SPAR
TwoBody (SPAR #1)
TwoBody (SPAR #2)
Wave Drift Force in Xdirection
0.0E+00
2.0E+04
4.0E+04
6.0E+04
8.0E+04
1.0E+05
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Frequency (rad/s)
D
r
i
f
t
F
r
o
c
e
(
N
)
Single SPAR
TwoBody(at Body #1)
TwoBody(at Body #2)
153
Table 8.4 The analysis results for twobody model composed of two spars
SPAR 1 SPAR 2 SPAR 1 SPAR 2 SPAR 1 SPAR 2 SPAR 1 SPAR 2 SPAR 1 SPAR 2
mean 24.60 24.32 25.45 24.40 25.41 23.84 24.57 23.75 24.46 23.81 24.16
min. 31.54 30.63 33.66 30.71 33.59 30.73 31.69 29.73 31.45 30.48 30.95
max. 18.36 18.40 18.55 18.54 18.49 18.29 17.89 18.72 17.54 18.63 19.44
rms. 2.33 2.18 2.82 2.18 2.66 2.17 2.57 2.05 2.73 2.25 2.40
mean 6.36 6.46 5.71 6.46 5.80 6.44 5.91 5.86 5.56 6.37 6.47
min. 9.85 9.88 9.59 9.87 9.60 9.89 10.16 11.22 10.73 10.57 10.62
max. 2.78 2.91 1.91 2.90 1.91 3.18 1.47 0.04 0.20 3.73 3.78
rms. 1.40 1.40 1.64 1.40 1.53 1.40 1.60 1.96 2.03 1.50 1.49
mean 0.22 0.23 0.19 0.22 0.19 0.26 0.23 0.27 0.24 0.26 0.24
min. 0.54 0.49 0.46 0.49 0.60 0.31 0.24 0.19 0.65 0.15 0.26
max. 0.83 0.80 0.78 0.79 0.78 0.93 0.54 0.97 0.93 0.63 0.64
rms. 0.18 0.17 0.18 0.17 0.19 0.14 0.11 0.15 0.24 0.12 0.13
mean 0.67 0.67 0.64 0.67 0.64 0.67 0.34 0.65 0.63 0.67 0.68
min. 0.43 0.49 0.33 0.49 0.38 0.48 0.19 1.06 0.95 0.22 0.23
max. 1.82 1.83 1.70 1.83 1.74 1.85 1.48 2.78 1.89 1.50 1.53
rms. 0.43 0.45 0.39 0.45 0.41 0.44 0.34 0.57 0.39 0.38 0.39
mean 2.17 2.14 2.27 2.16 2.26 2.07 2.17 2.04 2.16 2.19 2.01
min. 6.54 6.31 6.36 6.31 6.73 6.04 4.57 5.87 4.46 5.23 5.03
max. 2.00 2.00 1.52 1.96 1.56 1.60 0.11 1.36 0.49 0.05 0.28
rms. 1.19 1.16 1.16 1.16 1.19 1.09 0.69 1.01 0.67 0.96 0.98
mean 0.05 0.04 0.05 0.04 0.04 0.04 0.04 0.13 0.09 0.04 0.04
min. 0.04 0.04 0.04 0.05 0.05 0.05 0.15 6.94 3.72 0.07 0.04
max. 0.16 0.15 0.17 0.15 0.16 0.16 0.27 7.05 3.85 0.18 0.16
rms. 0.03 0.03 0.04 0.03 0.04 0.03 0.09 2.85 1.48 0.05 0.04
mean 16,339 16,070 17,152 16,162 17,071 15,768 16,350 15,672 16,279 15,784 16,024
min. 10,587 10,678 11,374 10,686 10,629 11,423 12,196 11,117 12,170 12,173 12,092
max. 27,045 25,223 29,225 26,377 29,717 24,792 24,711 23,876 24,079 23,987 25,040
rms. 2,421 2,260 2,958 2,259 2,839 2,003 2,095 1,882 2,022 1,958 2,090
mean 9,807 9,815 9,710 9,823 9,723 9,823 9,745 9,743 9,702 9,817 9,831
min. 9,280 9,304 9,132 9,304 9,131 9,322 9,160 8,937 8,821 9,431 9,441
max. 10,389 10,393 10,375 10,391 10,377 10,382 10,436 10,629 10,597 10,568 10,578
rms. 215 237 245 215 228 215 238 292 290 236 236
mean 7,207 7,222 7,165 7,216 7,168 7,244 7,220 7,251 7,222 7,246 7,232
min. 6,093 6,382 6,214 6,152 6,217 6,113 6,778 6,333 6,769 6,553 6,638
max. 7,871 7,863 7,759 7,859 7,833 7,823 7,633 7,964 7,619 7,660 7,669
rms. 224 218 235 215 229 203 136 193 132 164 161
mean 8,356 8,354 8,412 8,348 8,403 8,351 8,395 8,403 8,426 8,356 8,348
min. 8,081 8,085 8,095 8,086 8,094 8,088 8,072 7,996 8,033 8,021 8,018
max. 8,678 8,658 8,774 8,658 8,774 8,650 8,790 8,978 9,119 8,579 8,573
rms. 116 125 137 115 128 115 133 166 174 121 120
mean 45,368 45,368 45,369 45,368 45,369 45,368 45,368 45,368 45,368 45,368 45,368
min. 45,360 45,360 45,360 45,360 45,360 45,360 45,361 45,360 45,361 45,610 45,361
max. 45,392 45,388 45,393 45,389 45,393 45,385 45,380 45,392 45,381 45,380 45,850
rms. 4 3 4 4 4 4 3 4 3 3 4
mean 66 66
min. 31 29
max. 171 207
rms. 21 25
Remarks:
pitch
(deg)
Riser
(kN)
yaw
(deg)
Hawser
(kN)
Mooring
line #1
(kN)
Mooring
line #2
(kN)
Mooring
line #3
(kN)
Mooring
line #4
(kN)
Single
SPAR
roll
(deg)
surge
(m)
sway
(m)
heave
(m)
Body Motion
Line Tension
w/o hawser with hawser w/o hawser
SPAR+SPRING+SPAR
2)
with hawser
SPAR+SPAR
1)
w/o interaction with interaction
1) Both SPARs have 4 equivalent mooring lines and 1 equivalent central riser.
2) A linear spring of the same stiffness as the hawser is put directly in the system stiffness matrix.
with a linear spring
w/o interaction
154
8.7 Summary and Conclusions
The multiple body interaction effects on the twobody model of two spars due to
the hawser connection and the hydrodynamic interaction effects are investigated by
comparative study using two numerical models.
When a linear spring is used, the results must be an ideal case. So, the statistical
results of the motions of two bodies are shown to be identical. With comparing this, the
results of the hawser connection model make the two bodies move a little differently. It
shows that the hawser acts as a compensator for the second body in the lee side. When
the second body tends to move out of phase against the first body motion, it makes the
second body move to the opposite direction. Therefore, the second body will be able to
move within a certain range.
The hydrodynamic interaction effect is exhibited well in the six DOF motions as
the motions of the second body, except the sway motions are a little bit smaller than
those of the other. It is why the flow route of the external forces of wind, wave and
current is restricted by the protection effect of the front structure. However, for the sway
motion, it is hard to say that the second body will move less that the first body. On the
whole point of view, the fluid interaction effect is clearly illustrated in the leeside
structure, and the front structure acts as a protector for the rear structure when the
environmental loads are applied to the first structure collinearly with the direction of the
body connection.
155
CHAPTER IX
CASE STUDY 5:
DYNAMIC COUPLED ANALYSES FOR TWOBODY SYSTEM COMPOSED
OF AN FPSOFPSO AND AN FPSOSHUTTLE TANKER
9.1 Introduction
In this chapter, an FPSOFPSO and an FPSOShuttle tanker are taken as the
multiplebody models for the verification of the program (WINPOSTMULT) for the
dynamic coupled analysis of the multiplebody floating platforms, and the results are
compared with the exact solution using a twomassspring model. An FPSOFPSO
model consists of two identical FPSOs. The other twobody model is composed of an
FPSO and a shuttle tanker. The conventional tandem moorings have been used for the
multiplebody connections in many cases of the operation of offloading in the sea. For
the multiplebody model of the FPSOshuttle tanker, the tandem mooring is considered
to investigate the interaction effect. The simplified massspring model will give a
compatible result to judge the validity of the multiplebody program.
In this study, the interaction characteristics for the tandemmoored vessels are
calculated in regular waves at several frequencies by using WAMIT. The body motions
and line tensions are mainly reviewed with the numerical calculations performed by
WINPOSTMULT, the dynamic coupled analysis program for multiplebody platforms.
The coupled analysis results for the model of two identical FPSOs by the WINPOST
MULT program are compared with the exact solution for the twomassspring model.
156
From this study, the effect of the hawser to connect two structures is also specified. For
this verification, models both with a hawser and without a hawser are made and analyzed.
The interaction effect is studied as well for this model.
For the mooring system, a tandem mooring is taken into account. The tandem
mooring has been used for many years. The distance of the tandem mooring system is
taken as 30 meters, which is the same as in the previous chapter.
9.2 Particulars of Models and Mooring Arrangements
The main particulars, including the principle data of spar, are listed in Table 9.1.
The main particulars and dimensions of the shuttle tanker are taken as the same as the
FPSO’s. The arrangement of the tandem is shown in Figure 9.1. The water depth is
6,000 ft (1828.8 m). The distance between the two FPSOs in the tandem mooring is
taken as 30 meters. The original FPSO studied in Chapter V has 12 taut mooring lines
and 13 steel catenary risers(SCR). Here, for simplification, they are equivalently
combined as 4 groups for mooring lines and 1 group for risers. Each mooring line group
has 3 legs, and one riser group is composed of all (13) risers. The riser group is
centralized on the geometrical center of the turret. The configuration for the mooring of
the equivalent mooring lines is shown in Figure 9.2. The mooring lines are fixed at the
sea floor. The WAMIT program is used for the calculation of the hydrodynamic
coefficients of the vessels. The validity of the numerical modeling was already proven in
the previous chapters by the static offset test and free decay tests. The numerical models
and the particulars of the mooring systems are the same as the FPSO’s reviewed in
157
Chapter V. The hawser connecting the two FPSOs and the FPSOShuttle tanker has the
stiffness of 1/100 of the mooring stiffness and the pretension of 1/10 of the mooring
pretension. Main particulars of the mooring systems are summarized in Table 9.2.
Table 9.1 Main particulars of the turret moored FPSO
Description Symbol Unit Quantity
Production level bpd 120,000
Storage bbls 1,440,000
Vessel size kDWT 200
Length between perpendicular Lpp m 310.0
Breadth B m 47.17
Depth H m 28.04
Draft (in full load) T m 18.09
Diaplacement (in full load) MT 240,869
Lengthbeam ratio L/B 6.57
Beamdraft ratio B/T 2.5
Block coefficient Cb 0.85
Center of buoyancy forward section 10 FB m 6.6
Water plane area A m
2
13,400
Water plane coefficient Cw 0.9164
Center of water plane area forward section 10 FA m 1.0
Center of gravity above keel KG m 13.32
Transverse metacentric height MGt m 5.78
Longitudinal metacentric height MGl m 403.83
Roll raius of gyration in air R
xx
m 14.77
Pitch raius of gyration in air R
yy
m 77.47
Yaw radius of gyration in air R
ζζ
m 79.30
Frontal wind area Af m
2
1,012
Transverse wind area Ab m
2
3,772
Turret in center line behind Fpp (20.5 % Lpp) Xtur m 63.55
Turret elevation below tanker base Ztur m 1.52
Turret diameter m 15.85
158
Figure 9.1 Configuration of the mooring systems (Tandem mooring system)
Table 9.2 Main particulars of the mooring systems
310.0 m
30.0 m
FPSO 1 FPSO 2 or Shuttle Tanker Wave
Wind
Current
Description Unit Quantity
Pretension kN 1,201
Number of lines 4*3
Degrees between 3 lines deg 5
Length of mooring line m 2,087.9
Radius of location of chain stoppers on turn table m 7.0
Length at anchor point m 914.4
Diameter mm 88.9
Weight in air kg/m 164.9
Weight in water kg/m 143.4
Stiffness, AE kN 794,841
Mean breaking load, MBL kN 6,515
Length m 1127.8
Diameter mm 107.9
Weight in air kg/m 42.0
Weight in water kg/m 35.7
Stiffness, AE kN 690,168
Mean breaking load, MBL kN 6,421
Length m 45.7
Diameter mm 88.9
Weight in air kg/m 164.9
Weight in water kg/m 143.4
Stiffness, AE kN 794,841
Mean breaking load, MBL kN 6,515
Segment 1 (ground position): chain
Segment 2: chain
Segment 3 (hangoff position): chain
159
Figure 9.2 Configuration of the arrangement of the mooring line groups
#3
#2
#1
#7
#8
#9
#10#11#12
#4 #5 #6
NORTH
EAST
Equiv. #3 Equiv. #1
Equiv. #4
Equiv. #2
160
9.3 Environmental Conditions
The environmental conditions correspond to the 100year storm conditions in GoM
and the sea condition of West Africa. The 100year storm conditions are used in the case
of tandem moored vessels of the two body model of an FPSO and an FPSO. For the
wind force, API RP 2T is referred to obtain the wind velocity spectrum. For the wave
force, JONSWAP spectrum is used. The wave frequencies are taken account of the range
from 0.5 rad/s to 1.2 rad/s. The wave is calculated at every frequency, dividing the range
by 100 intervals, and it is summed up with a random phase at every time. The current
velocity is 1.0668 m/s at the free surface, and it is reduced as 0.0914 m/s at the sea floor.
It varies linearly to the sea floor. The environmental conditions at GOM and at the west
Africa sea are summarized in Tables 9.3.a and 9.3.b, respectively. The incident wave
heading in hurricane conditions is
o
180 when the xcoordinate is set to the East and y
axis is set to the North.
The west Africa sea conditions are used for the twobody model of an FPSO and a
shuttle tanker. The API wind velocity spectrum is also used, but the wind speed is slower
than that in the 100yr. storm condition. The current speed in the West Africa is less than
that in GoM. The reason that the mild condition is taken for the FPSOShuttle tanker
model is that the tandem mooring system for transferring oil or gas from the FPSO to the
shuttle tanker in the real open sea has been tried in a rather mild sea condition for the
safety. The wave heading of this condition is
o
180 when the xcoordinate is set to the
East and yaxis is set to the North.
161
Table 9.3.a Environmental conditions (100year storm condition at GOM)
Table 9.3.b Environmental conditions (west Africa sea condition)
Description Unit Quantity
Significant wave height, Hs m 12.19
Peak period, Tp sec 14.0
Wave spectrum
Direction deg
180
1)
Velocity m/s 41.12 m/s @ 10m
Spectrum
Direction deg
210
1)
Profile
at free surface (0 m) m/s 1.0668
at 60.96 m m/s 1.0668
at 91.44 m m/s 0.0914
on the sea bottom m/s 0.0914
Direction deg
150
1)
Remark: 1) The angle is measured counterclockwise from the xaxis (the East).
Wind
Current
Wave
JONSWAP ( γ =2.5)
API RP 2T
Description Unit Quantity
Significant wave height, Hs m 2.70
Peak period, Tp sec 16.5
Wave spectrum
Direction deg
180
1)
Velocity m/s 5.0 m/s @ 10m
Spectrum
Direction deg
210
1)
Profile
at free surface (0 m) m/s 0.150
at 60.96 m m/s 0.150
at 91.44 m m/s 0.050
on the sea bottom m/s 0.050
Direction deg
150
1)
Remark: 1) The angle is measured counterclockwise from the xaxis (the East).
Current
Wave
JONSWAP ( γ =6.0)
Wind
API RP 2AWSD
162
9.4 Calculation of Hydrodynamic Coefficients Using WAMIT
The hydrodynamic coefficients are calculated by WAMIT. For the twobody
analysis, the wave force coefficients and wave drift force coefficients are calculated. The
hydrodynamic coefficients of added mass, wave damping and linear transfer function
(LTF) of diffraction potential force are calculated by WAMT. In Figure 9.3, the model
for the wave force coefficients is shown. The modeling is made only for the port side,
and the number of elements is 1684.
A turretmoored FPSO has been designed to weathervane in the sea so that the
mooring lines and risers are only connected at the bottom of turret. Under the
circumstances of applying the environmental conditions associated with wave, wind and
current load, it will pursue the dynamical equilibrium position corresponding to the
neutral location for the sum of the environmental loads to be zero and trace the path by
itself. After that, she will move and rotate freely. For a twobody model composed of
FPSO and FPSO, the mooring lines and risers are connected as what they are, and the
100year storm conditions at GoM are applied. But, for a twobody model composed of
FPSO and a shuttle tanker, the mooring lines and risers are installed only for FPSO, and
the shuttle tanker has no mooring line and riser. FPSO and the shuttle tanker are
connected with one hawser. For FPSO and shuttle tanker model, the West Africa sea
condition is applied. It is well known that the range of yaw angle in which she may
move in the 100year storm condition will be about 10~20 degrees. Accordingly, the
hydrodynamic coefficients at every angle should be calculated for the dynamic analysis.
However, in the timedomain simulation, it is not practical to calculate the coefficients at
163
every time step. In this study, at every 5degree interval, the coefficients are calculated
prior to the coupled analysis. So, when the coupled analysis of the body and the mooring
system is performed, at every time step the yaw angle is checked. If the yaw angle is
beyond 5 degrees from the starting position, the other coefficients are read from the pre
made files.
(a) A single–body FPSO model
(b) Twobody model of FPSO and FPSO ( or Shuttle tanker) in tandem
arrangement
Figure 9.3 Configuration of singlebody, twobody models and the mooring system
164
(c) Configuration of moorings for twobody model of FPSO and FPSO
(d) Configuration of moorings for twobody model of FPSO and Shuttle tanker
Figure 9.3 Continued
9.5 TwoMassSpring Modeling
The twomassspring model is devised to get an exact solution for the idealized
twobody FPSO model and is used for verifying the numerical analysis results by the
WINPOSTMULT program. The idealized model is shown in Figure 9.4. The
environmental loads are calculated using Morison’s equation for the wind and current
forces and the JONSWAP spectrum formula for the wave force. The masses are
FPSO #1 FPSO #2
SEA BED
(Tandem Arrangement)
FPSO Shuttle Tanker
SEA BED
(Tandem Arrangement)
165
determined to add the FPSO body mass and the added mass at around surge natural
frequency. Spring constants are calculated by considering the total top tension of the
mooring lines and risers in the horizontal direction. The hawser stiffness can be directly
converted to the linear spring in the middle of the idealized model.
Figure 9.4 Twomassspring model
The wind force in xdirection,
xw
F , is obtained from Morison’s formula and
OCIMF wind coefficient as:
2
2
1
w T w xw xw
V A C F ρ = (9.1)
where
xw
C is the wind force coefficient that can be read from the OCIMF document,
w
ρ
is the water density,
T
A denotes the projected area in the lateral direction of the vessel
against wind, and
w
V is the wind velocity. The wind force by API RP 2T, ) 1 (
ww
F ,
represents the force per unit area in the normal direction to the wind blowing, and is
given by:
2
2
1
) 1 (
w w ww
V F ρ = (9.2)
M
1
M
2
K
1
K
2
K
3
F
1
F
2
X
1
X
2
166
Here, in this study, the unit wind force, ) 1 (
ww
F , is calculated by a separate program, and
the resultant wind force is computed in the WINPOST program, since the force varies
according to the wind blowing direction. In WINPOST, the yaw angle of the body at
every time step is checked, and the wind force coefficient is interpolated by using the
reading data from the OCIMF document.
T
A is given by a user as an input data. In y
direction, the wind force is obtained in the same way by the following formula:
) 1 (
ww L yw yw
F A C F = (9.3)
where
yw
C is the wind force coefficient in ydirection obtained from the OCIMF
document, and
L
A denotes the projected area in the longitudinal direction to be normal
to the wind. As the initial wind direction is considered to be
o
210 counterclockwise
from the xaxis (true East), the coefficients of
xw
C and
yw
C are evaluated as 0.73 and
0.30, respectively, in the full load condition.
The current forces,
xc
F in xdirection and
xc
F in ydirection, are also calculated
from Morison’s formula as follows:
In xdirection: T L V C F
pp c c xc xc
2
2
1
ρ = (9.4)
In ydirection: T L V C F
pp c c yc yc
2
2
1
ρ = (9.5)
Where
pp
L and T are the same as in Table 9.1,
c
ρ is the water density, and
c
V is the
current velocity, and here current speed is used at the free surface. The current
167
coefficients,
xc
C and
yc
C are evaluated as 0.024 and 0.922, respectively, by considering
the initial current direction of
o
150 from the xaxis counterclockwise.
The formula of the JONSWAP wave spectrum was written in Chapter V
(equation (5.1)). If the significant wave height,
s
H , the peak period,
p
T , and
overshooting parameter, γ , are taken in Tables 9.3.a and 9.3.b, the wave can be
estimated at any time with random phases.
( )
∑
+ =
j
j i j j i
t A t F φ ω ω
φ
cos ) ( ) ( (9.6)
where i and j are the indices for representing the time instant and the frequency of any
wave component,
j
ω is the frequency of the incident wave component j , ) (
j
A ω is the
wave amplitude, and
j
φ is the random phase between wave components. The total force
is determined as the linear sum of the equation (9.2) ~ (9.6) as:
φ
F F F t F t F
c w
+ + = = ) ( ) (
2 1
(9.7)
where ) (
1
t F and ) (
1
t F are the applied forces to the mass
1
M and
2
M in the idealized
model, and
1
M and
2
M represent the virtual masses made of the mass weights and the
added masses of the FPSOs.
The body mass and stiffness are obtained by considering the mass weight of
FPSO, m, the added mass,
a
m , and the line top tension as follows:
a
m m M M + = =
2 1
(9.8)
risers and lines mooring of stiffness
3 1
= = K K (9.9)
hawser the of stiffness
2
= K (9.10)
168
wind
wave
velo
disp
current
[time, wf]
Wind Force2
[time, wf]
Wind Force 1
f2_wave
Wave Force 2
f1_wave
Wave Force 1
forces2
forces1
t
To Workspace1
res
x' = Ax+Bu
y = Cx+Du
StateSpace
Mux
Mux5
Mux
Mux4
Mux
Mux3
Mux
Mux2
Mux
Mux1
Mux
Mux
K
Gain3
K
Gain2
1
Gain1
1
Gain
F2
F1
Demux
Demux1
emu
Demux
f(u)
Current 2
f(u)
Current 1
Clock
Table 9.4 The system parameters for twomassspring model
Figure 9.5 The diagram of the time simulation in SIMULINK of MATLAB
ITEM Symbol Unit Magnitude
Added mass m
a
kg 1.466E+07
FPSO weight in mass m kg 2.397E+08
Mass of FPSO #1 M
1
kg 2.543E+08
Mass of FPSO #2 M
2
kg 2.543E+08
Stiffness of mooring #1 K
1
N/m 2.389E+05
Stiffness of hawser K
2
N/m 1.868E+03
Stiffness of mooring #2 K
3
N/m 2.389E+05
Natural period (Mode #1) sec 16.34
(Mode #2) sec 205.02
169
The calculated results to get the idealized twomassspring model are summarized in
Table 9.4. For the validity of the model data, the eigenvalues are checked using
MATLAB. The time simulation for the massspring model is performed using
MATLAB. The calculation diagram in MATLAB is depicted in Figure 9.5.
(a) The displacements at mass #1 and #2 of the massspring model by MATLAB
(b) The surge motion of FPSO+FPSO model by WINPOSTMULT
(without the interaction effect)
Figure 9.6 The surge motion of the FPSO and FPSO model by MATLAB for mass
spring model and by WINPOSTMULT for twobody model
Timesimulation results for FPSO+FPSO model
(without the interaction effect)
40.0
30.0
20.0
10.0
0.0
10.0
20.0
30.0
0 500 1000 1500 2000 2500 3000 3500 4000
Time (sec)
S
u
r
g
e
m
o
t
i
o
n
(
m
)
FPSO #1
FPSO #2
TimeSimulation Result Using MassSpring Model
50.0
40.0
30.0
20.0
10.0
0.0
10.0
0 500 1000 1500 2000 2500 3000 3500 4000 4500
Time (sec)
D
i
s
p
l
a
c
e
m
e
n
t
(
m
)
Mass #1 Mass #2
170
(c) The surge motion of FPSO+FPSO model by WINPOSTMULT
(with the interaction effect by iteration method)
(d) The surge motion of FPSO+FPSO model by WINPOSTMULT
(with the interaction effect by combined method)
Figure 9.6 Continued
Table 9.5 Analysis results of massspring model: displacement at mass #1 and #2
(unit: m)
Timesimulation results for FPSO+FPSO model
(with the interaction effect)
50.0
40.0
30.0
20.0
10.0
0.0
10.0
20.0
0 500 1000 1500 2000 2500 3000 3500 4000
Time (sec)
S
u
r
g
e
m
o
t
i
o
n
(
m
)
FPSO #1
FPSO #2
Mean Min. Max. RMS
Mass #1 15.47 38.99 11.71 14.46
Mass #2 15.45 42.97 8.55 14.08
Timesimulation results for FPSO+FPSO model
(with the interaction effect by iteration method)
50.0
40.0
30.0
20.0
10.0
0.0
10.0
20.0
0 500 1000 1500 2000 2500 3000 3500 4000
Time (sec)
S
u
r
g
e
m
o
t
i
o
n
(
m
)
FPSO #1
FPSO #2
171
Table 9.6 Summary of the analysis results for two body FPSO+FPSO
FPSO 1 FPSO 2 FPSO 1 FPSO 2 FPSO 1 FPSO 2 FPSO 1 FPSO 2 FPSO 1 FPSO 2 FPSO 1 FPSO 2
mean 14.63 14.19 13.98 13.70 13.36 13.86 10.95 14.97 7.89 14.72 10.34 13.24 9.32
min. 35.57 34.51 33.36 37.45 37.55 33.15 20.09 34.78 22.53 34.38 24.64 36.30 21.49
max. 3.07 3.55 3.25 7.89 7.50 1.37 1.98 3.63 4.07 0.50 2.18 6.24 1.81
rms. 8.01 8.55 8.59 9.23 9.20 7.25 4.06 8.05 5.93 7.27 4.19 8.06 4.40
mean 4.41 4.59 4.19 3.65 4.06 3.76 4.07 1.81 1.43 4.56 3.34 3.23 3.51
min. 0.91 0.98 0.56 1.13 1.48 2.35 1.73 3.03 5.53 2.84 3.13 3.09 3.53
max. 12.59 13.93 10.74 8.77 10.88 11.43 12.23 7.04 10.94 13.89 9.82 11.61 14.41
rms. 2.68 2.87 2.47 1.42 1.88 2.81 3.33 2.17 3.61 2.98 2.96 2.11 3.01
mean 1.32 1.31 1.30 1.27 1.28 1.18 0.71 1.19 0.67 1.29 0.69 1.24 0.70
min. 9.58 10.30 9.95 9.44 9.65 8.68 3.29 8.26 3.22 9.43 3.41 10.42 3.83
max. 5.79 6.39 6.37 5.52 5.72 5.50 1.25 5.26 1.37 5.91 1.52 6.49 2.01
rms. 2.60 2.57 2.54 2.47 2.51 2.32 0.72 2.28 0.62 2.55 0.67 2.43 0.72
mean 0.00 0.01 0.01 0.00 0.00 0.01 0.07 0.04 0.02 0.03 0.00 0.00 0.04
min. 4.87 4.54 5.35 2.97 5.15 5.93 3.15 1.38 1.83 8.11 3.12 4.83 4.09
max. 4.70 4.36 5.66 2.95 5.08 6.02 2.91 1.54 1.57 7.53 2.85 4.67 3.20
rms. 1.50 1.45 1.34 0.90 1.34 1.76 0.83 0.37 0.53 2.42 0.66 1.38 0.82
mean 0.45 0.45 0.45 0.43 0.44 0.41 0.25 0.40 0.23 0.44 0.24 0.42 0.24
min. 3.12 3.44 3.55 3.36 3.45 2.82 0.79 2.89 0.78 3.09 0.87 3.40 0.97
max. 4.93 5.48 5.41 5.08 5.18 4.47 1.18 4.20 1.15 4.87 1.23 5.25 1.44
rms. 1.45 1.44 1.42 1.39 1.41 1.28 0.31 1.28 0.26 1.40 0.29 1.34 0.31
mean 9.52 9.85 8.65 6.56 8.46 12.92 18.75 1.82 12.48 11.47 18.59 9.33 16.67
min. 0.80 3.79 0.47 2.53 2.45 3.65 10.73 2.37 5.61 3.52 14.83 0.62 8.24
max. 17.85 17.23 16.14 11.49 13.57 21.87 26.46 5.53 16.86 20.19 21.72 16.77 23.20
rms. 4.08 2.82 3.61 1.56 2.29 5.18 4.07 1.99 2.41 4.26 1.68 3.43 2.75
mean 6,399 6,349 6,313 6,271 6,216 6,285 5,873 6,477 5,413 6,416 5,780 6,193 5,619
min. 3,516 3,480 3,373 3,041 3,025 4,001 4,369 3,543 3,634 3,859 4,312 3,330 3,802
max. 10,570 10,430 9,757 10,480 10,490 10,110 7,601 10,080 7,932 10,330 8,263 10,700 7,818
rms. 1,306 1,377 1,373 1,470 1,466 1,167 654 1,297 927 1,184 673 1,291 701
mean 3,537 3,506 3,553 3,617 3,565 3,621 3,642 3,872 3,994 3,512 3,728 3,679 3,710
min. 1,759 1,884 2,098 2,286 2,033 2,102 2,455 2,805 2,631 1,788 2,604 1,989 2,237
max. 4,768 4,889 4,968 4,783 4,734 4,685 4,672 5,350 5,098 5,040 4,784 4,923 4,792
rms. 488 496 460 383 409 473 440 435 500 500 405 427 405
mean 2,585 2,634 2,662 2,704 2,730 2,639 2,847 2,556 3,208 2,554 2,929 2,700 3,019
min. 570 535 608 558 530 785 1,868 622 1,798 693 1,754 668 1,828
max. 4,853 5,085 5,051 5,724 5,704 4,496 3,879 4,857 4,780 4,562 3,995 5,284 4,455
rms. 767 866 878 913 920 677 417 766 669 709 431 815 484
mean 4,765 4,809 4,751 4,667 4,728 4,701 4,796 4,411 4,419 4,803 4,691 4,609 4,711
min. 3,349 3,193 3,194 3,326 3,345 3,384 3,697 2,887 3,404 2,937 3,625 3,328 3,677
max. 6,906 7,073 6,613 6,224 6,704 6,747 6,335 5,550 5,900 7,231 5,956 6,580 6,598
rms. 561 591 542 430 483 563 513 462 534 619 452 492 462
mean 109,800 109,800 108,700 107,300 108,100 102,900 75,360 103,700 73,270 110,500 73,870 106,400 74,360
min. 0 0 0 0 0 0 0 0 0 0 0 0 0
max. 676,700 724,600 721,900 655,700 671,900 663,300 255,100 638,900 254,800 703,200 274,300 734,500 316,400
rms. 132,300 131,100 130,300 127,000 128,400 120,000 5,006 120,100 45,060 131,000 48,330 125,300 49,850
mean 101 101 102
min. 100 100 100
max. 103 104 106
rms. 0 1 1
Remarks: 1) Both FPSOs have 4 equivalent mooring lines and 1 equivalent central riser.
Body Motion
Line Tension
Riser
(kN)
Hawser
(kN)
Mooring
line #1
(kN)
Mooring
line #2
(kN)
Mooring
line #3
(kN)
Mooring
line #4
(kN)
Single
FPSO
with hawser
FPSO+FPSO
1)
w/o interaction
with interaction
(by combined method)
yaw
(deg)
surge
(m)
sway
(m)
heave
(m)
roll
(deg)
pitch
(deg)
w/o hawser with hawser w/o hawser
with interaction
(by iteration method)
w/o hawser with hawser
172
(a) The time simulation results of FPSO+shuttle tanker model
(without the interaction effect)
(b) The time simulation results of FPSO+shuttle tanker model by the iteration method
(with the interaction effect)
(c) The time simulation results of FPSO+shuttle tanker model by the combined method
(with the interaction effect)
Figure 9.7 The time simulation results of the FPSO and shuttle tanker model
Timesimulation of surge motion for FPSO+Shuttle Tanker
10.0
0.0
10.0
20.0
30.0
40.0
0 500 1000 1500 2000 2500 3000 3500 4000
Time (sec)
S
u
r
g
e
m
o
t
i
o
n
(
m
)
FPSO Shuttle Tanker
Time simulation of surge motion for FPSO+Shuttle Tanker
20.0
0.0
20.0
40.0
60.0
80.0
0 500 1000 1500 2000 2500 3000
Time (sec)
S
u
r
g
e
m
o
t
i
o
n
(
m
)
FPSO Shuttle Tanker
Time simulation of surge motion for FPSO+Shuttle Tanker
10.0
0.0
10.0
20.0
30.0
40.0
0 500 1000 1500 2000 2500 3000
Time (sec)
S
u
r
g
e
m
o
t
i
o
n
(
m
)
FPSO Shuttle Tanker
173
Table 9.7 Summary of the analysis results for the twobody FPSO+shuttle tanker
FPSO Shuttle FPSO Shuttle FPSO Shuttle
mean 0.46 0.91 21.72 0.67 16.86 0.39 17.51
min. 2.01 5.74 6.13 2.23 6.11 1.52 8.14
max. 0.81 1.81 54.15 0.80 33.16 0.41 24.26
rms. 0.51 1.54 17.69 0.62 8.10 0.35 5.09
mean 0.12 0.03 0.12 0.05 2.50 0.01 3.50
min. 0.65 0.79 8.57 1.16 2.59 1.26 2.81
max. 0.85 0.84 5.44 1.38 8.74 1.41 9.25
rms. 0.28 0.39 3.62 0.48 3.70 0.47 4.11
mean 0.60 0.60 0.77 0.60 0.77 0.60 0.77
min. 1.58 1.48 2.66 1.40 1.62 1.44 1.73
max. 0.43 0.27 4.19 0.23 3.34 0.28 3.41
rms. 0.27 0.26 1.15 0.27 0.86 0.26 0.87
mean 0.00 0.00 0.00 0.00 0.00 0.00 0.00
min. 0.47 0.26 0.66 0.11 0.23 0.32 0.34
max. 0.51 0.28 0.65 0.10 0.23 0.33 0.34
rms. 0.14 0.05 0.13 0.01 0.06 0.11 0.08
mean 0.21 0.21 0.27 0.21 0.27 0.21 0.27
min. 0.51 0.39 1.66 0.36 1.28 0.33 1.31
max. 0.97 0.84 1.16 0.81 0.69 0.79 0.73
rms. 0.20 0.20 0.48 0.20 0.34 0.19 0.35
mean 0.98 0.48 3.20 0.38 5.02 5.71 10.62
min. 1.21 2.99 2.46 2.56 0.62 7.67 4.50
max. 2.52 2.72 7.54 2.34 10.17 0.07 14.75
rms. 1.11 0.16 3.11 1.13 2.69 1.68 2.84
mean 4,268 4,339 4,298 4,257
min. 4,086 3,944 4,094 4,122
max. 4,487 5,050 4,509 4,428
rms. 74 232 89 51
mean 4,174 4,187 4,184 4,189
min. 3,974 4,018 3,946 3,965
max. 4,350 4,375 4,408 4,397
rms. 57 67 78 75
mean 4,115 4,051 4,086 4,126
min. 3,811 3,374 3,779 3,918
max. 4,367 4,508 4,375 4,353
rms. 93 225 104 74
mean 4,210 4,197 4,200 4,195
min. 4,041 4,019 3,967 3,991
max. 4,422 4,353 4,449 4,433
rms. 54 67 78 78
mean 69,550 69,530 69,490 69,560
min. 0 0 0 0
max. 164,900 150,600 146,600 151,300
rms. 24,730 24,170 24,410 23,730
mean 254 119 79
min. 5 6 6
max. 844 296 252
rms. 254 86 77
with interaction
by the combined
method
with hawser with hawser with hawser
2) The loading condition is changed for this calculation, which is intended to investigate the
difference with the results by three methods in a mild loading condition (West Africa sea
condition).The wind velocity is 10 m/s at 10 m height, the current speed is 0.15 m/s at free
surface, and the wave has Hs of 2.7 m, Tp of 16.5 sec, and gamma of 6.0.
Riser
(kN)
Hawser
(kN)
Single FPSO
Mooring
line #1
(kN)
Mooring
line #2
(kN)
Mooring
line #3
(kN)
roll
(deg)
Line Tension
Mooring
line #4
(kN)
FPSO+Shuttle Tanker
2)
w/o interaction
with interaction
by the iteration
method
pitch
(deg)
yaw
(deg)
Body Motion
surge
(m)
sway
(m)
heave
(m)
174
9.6 Results and Discussion
In Table 9.5, the statistics of the analysis results for the massspring model is
shown. The analysis results for the FPSO and FPSO model are summarized in Table 9.6
The two tables show that the statistical results are well matched with each other. In
Figure 9.6(a)~(d), the displacements in xdirection (surge motion) by the time simulation
analyses for the massspring model and the FPSO and FPSO model when the mooring is
in tandem arrangement are depicted. The hawser stiffness used for this analysis was
1/100
th
of the mooring stiffness, and the top tension of the hawser was taken as 1/10
th
of
the mooring line tension.
The surge motion amplitude for each case is very similar, so
that the validity of the program WINPOSTMULT for the twobody analysis with one
hawser is proved. However, whether the interaction effect is considered or not affects the
shape and the phase difference between surge motions of two bodies in the time
simulation. The time simulation results are shown for the purpose of comparison in
Figure 9.7.
In Table 9.7, the analysis cases for the twobody model of an FPSO and a shuttle
tanker are summarized for three different cases. The hawser stiffness used for this
analysis was 1/1000
th
of the mooring stiffness, and the top tension of the hawser was
taken as 1/10
th
of the mooring line tension. In the case of “no interaction”, the
hydrodynamic coefficients induced by wave, the body stiffness matrix and mass matrix
have only the terms for the single body, and the interaction terms are set to zero. That
means, in this case, the interaction effect between two vessels of the fluid and the
structures is not considered. In the case of the “with the interaction effect by iteration
175
method” for the twobody model, the selfcoupling terms in the hydrodynamic
coefficients, the twobody stiffness matrix and the twobody mass matrix are only
considered. Thus, the interaction terms between two bodies are set to zero. In the case of
the “with the interaction effect by the combined method”, the fully coupled matrices are
used for the analysis. The purpose of this study is to compare the analyzed results by the
developed program with the results produced by the methods used in the industry. The
program WINPOSTMULT has the kind function of performing the above three cases
by handling the system matrix or the hydrodynamic coefficient matrices. In Table 9.7, to
review the results of all cases can make some clues drawn about the hawser connection
effect and the hydrodynamic interaction effect between two bodies. In all motions at the
rear side vessel, the interaction and hawser effects are clearly illustrated. In the two
body model of the FPSO and shuttle tanker, the analysis results for the case of “with
interaction by the iteration method” give medium values among the results for the cases
of “with no interaction” and “with interaction by the combined method”. It means that it
is significant to consider the fully coupled interaction effect for the twobody analysis.
From Figures 9.8a through 9.10d, the time histories and the motion amplitude
spectra are shown for all analysis cases. To review the motion amplitude spectrum for
each case, the vessels have almost the same characteristics in their dynamic behaviors.
176
500 1000 1500 2000 2500 3000 3500 4000 4500 5000
10
5
0
5
time (sec)
S
u
r
g
e
(
m
)
1.932
7.197 −
surge1
i
4.595 10
3
×
500 t
i
500 1000 1500 2000 2500 3000 3500 4000 4500 5000
2
0
2
time (sec)
S
w
a
y
(
m
)
1.546
1.446 −
sway1
i
4.595 10
3
×
500 t
i
500 1000 1500 2000 2500 3000 3500 4000 4500 5000
2
1
0
1
time (sec)
H
e
a
v
e
(
m
)
0.423
1.574 −
heave1
i
4.595 10
3
×
500 t
i
Figure 9.8.a Time simulation for the two body model of the FPSO and shuttle
tanker (at body #1=FPSO; tandem; without interaction effect)
177
500 1000 1500 2000 2500 3000 3500 4000 4500 5000
0.5
0
0.5
time (sec)
R
o
l
l
(
d
e
g
)
0.434
0.485 −
roll1
i
4.595 10
3
×
500 t
i
Figure 9.8.a Continued
500 1000 1500 2000 2500 3000 3500 4000 4500 5000
1
0
1
time (sec)
P
i
t
c
h
(
d
e
g
)
0.968
0.508 −
pitch1
i
4.595 10
3
×
500 t
i
500 1000 1500 2000 2500 3000 3500 4000 4500 5000
10
5
0
5
time (sec)
Y
a
w
(
d
e
g
)
3.167
5.852 −
yaw1
i
4.595 10
3
×
500 t
i
178
500 1000 1500 2000 2500 3000 3500 4000 4500 5000
50
0
50
100
time (sec)
S
u
r
g
e
(
m
)
64.348
13.921 −
surge2
i
4.595 10
3
×
500 t
i
500 1000 1500 2000 2500 3000 3500 4000 4500 5000
10
0
10
20
time (sec)
S
w
a
y
(
m
)
11.648
8.595 −
sway2
i
4.595 10
3
×
500 t
i
500 1000 1500 2000 2500 3000 3500 4000 4500 5000
5
0
5
time (sec)
H
e
a
v
e
(
m
)
4.194
2.666 −
heave2
i
4.595 10
3
×
500 t
i
Figure 9.8.b Time simulation for the two body model of the FPSO and shuttle
tanker (at body #2=shuttle tanker; tandem; without interaction effect)
179
500 1000 1500 2000 2500 3000 3500 4000 4500 5000
1
0
1
time (sec)
R
o
l
l
(
d
e
g
)
0.81
0.813 −
roll2
i
4.595 10
3
×
500 t
i
500 1000 1500 2000 2500 3000 3500 4000 4500 5000
2
0
2
time (sec)
P
i
t
c
h
(
d
e
g
)
1.159
1.69 −
pitch2
i
4.59510
3
×
500 t
i
500 1000 1500 2000 2500 3000 3500 4000 4500 5000
10
0
10
20
time (sec)
Y
a
w
(
d
e
g
)
11.843
4.103 −
yaw2
i
4.59510
3
×
500 t
i
Figure 9.8.b Continued
180
0 0.2 0.4 0.6 0.8 1
0
0.5
1
frequency(rad/s)
S
u
r
e
g
A
m
p
l
i
t
u
d
e
(
m
) 0.663
9.4 10
4 −
×
Asp
j
1.01 0 freq
j
0 0.2 0.4 0.6 0.8 1
0
0.1
0.2
0.3
frequency(rad/s)
S
w
a
y
A
m
p
l
i
t
u
d
e
(
m
) 0.234
2.432 10
4 −
×
Asp
j
1.01 0 freq
j
0 0.2 0.4 0.6 0.8 1
0
0.05
0.1
frequency(rad/s)
H
e
a
v
e
A
m
p
l
i
t
u
d
e
(
m
)
0.056
4.134 10
5 −
×
Asp
j
1.01 0 freq
j
Figure 9.8.c Amplitude spectrum density curve of the motion responses
for the two body model of the FPSO and shuttle tanker
(at body #1=FPSO; tandem; without interaction effect)
181
0 0.2 0.4 0.6 0.8 1
0
0.01
0.02
0.03
frequency(rad/s)
R
o
l
l
A
m
p
l
i
t
u
d
e
(
d
e
g
) 0.022
3.977 10
5 −
×
Asp
j
1.01 0 freq
j
Figure 9.8.c Continued
0 0.2 0.4 0.6 0.8 1
0
0.02
0.04
0.06
frequency(rad/s)
P
i
t
c
h
A
m
p
l
i
t
u
d
e
(
d
e
g
)
0.041
2.322 10
5 −
×
Asp
j
1.01 0 freq
j
0 0.2 0.4 0.6 0.8 1
0
0.5
1
frequency(rad/s)
Y
a
w
A
m
p
l
i
t
u
d
e
(
d
e
g
)
0.921
6.074 10
4 −
×
Asp
j
1.01 0 freq
j
182
0 0.2 0.4 0.6 0.8 1
0
5
10
15
frequency(rad/s)
S
u
r
g
e
A
m
p
l
i
t
u
d
e
(
m
) 12.615
0.015
Asp
j
1.01 0 freq
j
Figure 9.8.d Amplitude spectrum density curve of the motion responses
for the two body model of the FPSO and shuttle tanker
(at body #2=shuttle tanker; tandem; without interaction effect)
0 0.2 0.4 0.6 0.8 1
0
1
2
3
frequency(rad/s)
S
w
a
y
A
m
p
l
i
t
u
d
e
(
m
) 2.163
1.61 10
3 −
×
Asp
j
1.01 0 freq
j
0 0.2 0.4 0.6 0.8 1
0
0.1
0.2
0.3
frequency(rad/s)
H
e
a
v
e
A
m
p
l
i
t
u
d
e
(
m
)
0.254
1.669 10
4 −
×
Asp
j
1.01 0 freq
j
183
0 0.2 0.4 0.6 0.8 1
0
0.02
0.04
0.06
frequency(rad/s)
R
o
l
l
A
m
p
l
i
t
u
d
e
(
d
e
g
) 0.049
2.585 10
5 −
×
Asp
j
1.01 0 freq
j
Figure 9.8.d Continued
0 0.2 0.4 0.6 0.8 1
0
0.05
0.1
0.15
frequency(rad/s)
P
i
t
c
h
A
m
p
l
i
t
u
d
e
(
d
e
g
)
0.109
6.176 10
5 −
×
Asp
j
1.01 0 freq
j
0 0.2 0.4 0.6 0.8 1
0
1
2
frequency(rad/s)
Y
a
w
A
m
p
l
i
t
u
d
e
(
d
e
g
)
1.69
1.423 10
3 −
×
Asp
j
1.01 0 freq
j
184
500 1000 1500 2000 2500 3000 3500 4000 4500 5000
4
2
0
2
time (sec)
S
u
r
g
e
(
m
)
0.798
2.229 −
surge1
i
4.595 10
3
×
500 t
i
Figure 9.9.a Time simulation the for two body model of the FPSO and shuttle
tanker (at body #1=FPSO; tandem; with interaction effect
by iteration method)
500 1000 1500 2000 2500 3000 3500 4000 4500 5000
2
0
2
time (sec)
S
w
a
y
(
m
)
1.377
1.162 −
sway1
i
4.595 10
3
×
500 t
i
500 1000 1500 2000 2500 3000 3500 4000 4500 5000
2
1
0
1
time (sec)
H
e
a
v
e
(
m
)
0.337
1.499 −
heave1
i
4.595 10
3
×
500 t
i
185
500 1000 1500 2000 2500 3000 3500 4000 4500 5000
0.5
0
0.5
time (sec)
R
o
l
l
(
d
e
g
)
0.392
0.423 −
roll1
i
4.595 10
3
×
500 t
i
500 1000 1500 2000 2500 3000 3500 4000 4500 5000
5
0
5
time (sec)
Y
a
w
(
d
e
g
)
2.338
4.502 −
yaw1
i
4.595 10
3
×
500 t
i
Figure 9.9.a Continued
500 1000 1500 2000 2500 3000 3500 4000 4500 5000
1
0
1
time (sec)
P
i
t
c
h
(
d
e
g
)
0.922
0.462 −
pitch1
i
4.595 10
3
×
500 t
i
186
500 1000 1500 2000 2500 3000 3500 4000 4500 5000
0
20
40
time (sec)
S
u
r
g
e
(
m
)
33.155
3.031
surge2
i
4.595 10
3
×
500 t
i
Figure 9.9.b Time simulation for the two body model of the FPSO and shuttle
tanker
(at body #2=shuttle tanker; tandem; with interaction effect by iteration method)
500 1000 1500 2000 2500 3000 3500 4000 4500 5000
10
0
10
20
time (sec)
S
w
a
y
(
m
)
11.883
2.808 −
sway2
i
4.595 10
3
×
500 t
i
500 1000 1500 2000 2500 3000 3500 4000 4500 5000
5
0
5
time (sec)
H
e
a
v
e
(
m
)
3.466
1.843 −
heave2
i
4.595 10
3
×
500 t
i
187
500 1000 1500 2000 2500 3000 3500 4000 4500 5000
0.5
0
0.5
time (sec)
R
o
l
l
(
d
e
g
)
0.285
0.307 −
roll2
i
4.595 10
3
×
500 t
i
Figure 9.9.b Continued
500 1000 1500 2000 2500 3000 3500 4000 4500 5000
2
1
0
1
time (sec)
P
i
t
c
h
(
d
e
g
)
0.763
1.329 −
pitch2
i
4.595 10
3
×
500 t
i
500 1000 1500 2000 2500 3000 3500 4000 4500 5000
0
5
10
15
time (sec)
Y
a
w
(
d
e
g
)
10.172
0.617
yaw2
i
4.595 10
3
×
500 t
i
188
0 0.2 0.4 0.6 0.8 1
0
0.1
0.2
0.3
frequency(rad/s)
S
u
r
e
g
A
m
p
l
i
t
u
d
e
(
m
) 0.227
1.285 10
4 −
×
Asp
j
1.01 0 freq
j
Figure 9.9.c Amplitude spectrum density curve of the motion responses
for the two body model of the FPSO and shuttle tanker
(at body #1=FPSO; tandem; with interaction effect by iteration method)
0 0.2 0.4 0.6 0.8 1
0
0.1
0.2
frequency(rad/s)
S
w
a
y
A
m
p
l
i
t
u
d
e
(
m
) 0.185
1.691 10
5 −
×
Asp
j
1.01 0 freq
j
0 0.2 0.4 0.6 0.8 1
0
0.05
0.1
frequency(rad/s)
H
e
a
v
e
A
m
p
l
i
t
u
d
e
(
m
)
0.055
4.255 10
5 −
×
Asp
j
1.01 0 freq
j
189
0 0.2 0.4 0.6 0.8 1
0
0.01
0.02
0.03
frequency(rad/s)
R
o
l
l
A
m
p
l
i
t
u
d
e
(
d
e
g
) 0.024
2.97 10
5 −
×
Asp
j
1.01 0 freq
j
Figure 9.9.c Continued
0 0.2 0.4 0.6 0.8 1
0
0.02
0.04
frequency(rad/s)
P
i
t
c
h
A
m
p
l
i
t
u
d
e
(
d
e
g
)
0.039
2.411 10
5 −
×
Asp
j
1.01 0 freq
j
0 0.2 0.4 0.6 0.8 1
0
0.5
1
frequency(rad/s)
Y
a
w
A
m
p
l
i
t
u
d
e
(
d
e
g
)
0.584
1.419 10
3 −
×
Asp
j
1.01 0 freq
j
190
0 0.2 0.4 0.6 0.8 1
0
2
4
frequency(rad/s)
S
u
r
g
e
A
m
p
l
i
t
u
d
e
(
m
) 3.876
1.069 10
4 −
×
Asp
j
1.01 0 freq
j
Figure 9.9.d Amplitude spectrum density curve of the motion responses
for the two body model of the FPSO and shuttle tanker
(at body #2=shuttle tanker; tandem; with interaction effect by
iteration method)
0 0.2 0.4 0.6 0.8 1
0
1
2
3
frequency(rad/s)
S
w
a
y
A
m
p
l
i
t
u
d
e
(
m
) 2.212
2.148 10
3 −
×
Asp
j
1.01 0 freq
j
0 0.2 0.4 0.6 0.8 1
0
0.2
0.4
frequency(rad/s)
H
e
a
v
e
A
m
p
l
i
t
u
d
e
(
m
)
0.272
2.976 10
4 −
×
Asp
j
1.01 0 freq
j
191
0 0.2 0.4 0.6 0.8 1
0
0.01
0.02
0.03
frequency(rad/s)
R
o
l
l
A
m
p
l
i
t
u
d
e
(
d
e
g
) 0.021
5.436 10
6 −
×
Asp
j
1.01 0 freq
j
Figure 9.9.d Continued
0 0.2 0.4 0.6 0.8 1
0
0.05
0.1
0.15
frequency(rad/s)
P
i
t
c
h
A
m
p
l
i
t
u
d
e
(
d
e
g
)
0.106
1.16 10
4 −
×
Asp
j
1.01 0 freq
j
0 0.2 0.4 0.6 0.8 1
0
1
2
frequency(rad/s)
Y
a
w
A
m
p
l
i
t
u
d
e
(
d
e
g
)
1.438
6.851 10
4 −
×
Asp
j
1.01 0 freq
j
192
500 1000 1500 2000 2500 3000 3500 4000 4500 5000
2
1
0
1
time (sec)
S
u
r
g
e
(
m
)
0.41
1.522 −
surge1
i
4.595 10
3
×
500 t
i
Figure 9.10.a Time simulation for the two body model of the FPSO and shuttle
tanker
(at body #1=FPSO; tandem; with interaction effect by combined method)
500 1000 1500 2000 2500 3000 3500 4000 4500 5000
2
0
2
time (sec)
S
w
a
y
(
m
)
1.408
1.256 −
sway1
i
4.595 10
3
×
500 t
i
500 1000 1500 2000 2500 3000 3500 4000 4500 5000
2
1
0
1
time (sec)
H
e
a
v
e
(
m
)
0.347
1.487 −
heave1
i
4.595 10
3
×
500 t
i
193
500 1000 1500 2000 2500 3000 3500 4000 4500 5000
0.5
0
0.5
time (sec)
R
o
l
l
(
d
e
g
)
0.42
0.429 −
roll1
i
4.595 10
3
×
500 t
i
Figure 9.10.a Continued
500 1000 1500 2000 2500 3000 3500 4000 4500 5000
1
0
1
time (sec)
P
i
t
c
h
(
d
e
g
)
0.938
0.449 −
pitch1
i
4.59510
3
×
500 t
i
500 1000 1500 2000 2500 3000 3500 4000 4500 5000
10
5
0
5
time (sec)
Y
a
w
(
d
e
g
)
1.635
7.674 −
yaw1
i
4.59510
3
×
500 t
i
194
500 1000 1500 2000 2500 3000 3500 4000 4500 5000
0
10
20
30
time (sec)
S
u
r
g
e
(
m
)
24.257
8.137
surge2
i
4.595 10
3
×
500 t
i
Figure 9.10.b Time simulation for the two body model of the FPSO and shuttle
tanker
(at body #2=shuttle tanker; tandem; with interaction effect by combined method)
500 1000 1500 2000 2500 3000 3500 4000 4500 5000
5
0
5
10
time (sec)
S
w
a
y
(
m
)
9.656
2.027 −
sway2
i
4.595 10
3
×
500 t
i
500 1000 1500 2000 2500 3000 3500 4000 4500 5000
5
0
5
time (sec)
H
e
a
v
e
(
m
)
3.535
1.934 −
heave2
i
4.595 10
3
×
500 t
i
195
500 1000 1500 2000 2500 3000 3500 4000 4500 5000
0
10
20
time (sec)
Y
a
w
(
d
e
g
)
14.745
0.813
yaw2
i
4.59510
3
×
500 t
i
Figure 9.10.b Continued
500 1000 1500 2000 2500 3000 3500 4000 4500 5000
0.5
0
0.5
time (sec)
R
o
l
l
(
d
e
g
)
0.344
0.336 −
roll2
i
4.59510
3
×
500 t
i
500 1000 1500 2000 2500 3000 3500 4000 4500 5000
2
1
0
1
time (sec)
P
i
t
c
h
(
d
e
g
)
0.797
1.356 −
pitch2
i
4.59510
3
×
500 t
i
196
0 0.2 0.4 0.6 0.8 1
0
0.05
0.1
frequency(rad/s)
S
u
r
e
g
A
m
p
l
i
t
u
d
e
(
m
) 0.095
1.389 10
4 −
×
Asp
j
1.01 0 freq
j
Figure 9.10.c Amplitude spectrum density curve of the motion responses
for the two body model of the FPSO and shuttle tanker
(at body #1=FPSO; tandem; with interaction effect by combined method)
0 0.2 0.4 0.6 0.8 1
0
0.1
0.2
frequency(rad/s)
S
w
a
y
A
m
p
l
i
t
u
d
e
(
m
) 0.183
9.657 10
6 −
×
Asp
j
1.01 0 freq
j
0 0.2 0.4 0.6 0.8 1
0
0.05
0.1
frequency(rad/s)
H
e
a
v
e
A
m
p
l
i
t
u
d
e
(
m
)
0.056
3.336 10
5 −
×
Asp
j
1.01 0 freq
j
197
0 0.2 0.4 0.6 0.8 1
0
0.02
0.04
frequency(rad/s)
R
o
l
l
A
m
p
l
i
t
u
d
e
(
d
e
g
) 0.033
2.608 10
5 −
×
Asp
j
1.01 0 freq
j
Figure 9.10.c Continued
0 0.2 0.4 0.6 0.8 1
0
0.02
0.04
frequency(rad/s)
P
i
t
c
h
A
m
p
l
i
t
u
d
e
(
d
e
g
)
0.038
4.769 10
5 −
×
Asp
j
1.01 0 freq
j
0 0.2 0.4 0.6 0.8 1
0
0.5
1
frequency(rad/s)
Y
a
w
A
m
p
l
i
t
u
d
e
(
d
e
g
)
0.878
1.507 10
3 −
×
Asp
j
1.01 0 freq
j
198
0 0.2 0.4 0.6 0.8 1
0
1
2
3
frequency(rad/s)
S
u
r
g
e
A
m
p
l
i
t
u
d
e
(
m
) 2.599
1.94 10
3 −
×
Asp
j
1.01 0 freq
j
Figure 9.10.d Amplitude spectrum density curve of the motion responses
for the two body model of the FPSO and shuttle tanker
(at body #2=shuttle tanker; tandem; with interaction effect by combined method)
0 0.2 0.4 0.6 0.8 1
0
0.5
1
1.5
frequency(rad/s)
S
w
a
y
A
m
p
l
i
t
u
d
e
(
m
) 1.038
1.01 10
3 −
×
Asp
j
1.01 0 freq
j
0 0.2 0.4 0.6 0.8 1
0
0.2
0.4
frequency(rad/s)
H
e
a
v
e
A
m
p
l
i
t
u
d
e
(
m
)
0.284
3.438 10
4 −
×
Asp
j
1.01 0 freq
j
199
0 0.2 0.4 0.6 0.8 1
0
0.01
0.02
frequency(rad/s)
R
o
l
l
A
m
p
l
i
t
u
d
e
(
d
e
g
) 0.016
2.48 10
6 −
×
Asp
j
1.01 0 freq
j
Figure 9.10.d Continued
0 0.2 0.4 0.6 0.8 1
0
0.05
0.1
0.15
frequency(rad/s)
P
i
t
c
h
A
m
p
l
i
t
u
d
e
(
d
e
g
)
0.111
1.342 10
4 −
×
Asp
j
1.01 0 freq
j
0 0.2 0.4 0.6 0.8 1
0
1
2
frequency(rad/s)
Y
a
w
A
m
p
l
i
t
u
d
e
(
d
e
g
)
1.374
3.589 10
4 −
×
Asp
j
1.01 0 freq
j
200
9.7 Summary and Conclusions
The hydrodynamic interaction effects and the hull/mooring/riser/hawser coupling
for the multiple body system are investigated by numerical simulations. A simplification
by the massspring model is also considered. An LNG FPSO and a shuttle tanker are
taken as a multiple body system, and the tandem mooring is considered. The distance
effects on motions and drift forces of two vessels are already reviewed in Chapter VII.
The coupling and interaction effects are studied using the twobody model of an FPSO
and a shuttle tanker.
The comparison of the analysis results for the FPSO and FPSO model and the
massspring model has the validity of the program WINPOSTMULT. The comparative
study of an FPSO and a shuttle tanker illustrates the importance of including the
interaction effect between multiple bodies.
201
CHAPTER X
CONCLUSIONS FOR ALL CASE STUDIES
WINPOST program was developed for the hull/mooring/riser coupled dynamic
analysis of floating structures, such as SPAR, TLP, and FPSO. In this study, the program
was extended to multiple body problems, including hydrodynamic interactions.
5 case studies are presented for the verification of the developed program
WINPOSTMULT. The first two cases are for single FPSOs. The first one is a turret
moored FPSO in full load or ballast condition. In the second case, the intermediate
loading conditions and the simulated results are compared with OTRC experiment. In
the OTRC experiment, several platform parameters are not clearly identified. Thus, the
missing parameters are deduced from the free decay test. Even though the adjustment is
made, there exist several uncertainties to be clarified. For example, the wind force,
current force and the truncated mooring lines with buoys and springs may well not
match with our numerical modeling. Despite the uncertainties mentioned, the trend of
the numerical simulations follows that of experimental results.
The third case is to review the hydrodynamic characteristics of twobody
interaction. For the twobody model, an FPSO and a shuttle tanker are selected. They are
moored in a tandem arrangement and a sidebyside arrangement. Both mooring systems
are considered for this study. The interaction effect is much stronger in the sidebyside
mooring system than in the tandem mooring system. For example, if the distance closes
to a half of the original distance, the motion RAOs double.
202
The fourth case is for the twobody analysis with two identical SPARs. For the
validity of this analysis, the connecting hawser is modeled as a spring. The spring
stiffness is directly input in the system matrix in the program. The spring is programmed
to work in taut state, but not to work in slack state. The analysis results using the
simplified massspring model and twospar model show a reasonable agreement with
each other.
For the verification of the twobody module of the program WINPOSTMULT,
several cases are considered, i.e., FPSOs with and without hawsers and an FPSO and a
shuttle tanker with and without hawser. To verify the results, the connecting hawser,
mooring lines and two FPSOs are modeled as a simple twomassspring system, and an
approximate solution is obtained. The environmental loads are calculated in a simplified
form to apply to the massspring model. These analyses are conducted for the tandem
mooring system. When multiple floated dynamics are solved, a typical approach in
offshore industry is one of them, either completely neglecting or partially including the
hydrodynamic interaction effects. The existing methods used in the industry are
reviewed with the more sophisticated WINPOSTMULT program, which includes the
full hydrodynamic interactions. From the analysis results, the conclusion is drawn that
the interaction effects of the twobody problem can be very important. The WINPOST
MULT program is proved to be a useful tool for solving multiplebody interaction
problems.
203
REFERENCES
API RP 2T 1997 Recommended Practice for Planning, Designing, and Consulating
Tension Leg Platforms. 2
nd
Edition, American Petroleum Institute, N.W., Washington
D.C.
Arcandra, T. 2001 Hull/Mooring/Riser Coupled Dynamic Analysis of a Deepwater
Floating Platform with Polyester Lines. Ph.D. Dissertation, Texas A&M University.
Arcandra, T., Nurtjahyo, P. & Kim, M.H. 2002 Hull/Mooring/Riser Coupled Analysis of
a TurretMoored FPSO 6000 ft: Comparison between Polyester and BuoysSteel
Mooring Lines. Proc. 11
th
Offshore Symposium The Texas Section of the Society of
Naval Architects and Marine Engineers, SNAME, 18.
Baar, J.J.M, Heyl, C.N. & Rodenbusch, G. 2000 Extreme Responses of Turret Moored
Tankers. Proc. Offshore Technology Conference, OTC 12147 [CDROM], Houston,
Texas.
Buchner, B, van Dijk, A. & de Wilde, J.J. 2001 Numerical MultipleBody Simulations
of SidebySide Moored to an FPSO. Proc. 11
th
Int. Offshore and Polar Eng. Conference,
ISOPE, 1, 343353.
Choi, Y.R. & Hong, S.Y. 2002 An Analysis of Hydrodynamic Interaction of Floating
MultiBody Using HigherOrder Boundary Element Method. Proc. 12
th
Int. Offshore
and Polar Eng. Conference, ISOPE, 3, 303308.
204
Dean, R.G. & Dalrymple, R.A. 1992 Water Wave Mechanics for Engineers and
Scientists. Advanced Series on Ocean Engineering. 2, World Scientific Press, Dover,
D.E.
Faltinsen, O.M. 1998 Sea Loads on Ships and Offshore Structures. The Cambridge
University Press, Cambridge.
Garrett, D.L. 1982 Dynamic Analysis of Slender Rods. J. of Energy Resources
Technology, Trans. of ASME, 104, 302307.
Garrison, C.J. 2000 An Efficient TimeDomain Analysis of Very Large MultiBody
Floating Structures. Proc. 10
th
Int. Offshore and Polar Eng. Conference, ISOPE, 1, 65
71.
Huijsmans, R.H.M., Pinkster, J.A. & de Wilde, J.J. 2001 Diffraction and Radiation of
Waves Around SidebySide Moored Vessels. Proc. 11
th
Int. Offshore and Polar Eng.
Conference, ISOPE, 1, 406412.
Hong, S.Y., Kim, J.H., Kim, H.J. & Choi, Y.R. 2002 Experimental Study on Behavior of
Tandem and Sideby Side Moored Vessels. Proc. 12
th
Int. Offshore and Polar Eng.
Conference, ISOPE, 3, 841847.
Inoue, Y. & Islam, M.R. 2001 Effect of Viscous Roll Damping on Drift Forces of Multi
Body Floating System in Waves. Proc. 11
th
Int. Offshore and Polar Eng. Conference,
ISOPE, 1, 279285.
205
Kim, M.H. 1992 WINPOST V3.0 Users Manual. Dept. of Ocean Engineering, Texas
A&M University.
Kim, M.H., Arcandra, T. & Kim, Y.B. 2001a Validability of Spar Motion Analysis
against Various Design Methodologies/Parameters. Proc. 20
th
Offshore Mechanics and
Arctic Eng. Conference, OMAE01OFT1063 [CDROM], L.A., Califonia.
Kim, M.H., Arcandra, T. & Kim, Y.B. 2001b Validability of TLP Motion Analysis
against Various Design Methodologies/Parameters. Proc. 12
th
Int. Offshore and Polar
Eng. Conference, ISOPE, 3, 465473.
Kim, M.H. & Ran, Z. 1994 Response of an Articulated Tower in Waves and Currents.
International Journal of Offshore and Polar Engineering, 4, (4), 298231.
Kim, M.H., Ran, Z. & Zheng, W. 1999 Hull/Mooring/Riser Coupled Dynamic Analysis
of a Truss Spar in TimeDomain. Proc. 9
th
Int. Offshore and Polar Eng. Conference,
ISOPE, Brest, France, 1, 301308.
Kim, M.H. & Yue, D.K.P. 1989a The Complete SecondOrder Diffraction Solution for
an Axisymmetric Body. Part 1. Monochromatic Incident Waves. J. of Fluid Mechanics,
200, 235264.
Kim, M.H. & Yue, D.K.P. 1989b The Complete SecondOrder Diffraction Solution for
an Axisymmetric Body. Part 2. Bichromatic Incident Waves. J. of Fluid Mechanics, 211,
557593.
206
Lee, C.H. 1999 WAMIT User Manual. Dept. of Ocean Engineering, Massachusetts
Institute of Technology, Cambridge, M.A.
Ma, W., Lee, M.Y., Zou, J. & Huang, E. 2000 Deep Water Nonlinear Coupled Analysis
Tool. Proc. Offshore Technology Conference, OTC 12085 [CDROM], Houston, Texas.
Nordgen, R.P. 1974 On Computation of the Motions of Elastic Rods. ASME Journal of
Applied Mechanics, 777780.
OCIMF 1994 Prediction of Wind and Current Loads on VLCCs. 2
nd
Edition, Witherby &
Co. Ltd, London, England.
Pauling, J.R. & Webster, W.C. 1986 A Consistent LargeAmplitude Analysis of the
Coupled Response of TLP and Tendon System. Proc. 5
th
OMAE Conf., Tokyo, 3, 126
133.
Ran, Z. & Kim, M.H. 1997 Nonlinear Coupled Responses of a Tethered Spar Platform in
Waves. International Journal of Offshore and Polar Engineering, 7, (2), 2734.
Ran, Z., Kim, M.H. & Zheng, W. 1999 Coupled Dynamic Analysis of a Moored Spar in
Random Waves and Currents (TimeDomain versus FrequencyDomain Analysis).
Journal of Offshore Mechanics and Arctic Engineering, 121, (2), 194200.
Teigen, P. & Trulsen, K. 2001 Numerical Investigation of Nonlinear Wave Effects
Around Multiple Cylinders. Proc. 11
th
Int. Offshore and Polar Eng. Conference, ISOPE,
3, 369378.
207
Ward, E.G., Irani, M.B. & Johnson, R.P. 2001 The Behavior of a TankerBased FPSO in
Hurricane Waves, Winds, and Currents. Proc. 11
th
Int. Offshore and Polar Eng.
Conference, ISOPE, 4, 650653.
Wichers, J.E.W. 1988 A Simulation Model for a Single Point Moored Tanker. Ph.D.
Dissertation, Delft University of Technology, Delft, The Netherlands.
Wichers, J.E.W. & Develin, P.V. 2001 Effect of Coupling of Mooring Lines and Risers
on the Design Values for a Turret Moored FPSO in Deep Water of the Gulf of Mexico.
Proc. 11
th
Int. Offshore and Polar Eng. Conference, ISOPE, 3, 480487.
Wichers, J.E.W. & Ji, C. 2000a On the Coupling Term in the LowFrquency Viscous
Reaction Forces of Moored Tankers in Deep Water. Proc. Offshore Technology
Conference, OTC 12086 [CDROM], Houston, Texas.
Wichers, J.E.W & Ji, C. 2000b DeepStarCTR 4401 Theme Structure Benchmark
Analysis for Tanker Based FPSOGoM. Technical Rep. No. 156291OE, MARIN,
Wageningen, The Netherlands.
Wichers, J.E.W, Voogt, H.J., Roelofs, H.W. & Driessen, P.C.M. 2001 DeepStarCTR
4401 Benchmark Model Test. Technical Rep. No. 164171OB, MARIN, Wageningen,
The Netherlands.
208
VITA
YoungBok Kim was born in Incheon in the Republic of Korea on September 9,
1958. He graduated from Inha University with a Bachelor of Science degree in naval
architecture and ocean engineering in February 1981. After he served in the Korean
Army about for 10 months, he was employed by the Daewoo Ship Building and Heavy
Industry Co. Ltd. (DWSH) on Keoje Island, Korea. There he worked as a structural
engineer and also as a ship vibration analysis engineer. He was involved in ship design,
vibration analyses and measurements for newly built ships. After working for seven
years for DWSH, he moved to the Korean Register of Shipping (KR) in Seoul, Korea.
While he worked at KR, he entered the graduate school of Seoul National University in
1992. He majored in naval architecture and ocean engineering, and two years later he
received his Master of Science degree in February 1994. After that, he went abroad to
pursue the doctoral degree at Texas A&M University in January 1999. In May 2003, he
received his Ph.D. in the field of ocean engineering. He married DeockSeung Seo in
1983 and has two sons, Hayong and Harin. His permanent address is: 4596, Chowon
Villa 102, JeonminDong, YusungKu, Taejon, Republic of Korea, 305810.
DYNAMIC ANALYSIS OF MULTIPLEBODY FLOATING PLATFORMS COUPLED WITH MOORING LINES AND RISERS
A Dissertation by YOUNGBOK KIM
Submitted to the Office of Graduate Studies of Texas A&M University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY
Approved as to style and content by:
MooHyun Kim (CoChair of Committee)
Cheung H. Kim (CoChair of Committee)
Jun Zhang (Member)
Robert H. Stewart (Member)
Paul N. Roschke (Head of Department) May 2003 Major Subject: Ocean Engineering
iii ABSTRACT Dynamic Analysis of MultipleBody Floating Platforms Coupled with Mooring Lines and Risers. (May 2003) YoungBok Kim, B.S., Inha University; M.S., Seoul National University CoChairs of Advisory Committee: Dr. MooHyun Kim Dr. Cheung H. Kim
A computer program, WINPOSTMULT, is developed for the dynamic analysis of a multiplebody floating system coupled with mooring lines and risers in the presence of waves, winds and currents. The coupled dynamics program for a single platform is extended for analyzing multiplebody systems by including all the platforms, mooring lines and risers in a combined matrix equation in the time domain. Compared to the iteration method between multiple bodies, the combined matrix method can include the
6 N × 6 N full hydrodynamic interactions among N bodies. The floating platform is
modeled as a rigid body with six degrees of freedom. The first and secondorder wave forces, added mass coefficients, and radiation damping coefficients are calculated from the hydrodynamics program WAMIT for multiple bodies. Then, the time series of wave forces are generated in the time domain based on the twoterm Volterra model. The wind forces are separately generated from the input wind spectrum and wind force formula. The current is included in Morison’s drag force formula. In the case of FPSO, the wind and current forces are generated using the respective coefficients given in the OCIMF
iv data sheet. A finite element method is derived for the long elastic element of an arbitrary shape and material. This newly developed computer program is first applied to the system of a turretmoored FPSO and a shuttle tanker in tandem mooring. The dynamics of the turretmoored FPSO in waves, winds and currents are verified against independent computation and OTRC experiment. Then, the simulations for the FPSOshuttle system with a hawser connection are carried out and the results are compared with the simplified methods without considering or partially including hydrodynamic interactions.
for her support and encouragement during the period of this study. C. H. Dr. I would like to express my sincere gratitude to my advisors. Dr. M. E. Finally. and Dr. Stewart for serving as advisory committee members. B. I would like to thank my wife. Zhang and Dr. I also would like to thank Dr. Zhihuang Ran (Alex) and Dr. I deeply thank the sponsors for this support. I greatly appreciate Dr. H. Kim. H.v ACKNOWLEDGEMENTS This work was completed only because of the financial support of the OTRC and JIP (Joint Industry Project) for over four years. Kim and Dr. for their continuous encouragement and guidance during my studies. Arcandra Tahar for sharing their efforts to review the programming and to discuss the problem. Portis for supervising the procedure of the final defense as a GCR. Mercier for releasing the OTRC experiment data. R. This work could only be done under the merciful guidance and the tender love of God. DeockSeung Seo. R. I would like to devote this work to His Glory. . J.
…………….. 7 1.vi TABLE OF CONTENTS Page ABSTRACT ………………………………………………………………………… iii ACKNOWLEDGEMENTS ………………………………..4.6 Motions of the Floating Platform ……………………………………….2 Wave Theory …………………………………………………….…………….3 Hydrodynamic Forces ………………………………..…………….………….……………….1 Boundary Value Problem (BVP) of Surface Wave …….2..….2 Literature Review ………….3 Objective and Scope ……….…………….1 Introduction …………………………….1 The FirstOrder Hydrodynamic Forces and Moments ………….…….4 MultipleBody Interaction of Fluid …………….……….……………..2.…….1 Interpretation and Preparation of WAMIT Results and Wind/ Current Forces ………………………….…………………. 2.. v vi x LIST OF TABLES ………………………………………….2 SecondOrder Boundary Value Problem …….2.1 Wave Loads …………………………………….…………….1 FirstOrder Boundary Value Problem …………………….………….…………….….4.……….3 Comparative Studies …………………………….……………. 1 1. 2..……………….… 8 1. 2.………….……..… xiv CHAPTER I INTRODUCTION …………………………………….6. 2.…………………. 10 II DYNAMICS OF THE FLOATING PLATFORM ……….……………………………….1 Background………….…………. 2.…...………………… 2. 2.…….….3..………….5 Boundary Element Method …………………………………………….2 Morison’s Equation …………….………………… 1 1.………. LIST OF FIGURES …………………………………………. 2. 12 12 12 14 16 17 19 23 23 26 28 30 33 33 36 ....………..4.…………… 2.………..……………. 2..3..……..…………………..…….………. 7 1..…..3.4 Procedure …………………………………..3 Diffraction and Radiation Theory ……………. 3 1.………….2 Developing the Coupled Dynamic Program …………..………………………. 2. 2..3....…………………… TABLE OF CONTENTS ……………………………………. 12 2.6..2. 5 1.…………….2 The SecondOrder Hydrodynamic Forces and Moments …….2 Formulation of Surface Wave …………………. 2...2. 2.
2.. 5.…… 3.…….5 Formulation for Dynamic ProblemTime Domain Integration ……. 79 5.…………… 4.3.2 TimeDomain Analysis …………………………………………..6.6..... 3.. 4.5 Formulation for the Multiple Body System …………………….………………………… 4.vii CHAPTER 2. 5.6 Results and Discussion .2 Theory of the Rod ……………………………………………………… 3.1 Wave Force …………………………………….……… V Page 37 38 40 44 44 46 50 55 59 63 66 66 67 69 71 72 74 75 CASE STUDY 1: DYNAMIC ANALYSIS OF A TANKER BASED FPSO ………………………………………………………………….2 Design Premise Data of FPSO and Mooring Systems …….4 Formulation of Static Problem ………………………………………….3 Wind and Current Forces by OCIMF …………….1 Static Offset Test (in Calm Water without Current) ……….…………………………………… 3. 5... 4.3 Timedomain Simulation for Hurricane Condition ………..1 Introduction …………………………………………………………….4 Modeling of Connection between Lines and Seafloor …….….3 Finite Element Modeling ………………………………………………. 4.………………………………… 5.……………… 5.2 Freedecay Tests (in Calm Water without Current) …………….5 Time Domain Solution of the Platform Motions ……..3 Single Body Motion ………….………………………………… 2. 5.………….…………... 5...…….……………………….. 4.6 Modeling of the Seafloor ……………………………………………….7 Summary and Conclusions ……………………………………………..5 Coupled Analysis of FPSO ………….……………….1 Introduction ………………………………….3. 5. 5. 79 80 85 87 88 90 93 95 98 99 101 103 106 .2 Wind Force ………………………………………………………. 3. IV COUPLED ANALYSIS OF INTEGRATED PLATFORM AND MOORING SYSTEM …………………………………………………..…………………………….6.2 The Spring to Connect the Platform and the Mooring System…………..1 Introduction ………………………. 5.3 Modeling of Damper on the Connection …………..3 Environmental Data …………………………………………………… 5.…………………….4 Hydrodynamic Coefficients …………………………………………….…….. III DYNAMICS OF MOORING LINES AND RISERS …………….6.3..4 Multiple Body Motion ………………….2.………….…… 5.2.1 Static Analysis ………………………………………………….6. 4..6. 3.
3 Environmental Conditions …………………………………………… 7.3 Environmental Data …………………………………………………...1 Introduction ………………………………………………………….….5 Results and Discussion …….……….……………………………………….…………………….…. 8...……………….….1 Static Offset Test with Regenerated Model Data …………….. 108 6.5... 6....…………………. 6.………….…………..…..6 Summary and Conclusions …………………………………………..….………….4 Calculation of Hydrodynamic Coefficients Using WAMIT 1st and 2nd Order …………. VII 108 109 114 116 119 119 120 123 125 CASE STUDY 3: CALCULATION OF HYDRODYNAMIC COEFFICIENTS FOR TWO BODY SYSTEM OF FPSO AND SHUTTLE TANKER ……………………………………………….viii CHAPTER VI Page CASE STUDY 2: DYNAMIC ANALYSIS OF A TANKER BASED FPSO COMPARED WITH THE OTRC EXPERIMENT ……………….. 126 128 132 133 141 VIII CASE STUDY 4: DYNAMIC ANALYSIS FOR TWOBODY SYSTEM COMPOSED OF SPAR AND SPAR …………………….7 Summary and Conclusions ………………………………. 6.……...3 Environmental Conditions …………………………………………… 8.……….6 Results and Discussion ……….5..1 Introduction ………………………………………………………….. 6..4 Results and Discussion ………. 6.….. 6.…………….4 Regeneration of the Experimental Model …………………………. 142 8.……………………………… 7.. 8. 126 7.…….2 Particulars of Models and Arrangements for the Tests ……………… 7..1 Introduction ………………………………………………………….5 Summary and Conclusions ………………………………………….…….2 Particulars of Models and Arrangements for the Analyses ….2 Particulars of Models and Arrangements for the Analyses ……. 155 9.2 OTRC Experimental Results and Design Premise Data ………….…… 6.……………………..… 142 143 146 147 149 149 154 IX CASE STUDY 5: DYNAMIC ANALYSIS FOR TWOBODY SYSTEM COMPOSED OF AN FPSOFPSO AND AN FPSOSHUTTLE TANKER …………………………………………....……..……… 156 9..………..2 FreeDecay Test with Regenerated Model Data ….3 Time Simulation Results …………………………. 8. 8.3 Environmental Conditions …………………………………………..… 155 9.5 Linear Spring Modeling ……….1 Introduction …………………………. 8.…………… 7. 160 .5. 6..
………………….…… 201 REFERENCES ………………………..……….………. 162 9.5 TwoMassSpring Modeling ….…………………….ix CHAPTER Page 9..4 Calculation of Hydrodynamic Coefficients Using WAMIT ..... 164 9.……………… 203 VITA ………………………………….……. 174 9.………………….7 Summary and Conclusions …………………………………………… 200 X CONCLUSIONS FOR ALL CASE STUDIES ………………..6 Results and Discussion ……….……………… 208 .…………………….………………..……….…………………….
113 6. 5. 5.2 Roughmeshed numerical modeling for a LNG FPSO and a shuttle tanker …… 132 7. 5.5 Hull drag coefficients proposed by Wichers (1998 & 2001) ………………….4 JONSWAP wave spectrum ……………………………………………………. 1996) ……………………………… Page 46 82 84 85 88 89 94 95 97 5.6 Modeling of body surface of FPSO …………………………………………… 5.……………………………………………………. 6.2 Arrangement of mooring lines for turretmoored FPSO ……………………… .. 5.7 Modeling of body surface and free surface of the water ……………………… 5. 115 121 122 131 7...1 Configuration of the mooring system ………………………………………….. …………………………………. 6.1 The body plan and the isotropic view of FPSO 6. 7.2 Arrangement of the mooring lines for FPSO 6.000 ft.8 Hull drag damping coefficients (Wichers.x LIST OF FIGURES FIGURE 3.1 General arrangement and body plan of FPSO 6. 5.5 API wind spectrum …………………………………………………………….3 NPD wind spectrum curve ..000 ft..4 Comparison of the static offset test results …………………………………….000 ft ……………………….10 Freedecay test results for surge.3 Finemeshed numerical modeling for a LNG FPSO and a shuttle tanker ……… 132 .………. 102 110 6. heave and roll motions ………….3 Arrangement of the risers for FPSO 6..9 Static offset test results for surge motion …………………… ………………… 100 5.000 ft ……………………….1 Coordinate system of rod ………………………………………………………. ………………………… 5. 6.
.6 Longitudinal wave drift force of tandem moored vessels in the head sea …….. 140 8.4.3 Configuration of the modeling of a twobody spar …………………………… 8.1 Configuration of the mooring system and the environmental loads (Tandem arrangement.7 Longitudinal wave drift force of sidebyside moored vessels in the head sea … 137 7. 9.. 8. 135 7.3 Configuration of singlebody.5 Roll response operators of sidebyside moored vessels in the beam sea …….4.5 Comparison of the surge drift force ……………………………………………. d=30m)………………………………………………..4 Heave response operators of sidebyside moored vessels in the beam sea …… 134 7.xi FIGURE Page 7.1 Configuration of the mooring systems (Tandem mooring system)…………….. twobody models and the system ……………… 163 9.. 152 9.2 Configuration of the arrangement of the mooring line groups ………………… 159 9. 144 148 148 151 151 152 8.. 8.8 The distance effect on the longitudinal wave drift force for a twobody and a single body model in the head sea ………………………………………..a Comparison of the surge motion RAOs ……………………………………...b Comparison of the heave motion RAOs ……………………………………. 8.c Comparison of the roll motion RAOs ……………………………………….4. 138 7.… 139 7. 165 168 . 158 9..10 Lateral wave drift force of sidebyside moored vessels in the beam sea …….4 Twomassspring model ………………………………………………………. 8.2 Configuration of the modeling of a single spar ……………………………….9 Lateral wave drift force of sidebyside moored vessels in the head sea …….5 The diagram of the time simulation in SIMULINK of MATLAB …………….. 136 7.
184 9.a Time simulation for the two body model of the FPSO and shuttle tanker (at body #1=FPSO. with interaction effect by combined method) ………………………………………………………… 190 9.9. with interaction effect and by iteration method).a Page The surge motion of the FPSO and FPSO model by MATLAB for massspring model and by WINPOSTMULT for twobody model ……………………….7 9.xii FIGURE 9. with interaction effect by iteration method) …………………………………………………………. 169 The time simulation results of the FPSO and shuttle tanker model ………….. tandem. with interaction effect by iteration method) …… 188 9.9.. without interaction effect) ……………………. tandem.. without interaction effect) ……………..b Time simulation for the two body model of the FPSO and shuttle tanker (at body #2=shuttle tanker.9.9.8. without interaction effect) ……………………. 180 9. tandem.8.c Amplitude spectrum density curve of the motion responses for the two body model of FPSO and shuttle tanker (at body #1=FPSO.c Amplitude spectrum density curve of the motion responses for the two body model of the FPSO and shuttle tanker (at body #1=FPSO.d Amplitude spectrum density curve of the motion responses for the two body model of the FPSO and shuttle tanker (at body #2=shuttle tanker.b Time simulation for the two body model of the FPSO and shuttle tanker (at body #2=shuttle tanker..10. 172 Time simulation for the two body model of the FPSO and shuttle tanker (at body #1=FPSO. 182 9. tandem. 9. tandem.b Time simulation for the two body model of the FPSO and shuttle tanker (at body #2=shuttle tanker. tandem. tandem... without interaction effect) ….a Time simulation for the two body model of the FPSO and shuttle tanker (at body #1=FPSO.10. with interaction effect by combined method) … 192 9. tandem. with interaction effect by combined method) ……………………………………………………….8. tandem.d Amplitude spectrum density curve of the motion responses for the two body model of FPSO and shuttle tanker (at body #2=shuttle tanker. 194 . 186 9.6 9.. tandem.……….8.. 176 178 9.
10. with interaction effect by combined method) … 196 9..d Amplitude spectrum density curve of the motion responses for two body model of FPSO and shuttle tanker (at body #2=shuttle tanker. with interaction effect by combined method) ……………………………………………………….xiii FIGURE Page 9. tandem.10. 198 . tandem.c Amplitude spectrum density curve of the motion responses for the two body model of the FPSO and shuttle tanker (at body #1=FPSO.
112 6. 5. rope and wire for the OTRC FPSO …………………………………………………………………………………… 112 6.2 Main particulars of mooring systems for the OTRC FPSO ………………. 117 119 6..1 Main particulars of the turret moored FPSO 6. 5.……..000 ft …………………………. rope and polyester ………………….. 105 6.1 Main particulars of the turret moored for the OTRC FPSO ……………………...7 Environmental loading condition ……………………………………………… 5..4 Environmental loading condition for the OTRC FPSO ………………………… 114 6.xiv LIST OF TABLES TABLE 5.7 Comparison of the free decay test results ……………………………………… 122 .9 Damping from freedecay tests estimated from the first 4 peaks assuming linear damping ……………………………………………………… Page 81 83 83 84 84 85 86 103 103 5.11 The results of tensions on the mooring lines and risers ……………………. 5.10 Timedomain simulation results ……………….5 Hydrodynamic coefficients of risers …………………………………………. 104 5.2 Main particulars of mooring systems …………………………………………. 111 6.8 Natural periods from freedecay tests ………………………………………….3 Hydrodynamic coefficients of the chain.6 Reestimated data from WAMIT output and handcalculation ……………….3 Hydrodynamic coefficients of the chain. 5. 5... 5.6 Azimuth angles of risers bounded on the earth ……………………………….4 Main particulars of risers ……………………………………………………… 5. 6.5 WAMIT output and handcalculation …………………………………………...…………………………….
1 Main particulars of the turret moored FPSO …………………………………… 157 9..5 Analysis results of massspring model: displacement at mass #1 and #2 ……… 170 9..3. 131 8. 171 9.b Environmental conditions (west Africa sea condition) ………….2 Main particulars of the mooring systems ………………………………………..6 Summary of the analysis results for two body FPSO+FPSO ………………….1 Main particulars of two vessels ……………………………………………….3.. 130 7.3 Environmental conditions ……………………………………………………… 146 8.1 Main particulars of moored spar ……………………………………………….4 The analysis results for twobody model composed of two spars …………….7 Summary of the analysis results for the twobody FPSO+shuttle tanker ……….3 Comparison of the hydrodynamic coefficients obtained from the rough model and the fine models ……………………………………………………………. 129 7..a Environmental conditions (100year storm condition at GoM) ……………… 161 9.8 Comparison of time simulation results ………………………………………… 124 7.4 The system parameters for twomassspring model …………………………… 168 9. 173 .xv TABLE Page 6.. 144 145 8.2 Freedecay test results for a LNG FPSO and a shuttle tanker (heave and roll) ………………………………………………………………. 158 9.…………… 161 9.2 Particulars of the mooring systems ……………………………………………... 8. 153 9..
Neptune. and Marlin were fixed in position by means of the mooring lines or risers in 2. respectively.590 ft. and the tension leg platform(TLP). the truss spar is being considered more costeffective. The recent trend in the installation of floating structures shows the water depth getting deeper and deeper since the oil and gas fields are expedited and discovered in the deeper sea. 4.000 ft or more.1 Background Recently. In the case of TLPs. and Diana were installed in 2. These installations were made from 1996 to 1999. This means the more developed designs should be invented and studied realistically for the installation of the floating structures in deep water of 6. there were several built and installed in GoM. .000 ft of water depth. These structures include the shipshaped vessel called an FPSO(Floating Production Storage and Offloading Unit). the spar platform. the column stabilized semisubmergible platform. The last two types have been designed and installed in the Gulf of Mexico(GoM) for the last decade.800 ft to 4. Ursa. Floating structures are more attractive to the industrial companies This dissertation follows the style and the format of the Journal of Fluid Mechanics. floating structures have been invented and their installation has been attempted worldwide because of cost effectiveness.1 CHAPTER I INTRODUCTION 1.000 ft. In the case of spars. of which Auger.300 ft of water depth. 2. Nowadays. in an attempt to replace traditional fixed jacket platforms. Genesis. Mars.
they will have to be potentially attractive production systems in ultra deep water of the GOM. Nowadays... This results in reducing the heave response by decreasing the vertical wave load and shifting the heave natural frequency in the low part far apart from the wavedominant frequency. 1999b. Ran and Kim. 2000). Ran. They have more advantages in that they have been designed under the concept of optimization and minimization against the responses to environmental conditions. The surfaceproduction trees and rigid risers are allowed due to the abovementioned aspect of design. many researchers have proved that coupled dynamic analyses are indispensable to get more convincing results from the platform responses and the line tensions than those of conventional uncoupled analysis methods (Pauling and Webster. 1986. It results in avoiding the resonance between the motion of TLPs and the wave excitations so that it is able to stay more stable while operating during oil or gas extraction. the Mineral Management System (MMS) has approved the installation of an FPSO under the condition that the vessel has the construction of a double hull tanker in .2 because they can allow for environmental conditions more flexibly than the fixed structures. Ma et al. they have small water plane areas compared with other floating structures. 1999a. For the floating structures in deep water. Kim and Zheng. Kim et al. Ran. the highstrength vertical tethers are normally used. Kim and Zheng. For the TLPs. instead of the subsea trees and flexible risers that are more expensive. 1994. For the spars. costsaving problems and less risk of oil spills. Since the shipshaped floating structures called FPSOs have more advantages as the solutions to comparably large deck space. 1997. and it allows using the surfaceproduction trees.
In addition. The large storage capacity is the biggest advantage because no pipeline has to be laid out from the sea floor to the land. 1. He derived the equation of the motions of a singlepointmoored FPSO exposed to current. . so that it cannot be neglected. wind and longcrested irregular waves. The fully coupled dynamic mathematical model is necessary to estimate realistic motion responses and line tensions. they examined the coupling terms due to the combined modes of motion in still water and in the current. the development of a coupled dynamic analysis code for solving the large yaw motion and the interaction problem of multiplebody system becomes indispensable. and carried out the nonlinear analysis by uncoupled method.2 Literature Review The comprehensive studies about the viscous dampings for dynamic motion analysis of the turretmoored FPSO were performed by Wichers(1988). They proved the viscous part in a normal direction contributes significantly to the hull dynamics.3 the GOM. A kind of LNG carrier or oil shuttle tanker is substituted for the pipelines for the purpose of turning over the oil and gas. the coupling effect of rigid body motion and the motions of the mooring lines and risers was investigated by Wichers and Devlin(2001). which solves the motions of body and mooring lines. separately. The coupling effects of the low frequency component of a viscous reaction force were studied by Wichers and Chunqun Ji (2000). For the installation of FPSO in deep water such as GoM. By conducting a series of experimental studies.
(2001) presented the results of experiments conducted in OTRC(Offshore Technology of Research Center in Texas A&M University) for a turretmoored FPSO in collinear and noncollinear environmental conditions. the study was attempted to compare the effect on drift forces . the wave and current of a 100year return period storm was investigated so that it was verified that the response of a turret FPSO is sensitive to noncollinear environmental conditions. Ward et al. The aspects of the hydrodynamic characteristics of the multiplebody structure combined with a barge and a mini TLP were studied by Teigen (2000). (2002) using a coupled dynamic analysis tool for floating structures. The dynamic motion of FPSO on collinear.4 The extreme response of a turret moored FPSO in GoM was studied by Baar et al. He compared the hydrodynamic coefficients of the multiplebody and the singlebody and also conducted the convergence tests according to the mesh size of the multiple body. They investigated two types of mooring system of the polyester mooring lines and buoy type mooring lines through the time simulation of FPSO 6. developed by him.000 ft under the conditions of 100year hurricane. (2001) solved the drift force for a multiplebody system of the FPSOLNG carrier in parallel arrangement with zero forward speed waves. He emphasized the importance of hydrodynamic interaction for the motion response of two bodies and indicated the fact that neglecting the fluidcoupling effect may result in an erroneous and nonconservative prediction. noncollinear wind. By adding the viscous roll damping to the potential damping. Inoue et al. (2000). The hull/mooring/riser coupled analyses of a tankerbased turretmoored FPSO was carried out by Arcandra et al. Using a threedimensional source technique.
1. Buchner et al. The hydrodynamic interaction of forces and motions of the floating multiplebody was investigated using the WAMIT program (Clauss et al.. For the same model.3 Objective and Scope The main objective of this research is to develop a numerical program to analyze the hydrodynamic interaction responses of multiple bodies. 1999) considering the interaction effects of the multiplebody. The first stage consists of the evaluation and interpretation of the hydrodynamic interaction analysis results with WAMIT and the preparation of the wind and current force data (OCIMF. For a multibody system with a sidebyside mooring of an FPSO and an LNG carrier.5 with experimental results in regular and irregular waves. 2001). mooring lines and risers based on the hull/mooring/riser coupled dynamic program called WINPOSTFPSO(Arcandra. and the composition of the complete matrix of retardation function for the correct prediction of heave and pitch motions. (2001) conducted the numerical simulation for the prediction of hydrodynamic responses of an LNG FPSO with alongside moored an LNG carrier. 1994) for performing the coupled dynamic analysis program newly . 2002). 2002) and the higherorder boundary element method (Choi et al. using the hydrodynamic coefficients calculated by WAMIT (Lee. a linear potential solver was developed by Huijsmans (2001). and the mean and lowfrequency wave drift forces were calculated by using it. They used a free surface lid in this multiplebody diffraction analysis for the calculation of drift forces and a relative viscous damping in a horizontal plane..
). For the wind and current forces. The wave heading angle will be considered separately for each body at every small degree of angle and the relative angles between multiple bodies will be considered at every span in the same manner as for the wave heading angle. the original program (WINPOSTFPSO) will be developed to be able to perform the hull/mooring/riser coupled dynamic analysis for general multiple floating bodies. The former has the characteristics to deal with the close proximity problem of a sidebyside offloading system.6 developed (WINPOSTMULT) for the shipshaped multiplebody system (FPSO. LNG carrier etc. Buchner’s model (2001) and Choi’s model (2002) may be used for a comparative study about the results to be obtained from WINPOSTMULT. it will be considered that the multiple bodies can be laid in any relative position to the open sea. In the second stage of this research. The third stage is to prove the validity of the newly developed program through carrying out the numerical simulation after the proper models are selected. and used the higherorder boundary element method (HOBEM) while the constant panel method(CPM) was used in WAMIT. a modification in some parts of the original program (WINPOSTFPSO) will be needed. The interpretation program (WAMPOSTMULT) of the WAMIT results will be made for the preparing the properly formatted data for WINPOSTMULTI. In the new program. same sized vessels of an FPSO and a shuttle tanker to tackle the problems of both cases of the sidebyside system and the tandem system. The latter took two. The coupled dynamic analysis scheme adopted in the program WINPOST .
7 MULT will be proved as the robust tool for analyzing the interaction problem of the multiplebody floating structure.
1.4 Procedure 1.4.1 Interpretation and Preparation of WAMIT Results and Wind/Current Forces For the calculation of the hydrodynamic coefficients and wave forces, WAMIT (1999) will be used. WAMIT will give the results of 6 × N degree of freedoms (DOFs)
for N bodies in consideration of the N body interaction. WAMIT should be run for each contacting angle between N bodies at every small angle. It will give the hydrodynamic interaction coefficients of added mass and damping and wave forces. The added mass and wave drift damping will be given as a matrix sized by (NFREQ x 6N x 6N), where
NFREQ means the number of frequencies of the wave. The wave forces will be given as
the linear wave force transfer function (LTF), sized by (NFREQ x 6N) and as the sumand differencefrequency components sized by (NFREQ x NFREQ x 6N). WAMIT should be prerun for each contacting angle between Nbodies at every small angle of wave heading and at every small amount angle of contact with each body for the expected positions. These results will be converted as the input data (each input data file will be named as data000.wv) for WINPOSTMULT. For the preparation of the input data, one converting program (WAMPOSTMULT) will be made. The wind and current forces subject to any shipshaped floating structures can be referred to the OCIMF (1994). For the full loading and the ballast condition, wind and
8 current forces and moments can be read from the tables in the booklet published by OCIMF (1994). They also will be prepared prior to running the WINPOSTMULT. In the WINPOSTMULT, the two data files will be read, and the real drafts of the subjected vessels will be recognized as the draft ratio to the full draft. During the running of the program WINPOSTMULT, the angles against wave headings and the relative angles between multiple bodies will be checked at every time step. If the angles exceed the initial angle, the wind/current forces and moments for the updated angle will be read from the files of the hydrodynamic coefficients precalculated for every 5 degree of yaw angle.
1.4.2 Developing the Coupled Dynamic Program
The backborn program, WINPOSTFPSO, is already developed by Arcandra(2001). For the N bodies, the dealing DOF number should be set up as 6N and the related subroutines should be modified. WINPOSTFPSO is a coupled dynamic program that can treat the body and rods(mooring lines and risers). For N bodies, the total equations of motion for the total system will be combined with the mooring line dynamic equations. For a single body system, the final equation of motion with a combination of the coupling terms of a single body and mooring lines/risers is obtained as:
K L K C U L F L C T B B = B (K ) K U F
where, subscripts of r, c and b mean the rod, the coupled term and the body, respectively. If the total number of mooring lines and risers of the system is defined as n L , the
9 matrices in the above equation, where the equations and figures in the parentheses after the matrix name mean the matrix size, are defined as follows:
K L ( (n L × [8 × (n E + 1) − 1]) × (bandwidth ) ) = the stiffness matrix of mooring lines and
risers
K C ( nL × [8 × (nE + 1) − 1] × (6 × N ) ) = the stiffness matrix coupled with the body and
mooring lines/risers
K B ( 6 N × 6 N ) = the motion matrix of the body
U L ( (n L × [8 × (n E + 1) − 1]) ×1 ) = the motion vector of mooring lines and risers U B ( 6 N × 1 ) = the motion vector of the body
F L ( (n L × [8 × (n E + 1) − 1]) ×1 ) = the external force vector subject to mooring lines
and risers
F B ( 6 N × 1 ) = the external force vector subject to the body
where n E is the number of elements per one line, the bandwidth is 15, and N denotes the number of bodies to be considered. For the multiple body system of N bodies, the rigid bodies are lumped at N points with 6 N DOFs, which are connected with springs and dampers to the mooring lines and risers. The number of DOFs of U B will be enlarged to 6 N as much as the number of DOFs for multiple bodies. Furthermore, the part of the program to deal with multiplebody systems needs to be modified for reading
10 the hydrodynamic coefficients and wave forces for the proper contacting angle at every time step, and for evaluating and assigning to the external forces of the wind and current forces for the loading conditions of the subject vessels. At every time step, the program will check the yaw angle for each body, so that if the angle exceeds a certain amount, the proper wave data file will be read and used for next time step. The existing program is implemented to consider the connecting part of the vessel to the mooring lines and risers as stiff linear rotational springs, or dampers only at the position of starting points of mooring lines and risers. On the contrary, the ending points of the mooring lines and risers are to be regarded as jointing to the sea floor with assumed very huge stiffness of the seabed foundation. Some parts of the futuredeveloped program will be modified so that the flexible connections at both ends of the mooring lines and risers are available. The program will use the existing output format of the previous program except extending the columns of output file for 6 N DOF motions.
1.4.3 Comparative Studies
In this stage, the Buchner’s model(2001) and Choi’s model(2002) may be taken for the comparative study about the results to be obtained from WINPOSTMULT. The former is the multiple body system composed of the LNG FPSO tanker and the LNG carrier. The two vessels are located each at very close proximity to the other in the open sea. Buchner et al. (2001) has performed the calculation of hydrodynamic interaction coefficients, wave load coefficients with the linear potential program using a lid
Choi et al. for which two identical SPARs. The analysis results for those models are compared with the simplified springmass models. (2002) used the higherorder boundary element method not CPM(Constant Panel Method) used in WAMIT. The results will be good for comparison with WINPOSTMULT’s. Some examples are taken for verification of the hull/mooring/riser coupled dynamic analyses of twobody system using the WINPOSTMULT program. the 100year storm condition in GOM and the sea condition in West Africa are taken. . The latter used the combining model of the FPSO and shuttle tanker located at close proximity with the sidebyside arrangement and also at a distance with the tandem arrangement.11 technique and the motion analysis of a multiplebody system using the above results as input data. two identical FPSOs and also an FPSO and a shuttle tanker are selected as the test models. For the environmental conditions.
1) (2.12 CHAPTER II DYNAMICS OF THE FLOATING PLATFORM 2. The surface wave theory is derived from the solution of the BVP with the free surface. ∇u = 0 or ∇2Φ = ∂ 2Φ ∂ 2Φ ∂ 2Φ + 2 + 2 =0 ∂x 2 ∂y ∂z (2. Finally.2.2 Formulation of Surface Wave 2. linear and secondorder wave theories are reviewed in the consideration of the free surface boundary value problem. sequentially. and then the boundary element method is discussed as one of the solution schemes for the free surface boundary value problem. the multiplebody interaction of fluid is reviewed. and then the dynamic motions for single body and multiple body systems of the floating structure are described.1 Boundary Value Problem (BVP) of Surface Wave The fluid in the region surrounding the free surface boundary can be expressed as a boundary value problem in the domain. 2.2) . and Morison’s equation and the wave drift damping are considered.1 Introduction In this chapter. the wave loads and dynamic responses of floating structures are discussed. First. The fluid motion can be expressed by the Laplace equation of a velocity potential with the assumption of irrotational motion and an incompressible fluid.
t ) is the displacement on the plane of the free surface to be varied in space and time. the condition can be described as follows: ∂Φ 1 + (∇Φ ⋅ ∇Φ ) + gz = 0 at z = −η ∂t 2 (2.2). The dynamic free surface condition defines that the pressure on the free surface is constant as the equal value to the atmospheric pressure and normally the atmospheric pressure is assumed to be zero.4) where η ( x. ∂z ∂x ∂y φ is the velocity potential.5) where g is the gravitational acceleration. The kinematic condition is to represent that the fluid particle on the free surface at any instance retains at one position of the free surface. The equation of the kinematic free surface condition can be given by: ∂η ∂Φ ∂η ∂η +u +v − = 0 at ∂t ∂x ∂y ∂z z = −η (2. which are the dynamic free surface condition and the kinematic free surface condition. In addition. The bottom boundary condition is given by the condition that the sea bed is impermeable: ∂Φ =0 ∂z at z = −d (2.3) where d is the water depth.13 where u is the velocity in x. The bottom boundary condition is to be considered. there are two free surface conditions. the boundary condition should be considered. y. specifically. In order to solve the equation (2.1) is known as the perturbation method under the assumption that the wave . Thus. y or z direction of fluid. The most popular approach to solve the equation (2. so it becomes ∂Φ ∂Φ ∂Φ i+ j+ k.
2. the wave elevation (wave particle displacement) and the velocity potential are to be taken as the power series forms a very small nondimensional perturbation parameter.7) ∑ε (n) ( n) η The linear wave equations are obtained by solving the perturbation formulation formed with the velocity potential and that with the wave elevation are obtained by: The firstorder potential: . The firstorder solution leads the linear wave theory and the secondorder solution leads the second order wave theory.6) (2. the total velocity potential and the wave elevation are written in the following forms: Φ = ∑ ε (n) Φ (n) η= (2. The linear wave and the second order or higher order wave can be derived from the perturbation formula of the wave equation. to be represented by the wave elevation and the velocity potential in terms of the perturbation parameter.2 Wave Theory The perturbation formulation of the BVP with the first.and secondorder parameters can give the firstorder solution and the secondorder solution.2. Finally. In the method.14 amplitude is very small. The velocity potential is represented by the summation of all perturbation terms and the wave elevation by summation of the perturbative wave elevations. which can give the approximated solution to satisfy partially the free surface boundary conditions.
10) η ( 2) = A 2 k cosh kd (2 + cosh 2kd ) cos(2kx cos θ + 2ky sin θ − 2ωt ) (2.8) η (1) = A cos(kx cos θ + ky sin θ − ωt ) where k is the wave number expressed by (2.15 − igA cosh k ( z + d ) i ( kx cos θ + ky sin θ −ωt ) Φ (1) = Re e cosh kd ω The firstorder wave elevation: (2. The secondorder potential and the secondorder wave elevation are obtained by solving the perturbation formulations formed with the secondorder potential and the secondorder wave elevation are obtained as follows: The secondorder potential: cosh 2k ( z + d ) i ( 2 kx cos θ + 2 ky sin θ −2ωt ) 3 e Φ ( 2 ) = Re ωA 2 sinh 4 kd 8 The secondorder wave elevation: (2. ω is the L wave frequency. are established. A fully developed wave is normally modeled in terms of energy spectra combined with ensembles of wave trains generated by random phases. y. t ) = ∑ Ai cos(k i x cosθ + k i y sin θ − ωi t + ε i ) = Re ∑ Ai e i ( k x cos θ + k y sin θ −ω t +ε ) (2. the wave is irregular and random. A is the wave amplitude.11) sinh 3 kd In the real sea.Moskowitz and the JONSWAP spectra. such as the Pierson.12) i i i i N i =1 N i =1 . and θ is the incident wave angle.9) 2π when L is the wave length. The time series for a given input amplitude spectrum S (ω ) is obtained by combining a reasonably large number N of linear wave components with random phases: η ( x. Wellknown spectra in common usage.
On the other hand.2. To avoid the increase of wave components and to increase the computational efficiency for a long time simulation.3 Diffraction and Radiation Theory The total velocity potential is decomposed into the incident potential Φ I . the total potential can be written by: Φ = ∑ ε ( n ) (Φ (In ) + Φ (Dn ) + Φ (Rn ) ) (2. 2.16 where Ai = 2 S (ω i )∆ω is the wave amplitude of the i th wave. ∆ω is the interval of wave frequency. t ) = Re ∑ Ai e j =1 N i ( k j x cosθ + k j y sin θ −ω ′j t +ε j ) (2. The diffraction wave represents the scattered term from the fixed body due to the presence of the incident wave.14) The diffraction wave force and the radiation wave force have a significant effect on a floating platform in deep water. y. The forces . By applying the perturbation method. the radiation wave means the wave to be propagated by the oscillating body in calm water.13) where ω ′j = ω j + δω j and δω j is a random perturbation number uniformly determined between − ∆ω ∆ω and . the following modified formula is used: η ( x. and the radiation potential Φ R . and ε i is the random phase angle. the diffraction potential Φ D . including the diffraction and the radiation. The total potential and the wave elevation are given by adding 2 2 every solution of each order equation.
2. y.)R = 0 in the fluid ( z < 0 ) on the free surface ( z = 0 ) (2. z ) + φ R1) ( x.20) .19) ( ∂φ D1) ∂φ I(1) =− ∂n ∂n (1) ∂φ R (1) (1) = −iωn ⋅ (ξ + α × r ) ∂n on the body surface (2. z )}⋅ e −iωt ] φ (2. which can be obtained by solving the BVPs of them. the firstorder potential can be written by: 1 Φ (1) = ε (Φ (I1) + Φ (D ) + Φ (R1) ) ( ( = Re[{ I(1) ( x. R = 0 ∂z ( ∂φ D1.8).17 induced by them are evaluated by integration of the pressure around the surface of the floating structure using the diffraction and the radiation potential.17) (2.15) By referring to the equation (2.18) ∂ (1) 2 − ω + φ D . y. y.1 FirstOrder Boundary Value Problem By separation of variable for the firstorder component.3.16) The BVPs for the firstorder potential of diffraction and radiation are defined as the following formula: ( ∇ 2φ D1.)R ∂z =0 on the bottom ( z = −d ) (2. 2. the solution of incident wave velocity potential is inferred as follows: φ I(1) = Re − igA cosh k ( z + d ) cosh kd ω (2. z ) + φ D1) ( x.
ξ1(1) . and A (1) is the firstorder rotational motion of body. respectively.3 for j = 4.24) The radiation potential can be decomposed as follows: ( φ R1) = ∑ ς jφ j(1) j =1 6 (2.α 3(1) ) (2.α 2(1) . ξ3(1) are defined as the amplitude of surge.25) should satisfy the boundary conditions of equation (2.6 (2.5.2.18) to (2.3 means the x .)R = 0 ∂ζ at far field (2. Equation (2. sway and heave motion.22) (2.18 lim r ( ζ →∞ ∂ ( ± ik )φ D1.23) where 1.axis. y . nz ) is the outward unit normal vector on the body surface.25) where φ j(1) represents the velocity potential of rigid body motion with unit amplitude in the j th mode when the incident wave does not exist. α 3(1) are defined as the amplitude of roll.21). Ξ (1) = (ξ1(1) . ξ 2(1) . Ξ(1) is the firstorder translational motion of the body. n y . The Ξ (1) and A (1) can be expressed as follows: Ξ (1) = Re[ξ (1) e − iωt ] . α 2(1) . Thus. while α 1(1) . The six degrees of freedom of the first order motion are rewritten as: ξ j(1) ς j = (1) α j −3 for j = 1.21) where r is the position vector on the body surface. ξ 3(1) ) α (1) = (α 1(1) . pitch and yaw motion.2 . The body boundary condition of φ j(1) is written as: . R is the radial distance from the origin ( r 2 = x 2 + y 2 ). n = (n x . z . ξ 2(1) . A (1) = Re[α (1) e − iωt ].
21)).2) can be solved for each potential component of equation (2.14) and the separation of variable is applied. φ − is the differencefrequency potential. y. The differencepotential and sumfrequency potential can be solved independently. z ) + φ ( x.28) where ω − = ω m − ω n is the differencefrequency. z )}⋅ e φ + I + D + R [ − iω + t ] (2. z ) + φ ( x. 2.6 (2. t ) = ε 2 (Φ (I2 ) + Φ (D2 ) + Φ (R2 ) ) − − = Re { I− ( x. can be solved numerically in consideration of the boundary conditions (equation (2. equation (2.2. The diffraction potential problem. z .5) as follows: .27) These boundary conditions are valid on the body surface.17).5. The Volterra series method will be applied to solve the secondorder BVP.26) ∂φ j(1) ∂n = −iω (r × n ) j −3 for j = 4. z ) + φ R ( x. y. If the secondorder terms are taken from the perturbation formulation (2.3 (2.3) to (2. y. y. y. ω + = ω m + ω n is the sum frequency. z )}⋅ e −iω t φ − + { ( x.3. and φ + is the sumfrequency potential.2.19 ∂φ j(1) ∂n = −iωn j for j = 1.18)(2. y.2 SecondOrder Boundary Value Problem The secondorder boundary value problem is made by considering the interaction of bichromatic incident waves of frequency ω m and ω n with a floating body. y. z ) + φ D ( x. The governing equation (2.28) considering the boundary conditions. equation (2.1) or (2. the secondorder potential is derived by: Φ ( 2 ) ( x.
31) and γ −* mn 2 igAm An* k m (1 − tanh 2 k m d ) − 2k m k n (1 + tanh k m d tanh k n d ) =− 2ω m ν − − k − tanh k − d (2. The secondorder diffraction potential. φ D2.35) ∂ ± ± 2 ± − (ω ) + g ∂z φ D = Q . deal with the second interaction of plane bichromatic incident waves. Thus.32) and the asterisk represents a complex conjugate. and ν ± and k ± are defined respectively by: ν± = (ω ± ) 2 . g k ± = km ± kn (2. the governing equation of the secondorder diffraction potential is defined by: ± ∇ 2φ D = 0 in the quiescent fluid volume ( z < 0 ) on the free surface ( z = 0 ) (2. contains the contributions of the secondorder incident potential and the firstorder potential.29) − x (2.33) ( ) The secondorder diffraction and radiation potential. The governing equation of the secondorder radiation potential is only expressed by the outgoing waves propagated by the secondorder body motion.30) 2 igAm An k m (1 − tanh 2 k m d ) + 2k m k n (1 − tanh k m d tanh k n d ) 2ω m ν + − k + tanh k + d (2.R .34) (2. ( φ D2 ) .20 + 1 + + (γ mn + γ nm ) cosh k ( z+ + d ) e ik cosh k d 2 − 1 − − (γ mn + γ nm* ) cosh k ( z− + d ) e ik 2 cosh k d φ I+ = φ I− = where + γ mn = − + x (2.
+ q mn = − ( iω m (1) ∂φ (1) ∂ 2φ m1) φn − ω 2 m + g g ∂z ∂z 2 ( iω m (1) * ∂ 2φ m1) ∂φ (1) φn − ω 2 m + g g ∂z ∂z 2 1 + + (qmn + qnm ) . 1 + ( bmn = − n ⋅ (ς n(1) ⋅ ∇ )∇φ m1) 2 1 − ( bmn = − n ⋅ (ς n(1) * ⋅ ∇ )∇φ m1) 2 (2. 2 B− = 1 − − (bmn + bnm* ) 2 (2. The Q ± are symmetric and expressed as follows: Q+ = and.37) (2.41) The B ± are also symmetric and expressed as follows: B+ = and.39) ( + + iω n ∇φ m1) ∇φ n(1) − q II ( − + iω n ∇φ m1) ∇φ n(1) * − q II (2.36) on the body surface (2.38) Boundary condition at far field where Q ± are the sum and difference frequency components of the free surface force and B ± are the sum and difference frequency components of the body surface force.43) 1 + + (bmn + bnm ).44) .42) (2.40) − q mn = − (2. 2 Q− = 1 − − (qmn + qnm* ) 2 (2.21 ± ∂φ D =0 ∂z ± ∂φ D ∂φ ± = − I + B± ∂n ∂n on the bottom ( z = − d ) (2.
Therefore.45) (2. the secondorder radiation potential has the same formula as the firstorder radiation potential. since the boundary conditions for the radiation potential do not contain any other potentials: ± ∇ 2φ R = 0 in the fluid ( z < 0 ) on the free surface ( z = 0 ) (2.47) on the body surface (2.49) where ξ ± and α ± are the second order translations and rotational motions of the body at the sum and difference frequencies. The homogeneous term of the secondorder diffraction potential has the farfield propagating behavior.46) ∂ ± 2 − ω + φ R = 0 ∂z ± ∂φ R =0 ∂z ± ∂φ R = −iωn ⋅ (ξ ± + α ± × r ) ∂n on the bottom ( z = − d ) (2.37) for the secondorder diffraction potential needs to be applied to the decomposed diffraction potential into a homogenous term and a particular solution term due to the complication. The governing equation and boundary conditions for the secondorder radiation ± potential φ R are defined as the firstorder radiation BVP.48) lim R ( R →∞ ∂ ± ± ik )φ R = 0 ∂R at far field (2.22 The boundary condition (2. . while the free surface force Q ± are dominant in the particular equation term.
the three terms represent the different contributions to the body forces and moments. (n .. The hydrostatic restoring forces are defined as the multiplication of the restoring stiffness and the motion responses.51) where. n ) n= 1 2 3 (n 4.1 The FirstOrder Hydrodynamic Forces and Moments If all of the potentials are solved. and the last term ( FE1) ) is the exciting forces generated by the incident and the diffraction potentials. The first term ( FS(1) ) is the hydrostatic restoring force.5. the ( ( second term ( FR1) ) is the force term due to the radiation potential. the firstorder force and moment can be obtained from the integration over the whole surface pressure on the body. n6) = r × n for j = 1.52) In the above equation (2. The six components of forces and moments are calculated as follows: F j(1) (t ) = − ρg ∫∫ zn j dS ∂Ω B − ρ Re iως j e −iωt ∫∫ φ j n j dS − ρ Re iωAe −iωt ∫∫ (φ I + φ D )n j dS ∂Ω ∂Ω B B .23 2.6 (2.50) where ρ is the fluid density. n5. n .3 for j = 4. j = 1. 6 (2.3 Hydrodynamic Forces 2.51).. The pressure on the body surface ( ∂Ω B ) is obtained from the potential as follows: ∂Φ (1) P (1) = − ρ + gz ∂t (2. and the components of restoring stiffness are defined as the .2.3.
54) K 45 = − ρg ∫∫ xyn3 dS ∂Ω B K 46 = − ρg∀xb + mgxcg K 55 = ρg ∫∫ x 2 n3 dS + ρg∀zb − mgz cg ∂Ω B K 56 = − ρg∀yb + mgycg where K mn = K nm for all m and n . z b ) is the center of buoyancy of the body.53) K 34 = ρg ∫∫ yn3 dS = ρgAwp y f ∂Ω B K 35 = − ρg ∫∫ xn3 dS = ρgAwp x f ∂Ω B K 44 = ρg ∫∫ y 2 n3 dS + ρg∀zb − mgz cg ∂Ω B (2. ∀ is the buoyancy of the body. and ( xb .51) can be rewritten as the form: . x f and y f are the distances from the center of the water plane area to the center of gravity in xdirection and in ydirection.24 following surfaceintegral form over the wetted body surface at the mean position ( ∂Ω B ): FS(1) = −[K ]{ς ( 1 ) } where K 33 = ρg ∫∫ n3 dS = ρgAwp ∂Ω B (2. respectively. ycg . The hydrostatic restoring stiffness will be used for the motion analysis of the floating body. yb . ( xcg . Awp is the water plane area. The radiation potential forces and moments corresponding to the second term of the equation (2. z cg ) is the center of gravity.
57) They are symmetric and dependent on the frequency of the body motion.60) .mz cg my cg I 11 I 21 I 31 mz cg 0 . The last term of the equation (2.mxcg I 12 I 22 I 32 my cg mxcg 0 I 13 I 23 I 33 (2.mz cg 0 mz cg . They can be represented as follows: ∂φ j M a = ρ Re ∫∫ φ j dS ∂Ω ∂n B (2. and it can be rewritten as the form: ∂φ j ( FE1) = − ρ Re Ae −iωt ∫∫ (φ I + φ D ) dS ∂n ∂Ω B (2. and ς = ςe − iωt are the body motions of six degrees of freedom.(M a&& + Cς ) + FE1) ς ς (2.25 ∂φ j ( φ j dS FR1) = − ρ Re ς j e −iωt ∫∫ ∂Ω ∂n B & = Re(M a&&( 1 ) + Cς ( 1 ) ) = Re[(.my mxcg cg 0 0 m my cg .55) where M a is the added mass coefficients. the equation of motion is formed as: ( ( ( & M&&( 1 ) = FS(1) + FR1) + FE1) = − Kς .56) ∂φ j C = ρ Im ∫∫ φ j dS ∂Ω ∂n B (2.59) where M is the mass matrix of the body.mxcg 0 0 . C is the radiation damping coefficients.58) Therefore. which is described as: 0 m m 0 0 0 M= 0 .51) corresponds to the linear wave exciting force.ω 2 M a − iωC )ς j e −iωt ] ς (2.
FR2 ) is the radiation potential force. I mn = ∫∫∫ ρ B (x ⋅ xδ mn − xm xn )dV is the moment of inertia. ρ B is the density of the body. which denotes the incident and diffraction potential . ∀ and δ mn is the Kronecker delta function. 2.26 where V represents the body volume. m = ∫∫∫ ρ B dV ∀ is the body mass.2 The SecondOrder Hydrodynamic Forces and Moments The secondorder wave forces and moments on the body can be obtained by direct integration of the hydrodynamic pressure over the wetted surface of the body at the instantaneous time step.3. The components of ( ( FE2 ) are defined as Fp( 2 ) = FI( 2 ) + FD2 ) .61) In consideration of the bichromatic wave. and .63) ( where FS( 2 ) represents the secondorder hydrostatic force. The secondorder forces and moments are defined as: ( ( F ( 2 ) = FS( 2 ) + FR2 ) + FE2 ) (2. the secondorder pressure is modified as: + − P ( 2 ) = Re ∑∑ Am An pmn e −iω t + Am An* pmn e −iω t + − 2 2 m =1 n =1 [ ] (2.62) ± where pmn are defined as the sum and difference frequency quadratic transfer functions for the secondorder pressure. The secondorder pressure is defined as: P (2) = − ρ ∂Φ ( 2 ) 1 2 − ρ (∇Φ (1) ) ∂t 2 (2. FE2 ) = Fp( 2 ) + Fq( 2 ) is the second ( order wave exciting force.
70) ± ± g mn = ρiω ± ∫∫ (φ I± + φ D )ndS / ( Am An .67) ± where f mn denote the quadratic transfer function (QTF) of the sum and difference ± ± ± frequency exciting force. QTF is obtained by the addition of hmn and g mn . and Fq( 2 ) denotes the quadratic product of the firstorder forces. D FI(.68) (2.64) (2. The component forces are derived in the integration forms of potentials as follows: ( FS( 2 ) = ρgAwp (ξ z( 2 ) + y f α x( 2 ) − x f α y2 ) )k (2.69) ρ ρω mω n ( (1) (1) + hmn = − ∫∫ (∇φ m1) ⋅ ∇φ n(1) )ndS − ∫ φm φn NdL / Am An 4g L 4 ∂Ω B W ρ ρω mω n ( (1) (1)* * − hmn = − ∫∫ (∇φ m1) ⋅ ∇φ n(1)* )ndS − ∫ φm φn NdL / Am An 4g L 4 ∂Ω B W (2. Each component of the QTF is defined as: ± ± ± f mn = hmn + g mn (2.71) .2D) = ρ ∫∫ ∂Ω B ∂t 2 2 ndS (2.66) ( + − FE 2 ) = Re ∑∑ Am An f mn e −iω t + Am An* f mn e −iω t + − m =1 n =1 [ ] (2.27 forces.65) ( FR2 ) = ρ ∫∫ ∂Ω B ∂Φ (R2 ) ndS ∂t ) ∂Φ (I2. where hmn are ± the contribution of firstorder quadratic transfer function and g mn are the summation of the quadratic transfer function of the sum and difference frequency exciting force due to the incident potential and the diffraction potential. Am An* ) ∂Ω B (2.
2. The radiation potential for the isolated body can be decomposed in the similar manner to the equation (2.73) where S I .25) as follows: I φ R = ∑ ς jφ jI j =1 6 6 (2. the radiation potential and the incident potential are obtained as described in the above sections.4 Multiple Body Interaction of Fluid The boundary value problem of the multiple body interaction of fluid is explained that the effects of the incident potential and the scattered potential on the main body and the adjacent body are investigated.74) II φ R = ∑ ς jφ jII j =1 (2.28 where N = n/ (1 − n z2 ) . φ7II denotes the scattered potential to the isolated body I and II . For the single body system. respectively. S II denotes the wetted surface of the isolated body I and II . The diffraction problem for the isolated body can be defined by the incident potential as follows: ∂φ7I ∂φ =− I ∂n ∂n ∂φ7II ∂φ =− I ∂n ∂n on S I (2. φ7I . respectively.72) on S II (2. and φ I represents the incident wave potential of the isolated body.75) The radiation problem for the isolated body I and II can be given by: . and k is the unit vector in the zdirection.
80) Interaction problem – radiation/scatter from II near I: ˆ ∂φ jII ∂n =− ∂φ jII ∂n on S II ( j = 1.. n Ij .II I. II is a unit normal vector for the six degree of freedom for the isolated body I and II .II = I. and n Ij .2....77) where φ jI . respectively.77)..II (n1 . r The boundaryvalue equation and the boundary condition for each body of the interaction problem is defined in the form of the radiation/scatter potential and the derivative as follows: Interaction problem – radiation/scatter from I near II: ˆ ∂φ jI ∂n ˆ ∂φ jI =− ∂φ jI ∂n on S I ( j = 1.. n2 . II is given by: n I..6) (2.2. φ jII denotes the decomposed radiation potential components for the isolated body I and II ...II ~ (n4..3 for j = 4. n6) = r × n for j = 1.2.79) ∂n =0 on S II ( j = 1.5..7) (2. In equation (2.7) (2.. n3 ) I.76) and (2.6) (2..7) (2. n5...2. respectively.29 ∂φ jI ∂n ∂φ jII ∂n = n Ij on S I ( j = 1.2.76) = n II j on S II ( j = 1.6 (2.78) where ~ denotes the relative distance from the origin to each body center.2.81) .
The potential when j = 7 means the scatter term. 2. since the solution to satisfy all of the boundary conditions. φ is the exact solution of potential and G satisfies the following equation: . and ∂Ω denotes the boundary of the domain. except the body boundary condition for the firstorder potential and the body boundary condition and the free surface condition for the secondorder potential. If the firstorder radiation/scatter potential is used when the above BVP is solved.30 ˆ ∂φ jII ∂n =0 on S I ( j = 1. BEM is generally called the inverse formulation. Ω denotes the fluid domain. the resultant potential would be the firstorder interaction potential.2..5 Boundary Element Method The boundary element method is proper for solving the boundary value problem of the fluid potential around the floating body since there is no analytic solution except for some special geometric bodies.83) where G is the Green function to satisfy all of the boundary conditions. respectively...82) ˆ where φ jI . is used as a weighting function. while the secondorder radiation/scatter potential leads the secondorder interaction potential. It is also based on GreenLagrange’s Identity given by: ∫∫∫ (G∇ φ − φ∇ G )dΩ = ∫∫ G ∂n − φ ∂n dS 2 2 Ω ∂Ω ∂φ ∂G (2. II denotes the interaction potential affected by radiation/scatter potential from the body I to the body II .7) (2. and vice versa.
it becomes as: G (ξ. ∂Ω B represents the body boundary. x ) − I ∂n ∂n ∂Ω B (2. If the body geometry has a smooth surface.85) is a fundamental equation called the Inverse Formulation.6 (2.31 ∇ 2 G = δ (x) (2. x )(r × n )k −3 dS (ξ ) ∂Ω B B B for k = 1. it becomes as: .83) becomes: c( x)φ ( x ) = ∫∫ G ∂n − φ ∂n dS + ∫∫ G ∂n − φ ∂n dS ∂Ω F ∂φ ∂G ∂φ ∂G (2. and x means the position coordinates. Since φ and G satisfy all of the boundary conditions except the body or the free surface.2.85) ∂Ω B where c( x ) means a shape factor depending on the body geometry.86) where ξ denotes the source point coordinates. and ∂Ω F is the free surface boundary.5.87) If the formulation is applied to the secondorder diffraction potential problem for the flat surface of body. the right hand side of the equation (2. x )nk dS (ξ ) ∂∫∫ ∂G (ξ.84) where δ is Dirac delta function. The equation (2.85) becomes a second kind of Fredholm integral equation such as: ( 2πφ D1) ( x ) + (1) ∫∫φ D (ξ) ∂Ω B ∂φ (1) (ξ ) ∂G (ξ. x ) dS (ξ ) dS (ξ ) = ∫∫ G (ξ.3 for k = 4. the equation (2. x ) Ω (1) (1) dS (ξ ) = 2πφR ( x ) + ∫∫φ R (ξ ) ∂n ∂Ω ∫∫ G (ξ. If it is applied to the firstorder radiation potential problem. c( x ) becomes 2π . If the formulation is applied to the firstorder diffraction potential problem for the smooth surface of body.
92) where N j is the shape function..88) If it is applied to the secondorder radiation potential problem for a far field.90) If the equation is applied for the discretized model. x2 ) is the local coordinate. of Interpolation points) j =1 M ∂φ H ijφ j = ∑ Gij .3 ∫∫ R →∞ ∂∫∫ ∂Ω Ω ± ± ∂G 2πφR + ∫∫φ R dS = ± ± ± 2 ± ∂n ∂Ω ∫∫ G (r × n )k −3 dS + ∫∫ ω G φ R m lim ik Rφ R dS for k = 4. j ≠ i ∫∫ G(ξ.. j ≠i (2.2.93) ∂Ω B .. it becomes as: ± ± G ± nk dS + ω 2 G ±φ R m lim ik Rφ R dS for k = 1.2. the simplest form is shown as: 2πφ ( x ) + ∂Ω B ∫∫φ (ξ) ∂G (ξ.6 R →∞ ∂Ω ∂Ω ± B F B ( ) ( ) B F (2. x)dS (ξ) (2. x2 )φ j .2.91) (2. x ) ∂φ (ξ ) dS (ξ ) = ∫∫ G (ξ. x ) dS (ξ ) ∂n(ξ ) ∂n(ξ ) ∂Ω B (2. L = 1. ( No of pannels) ∑ j =1 j =1 ∂n j M L (2..32 ∂φ ± ∂G ± dS = ∫∫ G ± B ± − I ∂n ∂n ∂Ω B ± ± 2πφ D + ∫∫φ D ∂Ω B 1 dS + ∫∫ Q ± G ± dS g ∂Ω F (2. M = 1. x ) dS (ξ ) ∂n(ξ ) .. If the Constant Panel Method (CPM) of BEM is taken. and H ij and Gij are as follows: 1 1 H ij = δ ij + 2 4π Gij = 1 4π ∂Ω B ∫∫ ∂G (ξ. ( No. it is modified as: φ (ξ ) = ∑ N j ( x1 ... ( x1 . it is noted that the integration term for the free surface remains.89) In this formulation..5.94) .
In this study.95) where h1 (τ ) is the linear impulse response function. so that it is essential to make the random waves for applying the external wave loads to the floating body.33 ∂φ ∂n In the equations of (2. Thus. the secondorder exciting force at time t for the two .20) and G (ξ.6.τ 2 )η (t − τ 1 )η (t − τ 2 )dτ 1 dτ 2 (2. ∂G (ξ. and the secondorder sum and difference frequency wave loads are computed by considering the bichromatic wave interactions. and h2 (τ 1 .92) and (2.e. τ 2 ) is the quadratic impulse response function. the WAMIT will be taken for solving the fluid interaction problem of the multiplebody system.92) can be solved ∂n(ξ ) for the whole panels. 2.94).6 Motions of the Floating Platform 2. i. For the BEM program. the equation (2. The linear and the secondorder hydrodynamic forces can be rewritten as the form of a twoterm Volterra series in time domain: F (1) (t ) + F ( 2 ) (t ) = ∫ h1 (τ )η (t − τ )dτ + ∫ −∞ ∞ ∞ −∞ −∞ ∫ ∞ h2 (τ 1 .1 Wave Loads The linear wave forces are calculated in the frequency domain. x ) are known as the exact forms. the WAMIT can be applied to the firstorder and secondorder diffraction/radiation potential problem. x ).. The real sea is made of random waves. the WAMIT (Lee et al. is given by the equation (2. 1991) of CPM is well known in this field.
97) can be easily changed into the energy spectra given by: ( S F1) (ω ) = Sη (ω ) q L (ω ) 2 (2. ω 2 2 − µ ) Sη ( ω 2 + µ ) Sη ( ω 2 − µ )dSη ( µ ) (2.95) can be rewritten in the form of the summation of the frequency components as follows: N FI(1) (t ) = Re ∑ A j q L (ω j )e iωt j =1 N N N N FI( 2 ) (t ) = Re ∑∑ A j Ak* q D (ω j . η (t ) is the ambient wave free surface elevation at a reference position. h1 (τ ) and h2 (τ 1 . and q D (ω j .98) 2 − S F (ω ) = 8∫ q D ( µ . the unidirectional wave exciting forces induced by the incident potential and the diffraction potential to have the similar form of the equation (2.99) S (ω ) = 8∫ + F ω/2 0 qS ( ω 2 + µ. Using the Fourier transform. ω k )e iω t j =1 k =1 j =1 k =1 − + (2. respectively.96) (2. ω − µ ) Sη ( µ ) Sη (ω − µ )dSη ( µ ) 0 ∞ (2.97) where q L (ω j ) represents the linear force transfer function (LTF).−ω k ) and q S (ω j .100) . ω k ) are the difference and the sum frequency quadratic transfer functions (QTF).−ω k )e iω t + ∑∑ A j Ak q S (ω j . τ 2 ) can be expressed in the functions of frequency. Since η (t ) .96) and (2. the equation (2.34 different unit amplitude inputs at time τ 1 and τ 2 .
55) at frequency ω . . FI = FI(1) + FI( 2 ) is the sum of the equation (2.101). Fc is the last term of the right hand side of the equation ~ (2.101) as the same form as the summation of the equation (2.and differencefrequency wave force spectrum.63) as follows: ~ FT = FI + Fc + FR (2. respectively.101) where M a (ω ) is the added mass coefficient as defined in the equation (2.56) at frequency ω .96) and (2.59) and (2. The total wave forces and moments can be obtained by summation of the equation (2. The first.96).101).and secondorder radiation potential forces are calculated by the following formula: t ∞ a M (ω ) − ∫ R (t ) cos ωtdt ς&(t) − ∫ R(t − τ )ς(τ )dτ & & FR (t ) = 0 −∞ (2.103) where FT = F (1) + F ( 2 ) is the total wave exciting force. and R (t ) is called a retardation function as defined below: R (t ) = C (ω ) π∫ 0 2 ∞ sin ωt ω dω (2.35 ( where Sη (ω ) is the wave spectrum. (2.102) where C (ω ) is the radiation damping coefficient in the equation (2. and FR is the first term of the equation (2.97) and (2.97). S F1) (ω ) is the linear wave force spectrum. and − + S F (ω ) and S F (ω ) are the secondorder sum.
the first term is called FroudeKrylov force. DS is the breadth or & diameter of the structure. D is the diameter of the slender body. Ca is the added mass coefficient. V = πD 2 4 is the volume per unit length of the structure. The drag force on the floating structure cannot be neglected.104) is added to the wave forces of the equation (2.2 Morison’s Equation For the slender cylindrical floating structure.103) to get the total force.104) where Fm denotes Morison’s force. and the last term the drag force. respectively.6. . and ς&n and ς n are the acceleration and the velocity of the body. (1950) proposed that the total force is the sum of drag force and inertia force as follows: & & Fm = C m ρVu n − C a ρVς&n + 1 & & ρC D DS (u n − ς n ) u n − ς n 2 (2. C m = 1 + C a is the inertia coefficient. In the above equation. C D is the drag coefficient. The derived force by the equation (2. the inertia and added mass effect and the damping effect of the drag force on the slow drift motion can be evaluated by using Morison’s equation. the second term the added mass effect. u n and u n are the acceleration and the velocity of the fluid & & normal to the body. because the slenderness ratio of the structure (the ratio of breadth or diameter to the length of the structure) is small compared to the wavelength so that the viscous effect cannot be negligible. Morison et al. respectively.36 2.
109) 0 .108) where M a (∞) is a constant.3 Single Body Motion The equilibrium equation using Newton’s second law called the momentum equation for the floating structure can be given as: M d 2 x cg dt 2 =f (2. In the time domain. the above equation is expanded as: [M + M and can be expressed by : a & & (∞) && + Kς = FI (t ) + Fc (ς .105) I dϕ + ϕ × ( Iϕ ) = m dt (2. t ) + Fm (ς . the equation (2. The second term of the lefthand side of the equation (2.106) becomes a linear equation as follows: M&& = F (t ) ς (2. equivalent added mass of the body at the infinite frequency M (∞) = M (ω ) − ∫ R(t ) cos ωtdt a a ∞ (2. and ϕ is the angular velocity.37 2.106) where M is the mass of the floating structure.107) ς where && is the normal acceleration of body motion. f and m are the external force and moment. I is the moment of inertia.104) and the relative angular motion of the body to the wave motion are nonlinear. x cg is the coordinates of the center of gravity of the floating body. If the rotation is assumed to be small. t ) ς ] (2.6. M is the 6 × 6 body mass matrix to be the same as equation (2.59) and F(t) is the external force vector.
96) and (2. the matrix sizes are extended accordingly. the body motion vector and the force vector in the equation (2. And also in the total system equation (2. N of which is the number of bodies. the total equation of motion in the global coordinate system is assembled for the combined system. and radiation/scatter from body II in the presence of body I.56). Finally. Fc is the same as the second term of the equation (2. the local coordinate system is used for each body. For the formulation of motion. t ) = − ∫ R (t − τ )ςdτ −∞ t (2. . 2.106).6.4 Multiple Body Motion For the multiple body system.103) and defined as: & & Fc (ς .97). ς is the normal velocity of the body.104). The hydrodynamic coefficients are premade in consideration of the fluidinteraction terms influenced on each body by using WAMIT.106) are changed to 6 N × 6 N .110) FI is the same as the equation (2. 6 N and 6 N . and Fm is the force by Morison’s & equation such as the equation (2. the coordinate transformation is needed. After forming the equation of motion for each body.38 where M a (ω ) is the same as defined in the equation (2. The hydrodynamic coefficients are solved in the sequence as follows: 1) The radiation/diffraction problem for each body in isolation 2) The interaction problem resulting from radiation/scatter from body I in the presence of body II. the number of the degrees of freedom of the mass matrix.
L. j = 1.109). II = − ∫ aφ7II n j dS . S II .6 ) ( j = 1. S II I . The boundary–value problem of fluid interaction is solved using the equation (2. SI ( j = 1.6 ) ( j = 1. II = − ∫ a (φ7II + φ7II )n j dS . but other boundary conditions for the bodies are the same as those in the isolated body. L . L . SI ˆ C jII . SI II . L .2. S II where the superscript I and II represent the body I and II.2. II ˆ (∞) = − ∫ φ jII ni dS .2. the twobody problem should be addressed for each unique pair of bodies.6 i. If there are several bodies. SI i.39 Where body I and II represent one pair of bodies which interact hydrodynamically. I ˆ (∞) = − ∫ (φ jI + φ jI )ni dS . the hydrodynamic coefficients are obtained by: I . j = 1. I = − ∫ a (φ7I + φ7I )n j dS . j = 1.112) (2.6 ) ( j = 1.L.116) (2.6 i.I ˆ M a (∞) = − ∫ φ jI ni dS .81) and (2. L .L. SI ˆ C jI . II ˆ (∞) = − ∫ (φ jII + φ jII )ni dS . If the coefficients are written in the form of equation (2.6 i. The boundaryvalue problem is formed differently due to the different kinematic boundary condition on the immersed surface of bodies.2.4 in the form of an excitation force coefficient as follows: ˆ C jI . I = − ∫ aφ7I n j dS . j = 1.114) ˆ C jII .2.2.113) (2.111) (2.L.82) in the section 2.2.2.118) Ma Ma Ma II .115) (2.6 ) (2.6 (2.117) (2.
40 Then.II . Thus. for the twobody problem. Fm (2.121) (2. II (2. K II . 2.6.120) (2.223) where the superscript I and II represent the body I and II. The equation of motion in time domain for a singlebody system and/or a twobody system is expressed as the . the other matrices contain the terms for two bodies.106). the equation (2. M I 0 .107). FI FC I FC = II .I K FI I FI = II . the matrices are of the size of 6 N × 6 N . M= II 0 M K I . FC Fm I Fm = II .116) are replaced for the equation (2.106). the numerical scheme of the iterative procedure in the time domain is commonly used.5 Time Domain Solution of the Platform Motions Since the system contains the nonlinear effect. I K = II.113) to equation (2. In the equation (2.119) K I.106). and the replaced equations mean the matrix M a (∞) in the equation (2. The total equation of motion of the system has the same form of equation (2.122) (2. but for the Nbody with 6 DOF for each body.
the following equation is obtained: t ( n +1) t ( n +1) ~ ~ Mη ( n+1) = Mη ( n ) + ∫ ( n ) (FI + Fc + Fm )dt − ∫ ( n ) Kςdt t t t ( n +1) (2. there are several kinds of implicit methods developed.125) & η =ς ~ where M = M + M a (∞) denotes the virtual mass matrix.127). the following equation is obtained after the resultant equation rearranged: ∆t ( n +1) ∆t ~ ~ ( n +1) ( n +1) (n) (n) (n) Mη ( n+1) = Mη ( n ) + (FI + Fc + Fm + FI + Fc + Fm ) − K (ς ( n+1) + ς ( n ) ) 2 2 (2. RungeKuta method and the AdamsMoulton method (or midpoint method). If the integration from time step t (n ) to t ( n+1) is performed. such as the NewmarkBeta method.124) (2. ς ) + Fm (t . In the first step.41 equation (2.108) is derated to the first order differential equation: ~ & Mη = FI (t ) + Fc (t . For the numerical integration in the time domain.126) (2. The last is used for the purpose of the guarantee of the secondorder accuracy. the equation (2. the AdamsBashforth method is also used for the time integration of the nonlinear force.110). Furthermore.127) ς ( n+1) = ς ( n ) + ∫ t(n) ηdt If the AdamMoulton method is applied to the equation (2.109) and (2.108) with the equation (2. Another reason to use it is that the method has the merit to solve together the coupled equations of the platform motion and mooring line motions at each time step.228) . ς ) − Kς (2.126) and (2.
To solve the above equations.229) The equations (2. Thus.124) and (2.233) Eventually. the time integration of the nonlinear term of drag force in Morison’s formulation is as follows: ∫ ∫ t ( n +1) (n) t Fm dt = ∆t (n) ( n −1) (3Fm − Fm ) 2 for n = 0 (2. the equation (2.231) In the same sense.229) are the combination of two linear algebraic equations with the unknowns of η ( n+1) and ς ( n+1) .234) − Fm ( n −1) ) − 2Kς (n) + 2F0 (2. the assumption of the first terms is needed.125) are derived as follows: 4 ~ (n) 4 ~ ( n +1) (n) (n) ( n −1) + FI ) + (3Fc − Fc ) ∆t 2 M + K ∆ς = ∆t Mη + (FI + (3Fm ∆ς = ς ( n+1) − ς ( n ) (n) (2.228) and (2.235) .230) t ( n +1) t(n) Fc dt = ∆tFc( 0 ) (2.232) t ( n +1) t(n) ( Fm dt = ∆tFm0) (2. the AdamsBashforth scheme is used. the time integration of the nonlinear term of radiation damping force is as follows: ∫ ∫ t ( n +1) t(n) Fc dt = ∆t (n) ( n −1) (3Fc − Fc ) 2 for n = 0 (2.42 2 ( n+1) (ς − ς (n) ) −η (n) ∆t η ( n +1) = (2. It means that the time integration may have an error term due to the arbitrary adoption of the first term. For the evaluation of the first terms of time varying unknowns to avoid the abovementioned problem.
the equation (2.43 where F0 represents the net buoyancy force for balancing the system. .234) is solved for the unknown of ∆ς . Then.229) and (2. the time interval of ∆t may be small enough to solve the mooring line dynamics. η ( n+1) and ς ( n+1) can be obtained from the equation (2. To obtain the stability and the accuracy of the solution.235). Firstly. since the mooring line shows a stronger nonlinear behavior than the platform movement.
the theory and the numerical method for the dynamic analysis of the mooring lines and risers are explained.44 CHAPTER III DYNAMICS OF MOORING LINES AND RISERS 3. The main purpose of risers is not to fix the floating structure in . and to be connected to the TLP. Synthetic mooring lines made of polyester are now considered as a more efficient solution. For exporting and importing gas and water. the Single Point Mooring System (SPM) and the shuttle tankers with hawsers or fluid transfer lines(FTLs). The multiple body interaction problems are caused by those kinds of system arrangements.1 Introduction In this chapter. risers are taken into account. and for the production of gas. several types and different materials of mooring lines have been installed. To maintain the sea keeping. Sometimes FPSOs are needed to construct the mooring lines and risers. usually intended to be installed in the TLP. A steel wire rope with chains at both ends has been used for SPAR platform in deep water. For the sea keeping for FPSOs. There are also taut vertical mooring lines and tethers made of several vertical steel pipes. The platform is considered as a singlepoint floating system when the behavior of the mooring line is taken into account. the attempt is to use synthetic mooring lines for fixing those structures in very deep water(over 6.000 ft). The interaction problem between the floating platforms is the matter to be presolved before planning the deep water installation of FPSOs.
But. the bending stiffness of the tendon and the riser in a TLP has a restoring effect. For the consideration of the hydrodynamic force on the line. The advantage of the elastic rod theory is that the governing equation. The restoring forces of both lines result from gravitational forces. In the mooring lines and risers. The strain and the stress of a structure with geometric nonlinearity can be solved with the beam theory using the updated Lagrangian approach. the tensioned string theory is used. Here the FEM technique suggested by the latter is taken. geometries and line tensions. The analysis of line dynamics is developed on the basis of the theories of behaviors of slender structures. Garret proved line dynamics could be solved more accurately by the FEM. in the program. It is called the elastic rod theory. The finite difference method was applied to this problem by the former. including the geometric nonlinearity. can be treated in the global . The static position and the line tension are obtained by using the catenary equation. In this study. in that they don’t have bending stiffness and are the slender members. but in the theory the structural strain and stress contribution are missing. It tends that the steel catenary risers are used more and more because they are inexpensive. but to act the roles. Both mooring lines and risers are the same from the viewpoint of the installation.45 the station keeping position. the geometric nonlinearity is dominant on the line behavior. the tensioned string theory using the string modeled as the beam elements is adopted for its rigorous analysis. In the catenary equation. and the formula was derived by Nordgen(1974) and Garret(1982). no hydrodynamic force on the line is considered. Therefore. a threedimensional elastic rod theory containing line stretching and bending behavior is adopted.
where the prime means the derivative with respect to the arclength s . t) M Y X Figure 3. Z s F r (s.1 Coordinate system of the rod . the principal normal vector as r ′′ .2 Theory of the Rod The behavior of a slender rod can be expressed in terms of the variation of the position of the rod centerline. A position vector r ( s. the governing equation of the static equilibrium and the dynamic problem of the body and lines is constructed in a form of weak formulation based on the Galerkin method to apply the Finite Element Method.46 coordinate system without transforming the coordinate system. t ) is the function of the arc length s of the rod and time t . The space curve can be defined by the position vector r . In this chapter. 3. and the binormal as r ′ × r ′′ .1 shows the coordinate system of the rod. The unit tangential vector of the space curve is expressed as r ′ . Figure 3.
and H is the torque.5) where m ⋅ r ′ is the distributed torsional moment.2) q = applied force per unit length ρ = mass per unit length of the rod m = applied moment per unit length The dot denotes the time derivative.2) and (3. the bending moment and the curvature has the relationship as: M = r ′ × EI r ′′ + H r ′ (3. .3) can be combined as follows: ′ r ′ × EI r ′′ + F + H ′r ′ + H r ′′ + m = 0 ( ) (3.3) where EI is the bending stiffness.4) yields H ′ + m ⋅ r′ = 0 (3. the torque in the line is usually small enough to be negligible. For the moment equilibrium. Furthermore.4) The scalar product with r ′ for the equation (3. Equation (3. This means that the torque is independent on the arclength s.1) (3. Since there is no distributed torsional moment. m ⋅ r ′ = 0 and H ′ = 0 .47 F ′ + q = ρ && r M ′ + r′× F + m = 0 where F = resultant force acting along the centerline M = resultant moment acting along the centerline (3.
7) using the relation of (3. the equation (3.7) where λ is the Lagrangian multiplier.9) λ = T − EIκ 2 (3. then the following formula is obtained: ′ F = − EI r ′′ + λ r ′ ( ) (3.8) λ = F ⋅ r ′ + EI r ′′ ⋅ r ′ or ( )′ (3.11) If the stretch of rod is assumed to be linear and small.12) .6) and the product with r ′ is taken. which is also called Lagrangian multiplier.10) If the equation (3.8) can be approximated as: λ T 1 (r ′ ⋅ r ′ − 1) = ≈ AE AE 2 (3. (3. Thus.6) If a scalar function.1). the inextensibility condition (3. the following equation of motion is obtained: ″ ′ − EI r ′′ + λ r ′ + q = ρ && r ( ) ( ) (3.4) can be rewritten in the reduced form: ′ r ′ × EI r ′′ + F = 0 ( ) (3. t ) .48 Here the torque H and the applied moment m on the line are assumed to be zero.8). λ ( s. is introduced to the equation (3.7) is substituted into (3. r ′ should satisfy the inextensibility condition: r′ ⋅ r′ = 1 Applying dot product with r ′ to (3.
49 In the floating platforms. and &&n is the component of the rod acceleration normal to the rod centerline. Thus. V n is the normal acceleration to the rod centerline.15) where C A is the added mass coefficient (added mass per unit length ). The velocity and acceleration of the . r n is r the component of the rod velocity normal to the rod centerline. C M is the inertia coefficient (inertia force per unit length per unit normal acceleration of rod). the applied force on the rod comes from hydrostatic and hydrodynamic forces caused by the environmental excitation by the surrounding fluid. F s is the hydrostatic force on the rod per unit length. C D is the drag coefficient (drag force per unit length per unit normal velocity). and the gravitational force on the rod. and F d is the hydrodynamic force on the rod per unit length.14) where B is the buoyancy force on the rod per unit length. The hydrostatic force can be defined by: F s = B − Pr ′ ( )′ (3. V n is the normal & & velocity to the rod centerline.13) where w is the weight of the rod per unit length. q may be rewritten by: q = w+ Fs + Fd (3. and P is the hydrostatic pressure at the point r on the rod. The hydrodynamic force on the rod can be derived from the Morison formula as: & & & & & F d = −C A &&n + C M V n + C D V n − r n V n − r n r = −C A && + F r n d (3.
20) λ = T + P − EIκ 2 = T − EIκ 2 ~ w = w+ B ~ ~ (3.21) (3. Nordgren (1974) applied the finite difference method .22) (3. and their derivatives. (3.18) (3. The equation (3.50 rod can be derived from the fluid velocity vector.11) can be rewritten as: ~ ρ && + C a ρ w &&n + ( EI r ′′)′′ − (λ r ′)′ = w + Fd r r where ~ ~ (3.20) is nonlinear. and w is the effective weight or the wet weight of the rod. 3.14) and (3.20) with the equation (3.19) ( ) When the above equation (3.17) (3.23) ~ T =T +P ~ ~ T is the effective tension in the rod. then the equation (3. the line tangential vector.15) are used.16) (3.12) is the fundamental equation of motion for the elastic rod to be applied to the FEM formulation.13).3 Finite Element Modeling The governing equation (3. & & V n = V − r − V − r ⋅ r′ r′ & & V n = V − V ⋅ r′ r′ & & & r n = r − (r ⋅ r ′) r ′ &&n = && − (&& ⋅ r ′) r ′ r r r ( ) [( ) ] (3. and can be solved except for special cases with particular conditions.
27) λ ( s.51 to the governing equation and the inextensibility condition. His analysis results showed satisfactorily the dynamic behavior of the pipe on the sea floor. t ) = Al ( s )U il (t ) (3. and U il .24) λ 1 (rr′rr′ − 1) − =0 2 AE Here the unknown variable r . the FEM technique is taken due to its various merits. Al .26) (3. λ m are the unknown coefficients. The application of the FEM starts from describing the equation in the form of tensor such as: ~ ~ ~ − ρ&& − C A &&n − ( EIri′′)′′ + (λ ri′)′ + wi + Fi d = 0 ri ri and (3. 0 ≤ s ≤ L .29) 0 The following cubic shape functions for Al and quadratic shape functions for Pm are used on the basis of the relation of δri = Al δU il (t ) and δλ = Pm λδ m such as equation (3. Pm are the interpolation(shape) functions. t ) = Pm ( s )λ m (t ) where.27): . λ can be approximated as: ri ( s.28) 0 ∫ δλ 2 (r ′r ′ − 1) − AE ds = 0 r r L 1 λ (3.25) (3. By introducing shape functions for the solution.26) and (3. In this study. the weak formulations for applying the FEM technique are written by multiplying the weighting function of δri as follows: r ∫ δr [− ρ&& − C L i i A i ~ ~ ~ &&n − ( EIri′′)′′ + (λ ri′)′ + wi + Fi d ds = 0 r ] (3.
t ).30) P1 = 1 − 3ξ + 2ξ 2 P2 = 4ξ (1 − ξ ) P3 = ξ (2ξ − 1) where ξ = s .31) λ1 = λ (0. t ). t ) (3.33) Thus. L U i1 = ri (0. t ) (3. the equation (3.32) (3. λ 2 = λ ( .35) L 0 = ( EIri′′)′ Al L ds − EIri′′Al′ 0 + EIAl′′ri′′ δU il 0 ∫ L ∫ 0 L δri (λ ri′)′ ds = ~ ∫ ~ (λ ri′)′ Al δU il ds (3. L 2 λ 3 = λ ( L. t ). U i 2 = ri′(0. t ).52 A1 = 1 − 3ξ 2 + 2ξ 3 A2 = L ⋅ (ξ − 2ξ 2 + ξ 3 ) A3 = 3ξ 2 − 2ξ A4 = L ⋅ (−ξ 2 + ξ 3 ) (3. U i 3 = ri ( L.31) can be extended in term by term as follows: ∫ ∫ 0 L L 0 δri ( ρ&& + C A &&n )ds = ri ri ∫ L 0 ( ρ&& + C A &&n ) Al δU il ds ri ri (3.34) δri ( EIri′′)′′ ds = ( EIri′′)′′ Al δU il ds 0 ∫ L (3. U i 4 = ri′( L. t ).30) and (3.36) L ~ = (λ ri′Al ) − λ ri′Al′ ds δU il 0 0 ∫ L ~ .
41) ∫ ( L 0 L 0 L Al C A &&n ds = C A Al Ak δ ij ds − ri 0 ∫ ∫ ( A A A′ A′ )U U 0 l k s t it L js δ ij ds U jk && (3.37) are assembled and the term of δU il is canceled out in both sides of the above equations.38) If the equation (3.38).39). and δλ m is removed from both sides of the equation (3.34) to (3.38).43) .42) ∫ EIA′′r ′′ds = ∫ EIA′′A′′δ l i l k ij dsU jk (3. the following equation is obtained: ~ ~ ~ ∫ {A (ρ&r& + C &r& )+ EIA′′r ′′ + λ A′r ′ − A (w + F )}ds l i A i n l l l i i d 0 L ~ L = − EIri′′Al′ 0 + λ ri′ + (EIri′′)′ Al [ ] (3. then the following equations are obtained: ∫ L 0 && ρAl && ds = ρAl Ak δ ij dsU jk ri 0 ∫ ) L (3.40). and the boundary conditions satisfy the equation (3.53 L ∫ [ L ~ ~ ~ + F d ds = ( w + F d ) A ds δU ~ δri wi i i i l il 0 0 ] ∫ (3.40) If the partial integrations are applied twice term by term for the equation (3. the equation (3.39) L 0 If the same operation is done for the equation (3.39) and (3.37) ∫ 0 L δλ 1 (rr′ rr′ − 1) − λ ds = AE 2 ∫ L 0 λ 1 Pm (rr′ rr′ − 1) − dsδλ m AE 2 (3.38) becomes as: ∫P 0 L m λ 1 ′ ′ (rr rr − 1) − ds = 0 AE 2 (3.
48) Gm = AmilU klU ki − Bm − Cmn λn = 0 where.47) (3.51) 2 K nijlk = ∫ P A′ A′ δ n l 0 L k ij ds (3.46) Using the equation (3.44) ∫ ∫ 0 L 1 1 ′ Pm rr′ rr′ ds = Pm Al′ Ak dsU jl U jk 2 2 ∫ 0 (3. M ijlk = ∫ ρA A δ l k ij ds (3.45) Pm λ AE ds = 1 L λ n Pm Pn ds AE 0 ∫ (3.54 ∫ L 0 L ~ 0 ′ λ Al′r ′ds = λ n Pn Al′ Ak δ ij ds 0 L ∫ L (3.41) to (3.40) can be rewritten in a matrix form as follows: 1 2 a && ( M ijlk + M ijlk )U jk + ( K ijlk + λ n K nijlk )U jk − Fil = 0 (3. the equation (3.53) .50) = ∫ EIA′′A′′δ l 0 L k ij ds (3.49) L L a ′ M ijlk = C A ∫ Al Ak δ ij ds − ∫ Al Ak As At′ds U itU jsδ ij 0 0 1 K ijlk (3.52) ~ ~ Fil = ∫ ( wi + Fi d ) Al ds 0 L (3.39) and (3.46).
55 and 1 = 2 Amil ∫ P A′ A′ds m i l 0 L (3.54) 1 Bm = 2 ∫ P ds m 0 L (3. Fil .55) C mn 1 = AE ∫ P P ds m n 0 L (3. the total force and the stretching force are described as Ril and Gm as: . some numerical approaches for solving the nonlinear timedomain problem in time domain are needed.56) and δ ij is the Kronecker Delta function. So. contains nonlinear terms. the total equations are nonlinear. 3. Thus.48) are used for solving the rod dynamics. in addition to the above manipulation. The program is implemented for calculating the equation (3. If the residuals are taken from the system energy equation and the inextensibility equation. they should be zero. Since the force vector.49) to (3. using the system parameters and the integration of the shape functions. these schemes are introduced and explained.56).47) and (3.47) and (3. In the following sections.48) can be called the equilibrium equation of the system energy and the equation of the extensible conditions in the FEM.4 Formation of Static Problem The equations (3. The equation (3.
59) where Fil is a static forcing term formed by gravity force.62) (3.56 Ril = 0 Gm = 0 (3. drag force and uniform current and the other applied static force on the line.61) And. ∂Ril 1 2 = K ijlk + λn K nijlk ∂U jk ∂Ril 2 = K nijlk U jk ∂λn ∂Gm = 2 AmklU jk ∂U jk ∂Gm = −C mn ∂λn (3.58).57) and (3.58) In the static problem.64) (3.60) ( ( G mn +1) = Gmn ) + (3. with neglecting the higher order terms.36). It is a nonlinear force vector.57) (3. NewtonRaphson’s iterative method is used. It becomes as: 1 2 Ril = ( K ijlk + λ n K nijlk )U jk − Fil (3.65) . Using the Taylor series expansion. the equation (3. For solving the equation.63) (3. can be expressed by: ( ( Riln +1) = Riln ) + ∂Ril ∂R (∆U jk ) + il (∆λn ) = 0 ∂U jk ∂λn ∂Gm ∂Gm (∆U jk ) + (∆λ n ) = 0 ∂U jk ∂λ n (3. the dynamic term is removed in the equation (3.
74) ( n 1 2 Riln ) = ( K ijlk + λ n K nijlk )U (jk ) − Fil ( ( ( Gmn ) = AmilU kin )U kln ) − Bm − C mn λ( n ) = 0 n After renumbering.65) and is rewritten.57 If the equation (3.70) t 0( Dmjkn ) = n AmkpU (jp ) L n ′ = Pm Ak A′p ds U (jp ) 0 ∫ (3.68) (3.66) (3.67) ( 2 AmklU jl (∆U jk ) − C mn (∆λ n ) = −Gmn ) They can be rewritten in matrix form as follows: K t 0( n ) K t1( n ) ∆U − R ( n ) i ln jk il ijlk = t 0( n ) t1( n ) ∆λ ( D Dmn n − G mn ) mjk where.61) is rearranged by replacing the equation (3.73) (3.69) K it1( n ) ln = n 2 K nijlk U (jk ) L ′ n = Pn Al′ Ak ds U (jk ) 0 ∫ (3.72) (3.75) . they are given by: ( 1 2 2 ( K ijlk + λ n K nijlk )(∆U jk ) + ( K nijlk U jl )(∆λ n ) = − Riln ) (3. (3.71) t1( Dmnn ) = −C mn 1 =− AE ∫ P P ds m n 0 L (3. t 0( 1 2 K ijlk n) = K ijlk + λ( n ) K nijlk n (3.60) and (3.62) to (3. the assembly equation in matrix form is given by: K ( n ) (∆y ) = F ( n ) where.
and the total number of equations is (N + 1 ) × 8 − 1 . and M = L × r ′ is the nodal resultant moment. L = L1 L2 L3 } is the force relating to the T T { { nodal resultant moment.77) (3. In every time step.58 − [λr1′ + ( Br1′′ ′] Al s =0 ) [1] − N1 ′′ Al′ s =0 [ EIr1 ] [1] − L1 − [λr2′ + ( Br2′′)′] Al s =0 − N [1] [ EIr ′′] A′ [12] 2 s =0 − L2 − [λr3′ + ( Br3′′)′] Al − [1] s =0 N3 [ EIr3′′] A′ s =0 − L[31] 0 0 = 0 0 Fr = ) [λr1′ + ( Br1′′ ′] Al s = L N1[ 2 ] [ 2] ] − [ EIr1′′ A′ s = L L1 [ 2] [λr2′ + ( Br2′′)′] Al s = L N 2 [ 2] − [ EIr ′′] A′ L2 2 s= L [ 2] [λr3′ + ( Br3′′)′] Al N3 s=L L[32 ] − [ EIr3′′] A′ s = L 0 0 (3. The Gauss elimination method for solving .79) F T = [R11 R12 R21 R22 R31 R32 G1 G2 R13 R14 R23 R24 R33 R34 G3 ] y ( n+1) = y ( n ) + ∆y where [1] denotes the first end of element.76) y T = [ U 11 U 12 U 21 U 22 U 31 U 32 λ1 λ2 U 13 U 14 U 23 U 24 U 33 U 34 λ3 ] (3. N = N 1 N 2 N 3 } is the nodal resultant force. the stiffness K and the force vector F are recalculated to solve ∆y . and [2] the second end of element. The stiffness matrix is the symmetric and banded matrix. The bandwidth of the assembled stiffness matrix is 15.78) (3. where N is the number of elements for a line.
82) (3. the iterative solution scheme is used to get ∆y until it becomes smaller than a given tolerance.5 Formulation for Dynamic ProblemTime Domain Integration The equation of motion.81) Gm = AmilU klU ki − Bm − C mn λ n = 0 where. a ˆ M ijlk = M ijlk + M ijlk 1 ˆ Fil = − Fil − Fil2 + Fil 1 1 Fil = K ijlk U jk 2 Fil2 = λ n K nijlk U jk (3.80) 3.85) .84) (3. and the equation (3. In addition. The order of the equation (3.47) and the stretch condition (3.82) is an algebraic equation.81) is the second order differential equation.83) The equation (3.75) conforming the symmetry and band is used.81) is derated using the first derivative of the displacement of the rod. (3. so that the equation results in two first order differential equations as follows: ˆ & ˆ M ijlk V jk = Fil & U jk = V jk (3. F r = −F ( n+1) (3. The resultant force can be obtained from force vector F r .48) can be rearranged. 1 2 && ˆ M ijlk U jk = −( K ijlk + λ n K nijlk )U jk + Fil ˆ = Fil (3.59 the equation (3.
86) ∫ t(n) & U jk dt = ∫ t ( n +1) t(n) V jk dt (3. Then the integration is achieved with the 2nd order accuracy: 1 ( n+ ) ( ˆ M ijlk 2 V jkn+1) 1 ( n+ ) ( ˆ − M ijlk 2 V jkn ) = ∫ t ( n +1) t (n) ˆ F jl dt (3.87) is obtained from the following sequential calculations: n n U (jk +1) = U (jk ) + ∆t ( n+1) ( V jk + V jkn ) 2 ( ) ) (3. 2 ˆ When the time step is (n + 1 ) .91) .88) ( The V jkn+1) of the equation (3. then they are given by: ∫ t ( n +1) t (n) t ( n +1) ˆ & M ijlk V jk dt = ∫ t ( n +1) t (n) ˆ F jl dt (3.90) ( V jkn+1) = 2 ( (∆U jk ) − V jkn ) ∆t (3.87) ˆ In the equation (3.89) n n ∆U jk = U (jk +1) − U (jk ) = ∆t ( n +1) ( V jk + V jkn ) 2 ( (3.86).60 If the two equations are integrated.86). a constant mass is ˆ newly introduced. M ijlk is not a constant with respect to the time. M ijlk 2 is the mass at time t 1 ( n+ ) 1 ( n+ ) 1 ( n+ ) 2 = t (n) + ∆t and a constant mass. since it includes the added mass term. In order that the time integration is possible. M ijlk 2 can be used for the integration of the equation (3.
∆λ n = 1 (n+ ) λn 2 1 (n− ) − λn 2 .92) consists of three parts of integration: ∫ ∫ ∫ t ( n +1) (n) t ( n +1) t ( n) ˆ F jl dt = − ∫ t ( n +1) t (n) 1 Fil dt − ∫ t ( n +1) t (n) Fil2 dt + ∫ t ( n +1) t ( n) F jl dt (3.93) is the gravitational force and the hydrodynamic force.92) The integration of the right hand side of the equation (3. each term of the equation (3.95) where.61 2 to both sides.94) t ( n +1) t(n) Fil2 dt = ∆t 2 ( n+1) Fil + Fil2 ( n ) 2 ∆t ( n +1) 2 2 = λn K nijlkU (jkn +1) + λ(nn ) K nijlkU (jkn ) 2 ( ) [ 1 1 ( n+ ) ∆t ( n + 2 ) 2 ( n +1) 2 n ≈ λn K nijlkU jk + λn 2 K nijlkU (jk ) 2 1 1 ( n− ) ∆t ( n − ) 2 n 2 n 2 = 2λn 2 K nijlkU (jk ) + 2 K nijlkU (jk ) (∆λn ) + λn 2 K nijlk (∆U jk ) 2 (3. The third term of the right hand side of the equation (3.91) and multiplying rewritten as: 4 ˆ M ijlk 1 (n+ ) 2 ( ∆U ∆t 2 4 ˆ (n+ 2 ) (n) 2 M V jk + jk ) = ∆t ijlk ∆t 1 ∫ t ( n +1) (n) t ˆ F jl dt (3.93) is given by: 1 Fil dt = t ∆t 1( n +1) 1 Fil + Fil ( n ) 2 ∆t 1 n 1 = K ijlk (∆U jk ) + 2 K ijlk U (jk ) 2 ( [ ) ] ] (3. The gravitational force is a .93) If the trapezoidal integration rule is applied.88) can be ∆t Using the equation (3. the equation (3.
96) into the equation (3. The hydrodynamic force can be calculated by applying Morison’s formula and the AdamBashforth explicit integration scheme: ∫ ∆tF (0) . for step 1 il Fil dt = ∆t ( n) ( n −1) t (n) . The time integration of the equation (3. ˆ t 0( ˆ t1 ˆ( K ijlk n ) (∆U jk ) + K lin( n ) (∆λ n ) = − Riln ) (3.82): ( ( 0 = 2Gmn+1) ≈ 2Gmn ) + 2 ( ∂Gmn ) ∂G ( n ) (∆U jk ) + 2 m (∆λn ) ∂U jk ∂λn ( 2 = 2Gmn ) + 2 K mijlkU il (∆U jk ) − 2Cmn (∆λn ) t1 ( ˆ t 0( = 2Gmn ) + Dmjkn ) (∆U jk ) + 2 Dmn( n ) (∆λn ) (3.99) Using the equation (3. the equation of motion and the stretching condition can be written as follows.98) By applying Taylor expansion to the stretching condition of the equation (3.94) to (3.97) and (3.97) = 4 ∆t 1 (n+ ) ( ˆ M ijlk 2 V jkn ) 1 ( n+ ) 2 + 3Fil( n ) − ( 1 ( n− ) n n 1 2 Fil( n−1) − 2 K ijlk U (jk ) − 2λ n 2 K nijlk U (jk ) ) The mass at time t by: = t (n) + ∆t is approximated using the AdamBashforth method 2 1 ( n+ ) ˆ M ijlk 2 = 1 ˆ ( n ) ˆ ( n−1) 3M ijlk − M ijlk 2 ( ) (3. for other steps 3Fil − Fil 2 t ( n +1) ( ) (3.99).100) .62 constant with time.92) is represented by: 1 1 4 ( n+ ) ( n− ) 2 + K1 + λ 2 K 2 ( ∆U ) + 2 K 2 U ( n ) ( ∆λ ) ˆ M ijlk ijlk n nijlk jk nijlk jk n ∆t 2 (3.93).96) The integration of force can be obtained by replacing the equations from (3.
The interaction effect between the line and seafloor acts the important role on the line . ( n− ) 2 1 2 ˆ t0 ˆ (n ˆ (n K ijlk( n ) = 2 3M ijlk) − M ijlk−1) + K ijlk + λn 2 K nijlk ∆t 2 n ˆ t1 K lin( n ) = 2 K nijlkU (jk ) 2 ( t 0( ˆ t 0( Dmjkn ) = 2 K nijlkU iln ) = 2 Dmjkn ) (3.110) 3.107) (3.106) t1 ˆ t1 Dmn( n ) = −2C mn = 2 Dmn( n ) 2 ( ˆ( ˆ (n ˆ (n Riln ) = 3M ijlk) − M ijlk−1) V jkn ) + 3Fil( n ) − Fil( n−1) ∆t 1 n − 2 K ijlkU (jk ) − 2λ ( ˆ( Gmn ) = 2Gmn ) 1 ( n− ) 2 n 2 n K nijlkU (jk ) ( ) ( ) (3.108) The total equation in matrix form is written by: ˆ ˆ K ( ∆y ) = F ˆ F r = −F ( n+1) at time step n (3.109) (3.104) (3.103) (3.105) (3.63 ˆ( ˆ t 0( ˆ t1( Dmjkn ) (∆U jk ) + Dmnn ) (∆λ n ) = −Gmn ) If the equation is written in matrix form.102) ( ) 1 (3.6 Modeling of the Seafloor The anchors are used for fixing the mooring lines and risers on the sea floor.101) ˆ t1 K lin( n ) ∆U jk − R ( n ) ˆ = il ˆ t1( ˆ( D mnn ) ∆λn − G mn ) (3. it gives: ˆ K t 0( n ) ijlk D t 0( n ) ˆ mjk where.
f1 = 0 . the equation of motion is rewritten by. the dynamic stiffness matrix is modified as: . 0. and the friction force is not considered. f2 = 0 . 1 2 a && ( M ijlk + M ijlk )U jk + ( K ijlk + λ n K nijlk )U jk = Filf + Fil (3. f3 = 3 0. the interaction force f on the line from the sea floor can be expressed as. the seafloor is modeled as an elastic foundation. When the force from the sea floor is added.114) In the static analysis using Newton’s method. = ∫0 l i 3 3 0. for r3 − D < 0 for r3 − D ≥ 0 for r3 − D < 0 for r3 − D ≥ 0 (3.111) where D is the water depth or vertical distance between the sea floor and the origin of the coordinate. in the program. and. Thus. for i = 3 otherwise (Kronecker Delta) (3. and r3 is the zcomponent of the line position vector r . With the origin of the coordinate system located on the mean water surface and zaxis pointing upwards.64 movement.113) δ i3 = 1. L A δ c(δ A U − D) 2 ds . for r3 − D < 0 for r3 − D ≥ 0 (3.112) where Filf = ∫ Al f i ds 0 L L A δ c(r − D) 2 ds . c(r − D) 2 . = ∫0 l i 3 i 3 k jk 0.
and it is finally combined into K ijlk . the dynamic stiffness matrix is modified as: ∫ t ( n +1) t (n) Filf = ∆t f ( n +1) Fil + Filf ( n ) 2 ∆t 3 ≈ K ijlk (∆U jk ) + 2 Filf ( n ) 2 ( [ ) ] (3.116) is added to the RHS of the equation (3.65 ∂Filf ∂U jk 3 K ijlk = (3. Thus. 1 1 4 (n+ ) ( n− ) n 3 2 2 + K1 + λ 2 K2 ˆ − K ijlk (∆U jk ) + 2 K nijlk U (jk ) (∆λ n ) M ijlk ijlk n nijlk ∆t 2 = 4 ∆t 1 ( n+ ) ( ˆ M ijlk 2 V jkn ) + 3Fil( n ) − Fil( n−1) + 2 Filf ( n ) ( ) 1 K 1 + λ( n − 2 ) K 2 U ( n ) − 2 ijlk n nijlk jk (3. In time domain analysis using the trapezoidal rule. The second term in RHS of the equation (3.117) . for δ m3 AnU mn) − D ≥ 0 ∫ 3 t0 This K ijlk is added to K ijlk in order to form the tangential stiffness matrix in the equation (3.97).97).69).115) L (n) (n) 2 Al δ i 3 cδ j 3 Ak (δ m3 AnU mn − D)ds. for δ m3 AnU mn − D < 0 = 0 (n 0.116) The first term in the RHS of the above equation is added to the LHS of the equation ~ t0 (3.
since the system matrices of body and lines are assembled and solved together. the preobtained body motion cannot be evaluated .1 Introduction The statics and dynamics of the mooring lines and risers can be solved with the given data and the boundary conditions. of the lines are connected to the platform with strong springs. At both ends of the lines. the mooring lines and risers are treated separately to the body motion. in the postprocessing. In the quasistatic analysis.66 CHAPTER IV COUPLED ANALYSIS OF INTEGRATED PLATFORM AND MOORING SYSTEM 4. different boundary conditions are applied. if the cable is installed for the connection of the vessel to vessel (for the multiple body interaction problem). The body behavior is greatly influenced by the movement of the mooring lines and risers. But. the dynamics of the mooring lines and risers are analyzed with the motions of the end nodes that are assumed to be the same amount as the body motion. It has six degrees of freedom. The coupling effect of the body and the lines can be considered. the end nodes are moved with almost the same displacements as the floating platform. The upper ends or the upper/lower ends. The motion of the body is solved first. The other ends of the lines are connected to the anchors on the seafloor and constrained with the fixed conditions in six degrees of freedom. Thus. The platform is concentrated as a single point on the center of the global coordinate and moved as a rigid body. and then.
and solved together. NewtonRaphson’s iterative scheme was used.000 ft. and the scheme is extended to the interaction problem of multiple body systems of floating platforms. because the body motion is analyzed separately without considering the line dynamics. the body and lines are analyzed at the same time. He developed the mathematical formulation to be applied to solving the coupled system. the above numerical methods are also adopted as a numerical tool of the main solver. for the timedomain analysis. As the water depth gets deeper and deeper. the AdamBashforth method was adopted as an explicit numerical scheme. The coupling effects were studied by Ran(2000). in the coupled analysis. in depth. the applied force and moment to come from lines directly affects the body. the formulations of the forces and moments to be transferred to the body from the lines is given by: . If the spring is strong enough. All dynamic effects of body and lines are included in system matrices. In his study. the inertia effect increases.67 properly to consider the inertia effects and the hydrodynamic loads on the lines. for static analysis. But. 4. On the contrary.2 The Spring to Connect the Platform and the Mooring System The end connection is modeled numerically by the translational and rotational springs between the body and lines. In this study. The coupled analysis is to be an essential tool for solving the floating platform motion and line dynamics in ultra deep water over 8. The stiffness should be considered strong enough so that the body reacts with the same amount of motion as the lines’ in six DOFs (degrees of freedom). the interaction effect greatly influences body and line motions. If the angular motion is assumed small. So.
r1′ = U 12 .68 N iS = K iL (X i + pi + θ j × pk − ri ) (4. and ei is a unit vector of the reference direction of the rotational spring. z direction and in the θ x . r2′ = U 24 . X i and θ j are the translational and rotational motions of the body. r2′ = U 22 .2) S where N iS = N1S N 2S N 3S and LS = L1 LS LS are the nodal resultant forces and moments 2 3 i θ θ θ on the end node of the line.6) C ji and D ji are defined to make easy the numerical manipulation of the vector product with the position vector pi and the unit vector ei as: . ri′ is the space derivative of the position vector ri . ri is the position vector of the ending or the starting node of the line attached by the spring to the body. r2 = U 21 . r3 = U 31 r3′ = U 32 (4.5) (4. r3 = U 33 r3′ = U 34 (4. K iL = K1L K 2L K 3L and K iθ = K1 K 2 K 3 are the translational and the rotational spring constants in the x.3) (4. r1′ = U 14 . pi is the position vector of the node of the body connected to the spring. The ri vector at the end node of the line is defined as: When the connection point is the starting point of the line: r1 = U 11 . y.θ z direction.θ y . r2 = U 23 .4) When the connection point is the ending point of the line: r1 = U 13 .1) ri′r j′ ri′ LS = K θ ei + θ j × ek − − i 1/ 2 ′ ′ (rm rm ) (rn′rn′ )3 / 2 T (4.
8) are used in equations (4. the . the equations are rewritten as: N iS = K iL (X i + pi + θ j C ji − ri ) ri′r j′ ri′ − LS = K θ ei + θ j D ji − i 1/ 2 ′ ′ (rm rm ) (rn′rn′ )3 / 2 (4.2).9) (4.1 Static Analysis The connector force and moment on the end node of the line are included in the equation of motion of the integrated system as external forces.8) If the equations (4.1’) (4.7) and (4.7) (4. and M iθ = LS × e j k is the moment resulting from the rotational spring.1) and (4. In the static analysis. 4.69 0 − p3 p 2 [C ] = − p3 0 − p1 p2 − p1 0 0 − e3 e2 [D] = − e3 0 − e1 e2 − e1 0 (4.2.10) where M iL = N kS × p j is the moment resulting from the linear spring.2’) The resultant force Fi S and moment M iS transferred to the body are defined as follows: Fi S = − N iS M iS = M iL + M iθ = N kS C ki + LS Dki k (4. The force Fi S and the moment M iS act on the body.
11) (4.13) rX K ij = − r K ijθ = − r K ij ′r′ = − θθ K ij = − These equations that shows forces and moments will be expressed with the coupled terms between body and line motions.15) .12) ( n +1) (n) r r + K ij ′r ′ ∆r j′ + K ij ′θ ∆θ j ∂N iS = K iLδ ij ∂r j ∂N iS = − K iLδ ij ∂X j ∂N iS = − K iL Cij ∂θ j ri′r j′ δ ij ∂LS i = K iθ − ′ ′ 1 / 2 (rn′rn′ ) 3 / 2 ∂r j′ (rm rm ) ∂LS i = − K iθ Dij ∂θ j (4. so that the force and moment in (n+1) iteration are approximated as follows: For ri : For ri′ : Where.14) (4. the connector force and moment on the rigid body at iteration (n+1) are approximated as follows using Newton’s method: For X i : For θ i : Where. Fi ( n+1) = Fi ( n ) + K ijXr ∆r j + K ijXX ∆X j + K ijXθ ∆θ j θ θ θθ M i( n+1) = M i( n ) + K ijr ∆r j + K ijr′ ∆r j′ + K ij ∆θ j (4. Similarly. rr K ij = − N iS LS i ( n +1) = N iS = LS i ( n) θθ rr rX + K ij ∆r j + K ij ∆X j + K ij ∆θ j (4.70 NewtonRaphson method is applied.
r r θ θ K ijθ . the coupled assembly system equations are solved to obtain the behaviors for the body and lines simultaneously.2. K ijXX . K ijr .16) K ijXX = − K ijXθ = − θ K ijr = − θ K ijr′ = − θθ K ij = − [ ] rr r The stiffness coefficients K ij and K ij ′r′ are added the stiffness matrix of elements. At each iteration step. and the iteration continues until the norms of the solutions reach a specified tolerance.2 TimeDomain Analysis The integrations from time t (n ) to t ( n+1) of the connector forces and moments on the end node of the lines are expressed by applying Newton’s method as: . 4. K ij . K ij ′θ . form the coupling terms in the assembled system matrix as the symmetric matrices.71 K ijXr = − ∂Fi = K iLδ ij ∂r j ∂Fi = − K iLδ ij ∂X j ∂Fi = − K iL Cij ∂θ j ∂M i = K θ C ji j ∂r j ri′r j′ δ ij ∂M i = Kθ − j D ji 1/ 2 ′ ′ ∂r j′ (rn′rn′ ) 3 / 2 (rm rm ) ∂M i = − K L Cki Ckj + K θ Dki Dkj j j ∂θ j (4. θθ rX K ijXθ and K ij are included in the stiffness matrix of the platform. The other terms. and K ijr′ .
4. The damper is modeled as a linear damping force proportional to the vibratory velocity of the .16) in the static analysis.20) Where the notations and the expressions for the K matrices follow the same convention as the equations (4.13) and (4.18) The integrations from time t (n ) to t ( n+1) of the connector forces and moments on the rigid body are expressed as: For X i : ∫ t ( n +1) t(n) Fi dt = ∆t ( n+1) Fi + Fi ( n ) 2 ∆t = − K ijXr ∆r j − K ijXX ∆X j − K ijXθ ∆θ j + 2 Fi ( n ) 2 ( ) ( ) (4.17) For ri′ : ∫ t ( n +1) (n) t LS dt = i (n) ∆t S ( n+1) Li + LS i 2 (n) ∆t r r = − K ij ′r′ ∆r j′ − K ij ′θ ∆θ j + 2 LS i 2 ( ( ) ) (4.3 Modeling of the Damper on the Connection The damper on the connector is used for controlling the excessive resonance of the high frequency vibration of the tensioned line like the tether or the riser in the TLP.72 t ( n +1) (n) ∆t S ( n+1) Ni + N iS 2 (n) ∆t rr rX r = − K ij ∆r j − K ij ∆X j − K ijθ ∆θ j + 2 N iS 2 For ri : ∫ t(n) N iS dt = ( ) ( ) (4.19) For θ i : ∫ t ( n +1) t(n) M i dt = ∆t M i( n+1) + M i( n ) 2 ∆t θ θ θθ = − K ijr ∆r j − K ijr′ ∆r j′ − K ij ∆θ j + 2M i( n ) 2 ( ) ( ) (4.
1’). r is the velocity of the attached node of the line to the body. pk is the position vector of the attached node of the line at the connection point. There are coupled terms with the body and the lines on the connection point.21) becomes: & & & N iD = C d X i + θ j C ji − ri It acts on the rigid body as reaction force by: Fi D = − N iD ( ) (4. . N iD . and the & & & vector product of the θ j and p k can be rewritten in the tensor form as θ j × p k = θ j C ji . The coupled terms can be solved together for body and line motions in the assembled system matrix equations. the equation (4. the integration from time t ( n+1) to t (n ) is obtained as: For ri : ∫ t ( n +1) t(n) t & & & N iD dt = ∫ ( n ) C d X i + θ j C ji − ri dt t ( n +1 ) ( ) = C d ∆X i + C d C ji ∆θ j − C d ∆ri (4.21’) (4.21) & & where C d is the damping coefficient.23) and (4. on the connection node of the line is given by: & & & N iD = Cd X i + θ j × pk − ri ( ) (4. X and θ are the translational and rotational & velocity of the rigid body.73 line on the top connection node of the body and the line.24) The equations of (4. as shown in the equation (4. So. The damping force.22) In the time domain analysis.23) For X i : ∫ t ( n +1) t(n) t & & & Fi D dt = ∫ ( n ) Cd − X i − θ j C ji + ri dt t ( n +1 ) ( ) = −Cd ∆X i − C d C ji ∆θ j + Cd ∆ri (4.24) show the terms of the geometric stiffness matrix of the system.
the position vector is assigned as: r1 = U 11 . r1′ = U 12 . If the end connection of the line consists of the anchor.25) (4. r2′ = U 22 .29) .4 Modeling the Connection between Lines and Seafloor The lower ends of the mooring lines and risers are normally connected to the seafloor. and ri and r ′ are the position vector and the tangential vector of the attached node to the seafloor. the clamped or hinged boundary condition is needed. The connector force N iF and moment LF are defined by: i N iF = K iL piF − ri ( ) (4.74 4. and then it can be obviously replaced by considering a proper spring so that the spring constant in the corresponding direction is to be large enough to hold the rigidity of the anchor or the hinged boundary sufficiently. r2 = U 21 . eiF is the reference direction vector of the rotational spring fixed on the seafloor. r3 = U 31 r3′ = U 32 (4. The formulation for the connection part of the lines and the seafloor are very similar to the modeling of the connection part of the body and the line.26) ri′r j′ ri′ LF = K θ eiF − − i 1/ 2 ′ ′ (rm rm ) (rn′rn′ )3 / 2 The damping force is defined as: & N iFd = − K iL ri (4.28) (4. Since the numbering of the lines starts from the seafloor when the line is attached to the seafloor.27) where piF is the position vector of the attached point of the seafloor.
5 Formulation for the Multiple Body System The equation of motion and the equation of the stretching condition for the multiple body system combined with any types of vessels can be derived in the same way as the equation (3. {F} and {λ} are the vectors of the size of N L × [8 × ( N E + 1) − 1] . and they can be simplified as follows: && MU + KU = F (4. {U 2 } .47) and (3. After applying the Taylor expansion.31) AU 2 − B − Cλ = 0 The [M ] . 1 2 a && ( M ijlk + M ijlk )U jk + ( K ijlk + λ n K nijlk )U jk − Fil = 0 (3. and the Newton method of static and dynamic analysis. and the assembled matrix and system equations are dealt with in the next section.30) (4. the equations can be expressed in the matrix form as: .48) for a single body system. In the next step. the matrix of equations for the lines is combined with the matrix for the body motion including the coupled terms in the stiffness matrix.75 4. [K ] . the AdamsMoulton method. {U}. and the AdamsBashforth method. The global coordinate is used for composing each matrix.48) (3. regardless of the body to which the line is connected. {U}.49) G m = Amil U klU ki − Bm − C mn λ n = 0 The two equations for a multiplebody system has the same form. and [B ] . {U}. [A ] and [C] have the size of rows N L × [8 × ( N E + 1) − 1] and the && && bandwidth of 15. where N L is the total number of lines and N E is the number of elements per each line.
32) where.76 In static analysis: t0 1 K ijlk( n ) K itln( n ) ∆U jk − Riln ) ( = t 0( n ) t1( n ) (n) Dmjk Dmn ∆λn − Gm (4.34) ( ) 1 (4.35) 2 ( ˆ( ˆ (n ˆ (n 3M ijlk) − M ijlk−1) V jkn ) + 3Fil( n ) − Fil( n−1) Riln ) = ∆t − n 1 2 K ijlk U (jk ) ( ) ( ) − 1 ( n− ) n 2 2λn 2 K nijlk U (jk ) ( ˆ( Gmn ) = 2Gmn ) . ( n− ) 2 1 2 ˆ t0 ˆ (n ˆ (n K ijlk( n ) = 2 3M ijlk) − M ijlk−1) + K ijlk + λn 2 K nijlk ∆t n 2 ˆ t1 K lin( n ) = 2 K nijlk U (jk ) ( t 0( 2 ˆ t 0( Dmjkn ) = 2 K nijlk U iln ) = 2 Dmjkn ) t1( ˆ t1( Dmnn ) = −2C mn = 2 Dmnn ) ˆ1 K itln( n ) ∆U jk − Ril( n ) ˆ = ˆ( ˆ t1 Dmn( n ) ∆λn − Gmn ) (4. 1 2 t0 K ijlk( n ) = K ijlk + λ(nn ) K nijlk 1 2 K itln( n ) = K nijlkU (jkn ) t0 Dmjk( n ) = AmkpU (jpn ) t1 Dmn( n ) = −Cmn 1 2 Ril( n ) = ( K ijlk + λn K nijlk )U (jkn ) − Fil ( Gmn ) = 0 (4.33) In the dynamic analysis in time domain: ˆ t0 K ijlk( n ) ˆ t0 Dmjk( n ) where.
. where N is the number of the multiple bodies. For a singlebody system. and the remaining terms subtracting the nontrivial terms from [K C ] are filled with zeros. [K C ] has the size of [8 × (nE + 1) − 1] rows and 6 columns per line... [U L ] and [U B ] denote the displacement matrices of the lines and the body..77 The assembled equation of the coupled system of the rigid body and the lines can be expressed as: [ K L ] [K C ] U L F L . and [F ] and [F ] are the force and moment terms acting on the lines and the body. = ... The matrix (K C ) [ T ] is the transpose matrix of [K C ]... The size L B of [K B ] is 6× 6 for a single body system.... but for the multiplebody system 6 N × 6 N . . (K C )T [K B ] U B F B [ ] (4. When the multiplebody system is considered. and the hawser or the fluid transfer line (FTL) between one body and another body is connected to body.. [K ] is the stiffness matrix of the rigid body. the total number of rows of the matrix [K ] C becomes [8 × (nE + 1) − 1] rows and 6 × N columns per connecting line. It has nontrivial terms of the size of 7 × 6 at the last end rows of the matrix. [K ] and [(K ) ] are the coupled stiffness B C C T matrices and its transpose matrix including the coupling terms of the rigid body and the lines.. It makes two coupled terms on the starting node and the ending node of the connecting line..36) where [K L ] is composed with the stiffness matrix of the lines and the connector springs.. where nE is the number of elements per line.
After the forward elimination process is performed in the first step for solving the system matrix. [K L ]. For this study.78 Thus. The matrix equation of total system explicitly has the sparse matrix form. . of the lines has nL × [8 × (nE + 1) − 1] rows and the bandwidth of 15. it has the nontrivial terms twice of 7 × 6 N in size. and the remaining terms except the nontrivial terms are filled with zeros like those in a single body. the backward substitution follows it next. Nowadays. some updated sparse matrix solvers are developed and announced by many mathematical researchers. and modified slightly for the purpose of treating the sparseness of the system matrix effectively. The displacement vector [U B ] and the force vector [F B ] for the rigid body have the size of 6 N × 1 . the forward and backward Gauss elimination method as the rigorous and traditional solver is used. It means that a special consideration should be required to solve it. The stiffness matrix. where nL is the total number of lines.
Using this program. and the experimental data of many years. the following case studies were performed for verification of the program. rotational movement during operation in the sea. The wave loads induced by potential velocities are calculated by using WAMIT that is a program to solve the potential problem of the fluid interaction.000 ft. The test model is selected as a turret moored FPSO in 6. The tankerbased FPSO is designed for the purpose of installation in the sea at the water depth of 6. The environmental conditions of the GoM (Gulf of Mexico) are used for the design. . The results of the analysis are compared with MARIN’s experimental results. due to this kind of specific large yaw rotation.79 CHAPTER V CASE STUDY 1: DYNAMIC ANALYSIS OF A TANKER BASED FPSO 5. of water depth. where the environmental conditions are the extreme hurricane conditions in the Gulf of Mexico. the current and the wind force coefficients are specially considered.1 Introduction As mentioned in the previous chapter. based on many VLCCs investigated and developed by Oil Company International Marine Forum (OCIMF) is used. a tankerbased FPSO is taken.000 ft. In general. The FPSO has a large. The mooring system is a semitaut steel wire system. the hull/mooring line/riser coupled analysis program for solving the twobody interaction problem was developed. For the first case.
only 13 risers among them are modeled equivalently as to what .1. There are 4 groups of mooring lines. The vessel for this study is an FPSO in 6. and so the second group is toward the true North. wire and chain. In the figure. On the contrary. but for the simulation. The details of the design premise data are shown in Table 5.9 meters and the displacement is 240. for the riser system. and the hydrodynamic coefficients are depicted in Table 5. and 13 steel wire risers.000 ft of the water depth. The turret is located at 63. 19 lines are used in the prototype FPSO. The schematic plot of the arrangement for mooring lines is shown in Figure 5.55 meters aft of the forward perpendicular of the vessel.000 bpd.440. This vessel has an LBP of 310 meters.869 MT.1. The center of the first group is heading the true East. The body plan and the isotropic view of the vessel are shown in Figure 5. The mooring lines and risers are spread from the turret. Each mooring line has a studless chain anchor of grade K4. The main particulars of risers are shown in Table 5. The dead weight of this vessel is 200 kDWT.17 meters. and the production level is 120.2 shows the main particulars of mooring lines.2 Design Premise Data of FPSO and Mooring Systems The design premise data is described in this section.4.04 meters as the main dimensions. The capacity of the vessel storage is 1.000 bbls.5. Table 5. Table 5. and a depth of 28.80 5. each of which is normal to the other group. Each group is composed of 3 mooring lines 5 degree apart from each mooring line in the group. a molded breadth of 47. the bow of the vessel is heading toward the east.3 gives the hydrodynamic coefficients for mooring lines. In the full load condition. There are 12 combined mooring lines with chain. the draft is 18.2.
000 ft Description Production level Storage Vessel size Length between perpendicular Breadth Depth Draft (in full load) Diaplacement (in full load) Lengthbeam ratio Beamdraft ratio Block coefficient Center of buoyancy forward section 10 Water plane area Water plane coefficient Center of water plane area forward section 10 Center of gravity above keel Transverse metacentric height Longitudinal metacentric height Roll raius of gyration in air Pitch raius of gyration in air Yaw radius of gyration in air Frontal wind area Transverse wind area Turret in center line behind Fpp (20.09 240.6 and Figure 5.85 m m m .55 1.81 MARIN did in their experimental tests. Table 5.6 13. The risers are arranged nonsymmetrically with respect to the xaxis (the axis toward the East).5 % Lpp) Turret elevation below tanker base Turret diameter L/B B/T Cb FB A Cw FA KG MGt MGl Rxx Ryy Rζζ Af Ab Xtur Ztur m m m m m m m m m 2 2 Symbol Unit bpd bbls kDWT Quantity 120. The top view of the arrangement of risers is shown in Table 5.400 0.9164 1.0 13.57 2.000 1.772 63.440. With respect to the yaxis (the axis toward the North).78 403. the arrangement is also not symmetrical.52 15.0 47. the riser bending stiffness is not considered.3 on the horizontal plane based on the earth.83 14.30 1.17 28.5 0.85 Lpp B H T m m m m MT m m 2 6.47 79.77 77.1 Main particulars of the turret moored FPSO 6.04 18. In this study.000 200 310. But the risers are almost balanced in the viewpoint of top tension with respect to both axes.869 6.012 3.32 5.
1 The body plan and the isotropic view of FPSO 6.000 ft .82 Figure 5.
6 .4 88.515 deg m m Unit kN Quantity 1.9 164.2 0.4 794. rope and polyester Hydrodynamic Coefficients Normal drag Tangential drag Normal added inertia coefficient Tangential added inertia coefficient Coulomb friction over seabed Symbol Cdn Cdt Cin Cit F Chain 2.9 42. AE Mean breaking load.168 6.65 2. MBL Segment 3 (hangoff position): chain Length Diameter Weight in air Weight in water Stiffness.9 143.2 Main particulars of mooring systems Description Pretension Number of lines Degrees between 3 lines Length of mooring line Radius of location of chain stoppers on turn table Segment 1 (ground position): chain Length at anchor point Diameter Weight in air Weight in water Stiffness.841 6.7 690.0 Table 5.4 794.2 0.421 m mm kg/m kg/m kN kN 914.50 1.087.15 0.0 Rope/Poly 1.8 107.9 7.0 35. AE Mean breaking load.9 143. MBL Segment 2: Polyester Length Diameter Weight in air Weight in water Stiffness.515 m mm kg/m kg/m kN kN 1127.3 Hydrodynamic coefficients of the chain.45 0.201 4*3 5 2. MBL m mm kg/m kg/m kN kN 45. AE Mean breaking load.841 6.00 0.83 Table 5.3 1.7 88.9 164.
88 4.8 453.0 0.5 386.5 609.08E+07 1.0 1.88 4.414 1.9 Out Stiffness.52 1.14E+07 8.7 2020.88 m 1.4 Dry weight/ wet weight N/m 1927/1037 1708/526 2803/1898 1810/1168 1358/423 1.7 184.52 m 1829 1829 1829 1829 1829 Table 5. diameter AE mm 444.84 NORTH #6 #5 #4 #7 #8 #9 #3 #2 #1 EAST #10#11 #12 Figure 5.6 .88 4.52 1.88 4.52 1.83E+07 1.86E+07 3.2 Arrangement of the mooring lines for FPSO 6.0 1.5 Hydrodynamic coefficients of risers Description Normal drag Tangential drag Normal added inertia coefficient Coulomb friction over seabed Symbol Cdn Cdt Cin F Coefficients 1.0 1352.0 342.1 530.4 1.60E+06 Mass kg/m 196.9 287.4 174.5 138.52 1.0 Cdn Radius of riser connection Points on turnable Connection level Total length m 4.1 285.9 kN 1.414 1.4 Main particulars of risers Description Liquid production Gas production Water injection Gas injection Gas export No.000 ft Table 5. 4 4 2 2 1 Top tension kN 1112.0 0.
3 Environmental Data For the loading condition for the analysis. and the overshooting parameter of 2.5. The wave condition is composed of the significant wave height of 12 m.6 Azimuth angles of risers bounded on the earth Description Liquid production (LP) Gas production (GP) Water injection (WI) Gas injection (GI) Gas export (GE) Azimuth angle of riser #1 0 45 165 30 300 (North) X2 LP#14 GP#18 GP#17 (East) X1 #2 90 135 337. the peak period of 14 sec. The wind spectrum of API formulae is taken as . which is one of the severest in the world.5 210 #3 180 225 #4 270 315 WI#21 GI#23 LP#15 LP#13 GI#24 WI#22 GP#20 GE#25 LP#16 GP#19 Figure 5.3 Arrangement of the risers for FPSO 6.85 Table 5.000 ft 5. the 100year extreme hurricane condition at the GoM is used.
7.19 14 JONSWAP ( γ =2.96 m to 91. The velocity of current at the sea surface is 1.0668 m/s. The mean wind velocity at the reference height of 10 m for one hour sustained is 41. since the hurricane condition is more severe than the loop current case.7 Environmental loading condition Description Wave Significant wave height.0668 0.96 m at 91. Table 5. Tp Wave spectrum Direction Wind Velocity Spectrum Direction Current Profile at free surface (0 m) at 60.86 the design condition.44 m under the surface to the sea bottom. the loopcurrent condition will not be applied. The summary of the environmental condition for this study is shown in Table 5.44 m.0668 1.5) deg 180 1) 41. the current is assumed as a onedirectional current. the loop current in the GoM should be considered as a design loading condition. when the GoM environmental condition is applied to the platform design.44 m on the sea bottom Direction Unit m sec Quantity 12. The current speed is uniformly kept 0. and it keeps until 60.96 m to 91. Hs Peak period. While the storm wave and wind arise.05 m/s from 91.12 m/s @ 10m API RP 2AWSD deg 210 1) m/s m/s m/s m/s m/s deg 1.0914 0. however. the current profile is determined by the linear interpolation.12 m/s. But. the current speed is varied from 1. In this study. From 60.05 m/s. .0914 150 1) Remark: 1) The angle is measured counterclockwise from the xaxis (the East). For the intermediate region between 60.0668 m/s to 0.96 m under the sea surface.44 m under the sea surface. The current is mainly induced by the storm.
γ and ω 0 are determined by the following formulae: α = 0. In this study. The formula is given by: S (ω ) = α g ω 2 −5 ω exp − 1 .076( X 0 )−0. The formula is to be derived from the modified PiersonMoskowitz spectrum formula. X 0 = g X Uw 2 (5. g is the gravitational acceleration.1) where α is a parameter related to the prevailing wind field with the wind velocity of U w and a fetch length of X . .0081. The α . When X is unknown. 25 ω 0 −4 γ (ω − ω 0 ) 2 exp − 2 2τ 2ω 0 (5.3) ω 0 = 2π where. α is taken as 0.07 for ω ≤ ω 0 0.33 Uw (5.87 5. the wave frequencies are considered to be between in 0. Figure 5. γ is the overshooting or peakness parameter. and τ is the shape parameter.1 Wave Force The JONSWAP spectrum was developed to define the wave by Hasselman. et al.4) .2 rad/s and 1.4 shows the wave spectrum with the given data.3.5 rad/s. (1973) for the Joint North Sea Wave Project.22 (5.2) τ = 0.09 for ω > ω 0 g ( X 0 )−0.
5 f / f ] p f / fp 5/3 σ ( z) 2 (5.6 Frequency (rad/s) Figure 5.3.2 1.0 0.6 0.19 m/s.2 0.8 1.88 JONSWAP Spectrum (Hs=12. i.2 Wind Force The formulae of API wind spectrum is as follows: S uu ( f ) = where: [1 + 1. Tp=14 s) 60 Power Spectral Density (m^2sec) 50 40 30 20 10 0 0.4 0.025V z / z = the average value of the frequencies of the measured wind spectra σ (z ) = the standard deviation of wind speed. f = the frequency in hertz.0 1.4 JONSWAP wave spectrum 5. . f p = 0.4 1.5) S uu ( f ) = the spectral energy density at elevation z.e.
6) V z = V H ( z / H ) 0. API Wind Spectrum (V z=41.275 for z > z s (5.125 = the mean wind speed at elevation z for one hour VH I ( z) = = the mean wind speed at elevation 10 m for one hour σ ( z) Vz 0.5 shows the API wind spectrum of the given wind speed at the reference elevation.89 σ ( z) = I ( z) Vz (5.125 for z ≤ z s = 0. the actual wind force varying with the weathervaning angle (yaw) of the vessel should be reestimated by considering the force coefficients of the wind and the current in the OCIMF booklet. After the normal wind force is calculated using the above wind spectrum.15( z / z s ) −0.15( z / z s ) −0.5 API wind spectrum .12 m/s at 10 m) 12 10 8 6 4 2 0 0 5 10 15 20 25 30 35 40 45 50 Frequency (Hz) PSD of Wind Speed (m/s)^2sec Figure 5.7) = turbulence intensity over one hour where z s = 20 m is the thickness of the surface layer. Figure 5.
12) M xyc = (5. AT . .8) (5.90 5. ρ w and ρ c are the densities of air and fluid. The OCIMF is the international research committee that has been investigated the wind and current foresee subjected on VLCC.10) 1 C xc ρ cVc2 LPP T 2 1 C yc ρ cVc2 LPPT 2 1 C xyc ρ cVc2 L2 T PP 2 (5. and C xc . and Vw and Vc are the wind velocity and current speed at the free surface. the OCIMF booklet published in 1998 is referred to for calculating the wind and current force coefficients. C yc and C xyc are the current force and moment coefficients.9) M xyw = Fxc = Fyc = (5.3. M xyw is the yaw wind moment. C yw and C xyw are the wind force and moment coefficients. AL . and M xyc is the current yaw moment. In this study.3 Wind and Current Forces by OCIMF The FPSO is a kind of tankerbased vessel.13) where Fxw and Fyw are the surge and sway wind forces. They suggest the following formula of the wind and current force coefficients: Fxw = Fyw = 1 C xw ρ wVw2 AT 2 1 C yw ρ wVw2 AL 2 1 C xyw ρ wVw2 AL LPP 2 (5. Fxc and Fyc are the surge and sway current forces.11) (5. C xw .
91 T and LPP are the transverse area. For the bow shape. The OCIMF formula for the wind and current forces are to be expressed with respect to the center of the vessel. the coefficient for every 5 degree of attack angle is prepared in advance. so the center of the vessel movement should be the center of turret position. Therefore. They surveyed the force and moment coefficients on the varying attack angle. For the current force coefficients. the forces and moments give the localized components acting on the vesselwise coordinate. In this study. the draft and the length between perpendiculars of the vessel. the force and moment coefficients are shown in the variation of the attack angle with parameters of the loading condition and the bow configuration. Whenever the angle exceeds 5 degree. the wind and current force coefficients are recalculated using the premade coefficient data files. The subject vessel is a turretmoored tanker. which is located near the midship. the forces and moments are transferred to the global coordinate components according to the yaw angle at every time . to calculate the global motions of the vessel. But. The considered loading conditions are ballast and full load conditions. the longitudinal area. the water depth to draft ratio is also taken as a parameter. and at every time step during analyzing the yaw angle is swept. for the two loading conditions. respectively. Thus. The attack angle is measured from 180 degree on the bow to 0 degree on the stern. not the center of the vessel. the cylindrical bow and the conventional bulbous bow are taken. In the OCIMF booklet. the tanker area and drag coefficients are assumed unchanged during the time simulation. and for two kinds of bow shapes.
15) Coordinate transformation of force vector: FX F = FY M XY .17) cosθ −1 F = T f = sinθ 0 (5.Local force vectors: Fxw + Fxc Fx f = Fyw + Fyc = Fy M xyw + M xyc M xy − sin θ 0 Fx cosθ 0 Fy 0 1 M xy (5.14) Inverse of rotational matrix: T −1 cosθ − sin θ = sin θ cos θ 0 0 (5.Global force vectors: (5.92 step during simulation.16) . The force and moment are transferred by the inverse of rotation matrix as follows: Rotational matrix: cosθ T = − sin θ 0 sin θ cosθ 0 0 0 1 0 0 1 (5.20) .19) (5.18) Considering the translation of turret position: M XY = Fy xturret + M xy Resultant force vectors: cosθ F = T −1f = sinθ 0 − sin θ 0 Fx cosθ 0 Fy xturret 1 M xy (5.
The WAMIT is the program to solve the velocity potential on the wetted surface around the floating structure based on the potential theory by means of the Boundary Element Method (BEM) using the 3dimensional panel elements. the added mass and linear damping coefficients. Taking Green’s function to satisfy all other boundary conditions in the fluid domain as the weighting function in the integral equation of motion makes it possible to solve the potential in the fluid domain. which can solve the diffraction/radiation and the interaction problem of fluid and the platform structure. and the symmetric condition is used for the potential calculation in WAMIT. In the linear theory.4 Hydrodynamic Coefficients The hydrodynamic coefficients are calculated by using WAMIT.6 and 5. the quadratic transfer functions corresponding to the secondorder difference frequency forces and the secondorder sum frequency forces can be withdrawn. In the numerical model.7. BEM is the numerical technique for considering only the wetted body surface and/or the water free surface instead of considering the whole fluid domain.93 where θ is the yaw rotation angle of the vessel and xturret is the xcoordinate of the turret position in the body (local) coordinate system. By using the second order WAMIT. Only the port side of the vessel is modeled. 5. the number of elements on the body is 1870. The modeling of the subject vessel is shown in Figures 5. and mean drift forces can be obtained from the WAMIT. exciting forces by diffraction potential. Several models with other sized numberings are selected for convergence .
it also does not have much influence on the body or on the mooring system. So. Through the convergence study. For the hydrodynamic coefficients. Newman’s (1974) approximation method is used. the different frequency components are replaced by the mean part of the linear transfer function (LTF).6 Modeling of body surface of FPSO . Y Z X Figure 5. When two frequencies are quite large. the determined model was proved to be proper for the analysis. the different frequency is also large.94 study. the frequency is far away from the natural frequency of the body or mooring system. Then. In this method. It is well known that the difference frequency component of the quadratic transfer function is not sensitive to the frequency when two frequencies are close.
The results are compared with MARIN’s. The mooring lines and risers are modeled for preparing the input data of WINPOSTFPSO. The mooring lines consisted of three parts.7 Modeling of body surface and free surface of the water 5.8 m). So. The first part is divided into 5 elements.000 ft (about 1828. The connection boundary to the turret is modeled as a hinged joint. The water depth is 6. i. The hydrodynamic coefficients are calculated at every 5 degree of yaw angle by WAMIT. the midpart (wire) into eight elements. and the last chain part for the connection to the turret into 1 element.95 X Z Y Figure 5. a wire part of mid and a hangoff chain part. and WIMPOSTFPSO is used for the coupled analysis. the analysis case is explained for the turretmoored FPSO mentioned in the previous section. a chain anchor part. the .e..5 Coupled Analysis of FPSO In this study.
roll. sway. From this test. The boundary conditions for risers are the same as those for mooring lines. but no translation movement is allowed on that point. Through these tests. During this test. the static offset test is carried out for the surge motion. the free decay test is conducted for the surge. the static weight balance with the top tension of mooring lines and risers.0E+07 N. current force and wave loading are described in Table 5.96 rotations are free.7. the vessel weight and the buoyancy are checked. To review the surge stiffness is a measure to judge whether the vessel combined with mooring system is properly modeled or not. the Dirichlet boundary condition is applied. The input data for wind. The surge external force is increased up to the initial force . the footprints of mooring lines and risers are adjusted back and forth. the FPSO is kept heading to 0 degree. Secondly.02 sec. The stiffness of the combined system with the body and mooring system is reviewed as well. heave. At the first node of mooring line on the sea bed. a static and dynamic balancing test should be provided. Before the coupling dynamic analysis is performed. pitch and yaw motion in the calm water and in the 0 degree heading angle of the vessel. the stiffness and system parameters such as natural frequencies and damping factors of the numerical model can be judged whether they are equivalent to the real system or not. Firstly. The initial external force in the direction of the surge motion is set as 2. Until a wellbalanced state is obtained. The risers are divided uniformly into 12 elements. All Risers are treated as Steel Catenary Risers (SCRs). The time interval is defined as 0.
8 Hull drag damping coefficients (Wichers. 1996) .00 #0 #2 #4 #18 #20 Figure 5.15 rad/s to 1.02 sec. and then is released for 2. respectively. For the time simulation.8. and the total time to 3 hours.000 seconds. The damping coefficients for the hull drag forces are depicted in Figure 5. the ramping function is adopted to smoothly increase for 200 sec in order to avoid the peculiar transient state. These are corresponding to 42 sec and to 5. 2.64 1.2 rad/s.36 1. the time interval is set to 0.2 sec. In the beginning part of time duration. Finally. This test gives the critical damping coefficients in the still water. The firstorder and also the secondorder wave forces are calculated using the concept of a twoterm Volterra series model.97 level during four time steps. 51 wave components are combined to generate the time series wave data with random phases. Additional hull drag damping forces in the irregular state due to the current and waves are evaluated with reference to the paper produced by Wichers(1996). The frequency range for this combination is 0. the coupled analysis in the time domain is carried out in irregular waves.46 1.
1997 and Kim et al. 1991). and therefore. 1997). All the hydrodynamic coefficients were calculated in the frequency domain. Figure 5.24 to 1. The nearvertical riser is also hinged at the turret. The wave drift damping was expected to be small and thus not included in the ensuing analysis. firstorder wavefrequency forces. Taking advantage of symmetry.. The empirical coefficients for the viscous damping of the same FPSO hull in normal direction were obtained from the model test by Wichers(2000a). The frequencydependent radiation damping was included in the form of convolution integral to the time domain equation. ranging from 0.8 rad/sec and the intermediate values for other frequencies are interpolated. The methodology for hull/mooring/riser coupled statics/dynamics is similar to that of Ran and Kim.9 shows the distribution of panels on the body surface and free surface. 1999. and secondorder mean and differencefrequency forces are calculated from the secondorder diffraction/radiation program WAMIT (Lee et al.3686 × 10 8 kg at 62ft draft. The mooring lines are assumed hinged at the turret and anchor position. The hydrodynamic coefficients and wave forces are expected to vary . only half domain is discretized (1684 panels for hull and 480 panels for free surface). The calculated platform mass for the given condition is 2. The wave force quadratic transfer functions are computed for 9 wave frequencies. and then the corresponding forces were converted to the time domain using twoterm Volterra series expansion (Ran and Kim.6 Results and Discussion The added mass and radiation damping.98 5. riser tension is included in the vertical static equilibrium of the hull.
The surge staticoffset test shows a weakly softening trend. Typical results for surge offsets are shown in Figure 5. 1998) i.1 Static Offset Test (in Calm Water without Current) The surge static offset test was conducted by pulling the VCG (Vertical Center of Gravity) in the horizontal direction in calm water. The surge static offset . Therefore.6.e.99 appreciably with large yaw angles and the effects should be taken into consideration for the reliable prediction of FPSO global motions. In this paper. Therefore. The secondorder diffraction/radiation computation for a 3D body is computationally very intensive especially when it has to be run for various yaw angles. many researchers avoided such a complex procedure and have instead used simpler approach called Newman’s approximation(Faltinsen. The wind and current force coefficients on the vessel are read from OCIMF data. the offdiagonal components of the secondorder differencefrequency QTFs are approximated by their diagonal values (mean drift forces and moments). which is contrary to the typical hardening behavior of catenary lines. The dynamic wind loading was generated from the wind velocities obtained from the API wind spectrum. The yaw wind moments are increased by 15% considering the effects of superstructures. they are calculated in advance for various yaw angles with a 5degree interval and the data are then tabulated as inputs. 5.9. This approximation can be justified only when the relevant natural frequency is very small and the slope of QTFs near the diagonal is not large. the full QTFs are calculated and the validity of Newman’s approximation is tested against more accurate results with complete QTFs.
6E+07 1.0E+06 0.9 Static offset test results for surge motion . On the other hand. 1.0E+06 5.9.0E+06 0. The results are shown in Figure 5.0E+00 0 10 Full Load (w.0E+06 4.0E+07 8.0E+06 2.0E+06 4.2E+07 1.0E+06 Mooring line#2 tension [N] 7.0E+06 2.0E+06 6.100 curves with risers are in general greater than those without risers due to the contribution of riser tension. the effects of risers on individual mooring tension are less appreciable.0E+06 3.0E+06 1. risers) Full Load(w/o risers) Surge force [N] 1.4E+07 Full Load(w.0E+06 6. risers) Full Load (w/o risers) 20 30 40 50 60 70 80 90 100 Offset [m] (b) Static offset test results of #2 mooring line in the surge direction Figure 5.0E+00 0 10 20 30 40 50 60 70 80 90 100 Offset [m] (a) Static offset test results for surge motion 8.8E+07 1.
101
1.6E+06
Mooring line#8 tension [N]
1.4E+06 1.2E+06 1.0E+06 8.0E+05 6.0E+05 4.0E+05 2.0E+05 0.0E+00 0 10 20 30 40 50 60
Full Load (w. risers) Full Load (w/o risers)
70
80
90
100
Offset [m]
(c) Static offset test results of #8 mooring line in the surge direction Figure 5.9 Continued
5.6.2 Freedecay Tests (in Calm Water without Current) To see the effects of risers (mostly the amount of damping from risers) in the freedecay tests more clearly, a simpler riser model was developed i.e. all the 13 risers are replaced by a single equivalent massless riser at the center with the same total tension. The resulting surge/sway stiffness at the turret is then approximately calculated and added to the hydrostatic matrix. Figure 5.10 shows typical freedecay test results for surge, heave, roll, and pitch modes. The natural frequency and the damping coefficients obtained from the free decay test are summarized in Table 5.8 and Table 5.9.
102
100 80 60 40 20 0 20 40 60 80 100 120 0 200 400 600 800 1000 1200
Full Load (w. risers) Full Load (w/o risers)
Surge [m]
1400
1600
1800
Time [sec]
(a) Free decay test for surge motion
15 10
Heave [m]
Full Load (w. risers) Full Load (w/o risers)
5 0 5 10 0 20 40 60 80 100 Time [sec] 120 140 160 180 200
(b) Free decay test for heave motion
6 4
Roll [deg]
Full Load (w. risers) Full Load (w/o risers)
2 0 2 4 6 0 20 40 60 80 100 120 140 160 180 200
Time [sec]
(c) Free decay test for roll motion Figure 5.10 Freedecay test results for surge, heave and roll motions
103 Table 5.8 Natural periods from freedecay tests
Surge Full draft (with risers) Full draft (w/o risers) 209.8 s 225.9 s Heave 18.7 s 18.7 s Roll 13.0 s 13.4 s Pitch 18.6 s 18.6 s
Table 5.9 Damping from freedecay tests estimated from the first 4 peaks assuming linear damping
Surge Full draft (with risers) Full draft (w/o risers) 11.0 % (97.5 ~ 12.2 m) 5.8 % (96.7 ~32.7 m) Heave 6.5 % (10.9 ~3.2 m) 6.1 % (10.4 ~3.3 m) Roll 0.86 % (5 ~ 4.2 deg) 0.68 % (5 ~ 4.4 deg) Pitch 6.7 % (5 ~ 1.4 deg) 6.0 % (5 ~ 1.6 deg)
5.6.3 Timedomain Simulation for Hurricane Condition The current is assumed to be steady and the irregular wave unidirectional. A JONSWAP spectrum of significant wave height H s = 12.192 m, peak period T p =14s, and overshoot parameter γ =2.5 was selected to represent a typical 100yr storm in the Gulf of Mexico. The storm induced current flows from 30deg. right of wave direction. The current velocity is assumed to be 3.5ft/s between 0200ft and reduced to 0.3ft/s at 300ft3000ft. The wind speed used is 92mph@10m and its direction is 30deg. left of waves. The API wind spectrum is used for the generation of timevarying wind forces. The drag coefficients for wave forces are 1.0 for mooring lines, 1.0 to 1.414 for risers. The low and wavefrequency regions are defined as 00.2 rad/s and 0.21.3 rad/s, respectively. The timedomain simulation results are summarized in Table 5.10.
104 Table 5.10 Timedomain simulation results
Condition Newman’s Approx. (with risers) Newman’s Approx. (w/o risers) Full QTF (with risers) Newman’s Approx. (with risers) Newman’s Approx. (w/o risers) Full QTF (with risers) Newman’s Approx. (with risers) Newman’s Approx. (w/o risers) Full QTF (with risers) Newman’s Approx. (with risers) Newman’s Approx. (w/o risers) Full QTF (with risers) Newman’s Approx. (with risers) Newman’s Approx. (w/o risers) Full QTF (with risers) Yaw (deg.) Newman’s Approx. (with risers) Newman’s Approx. (w/o risers) Full QTF (with risers) Mean 13.9 13.9 14.7 4.7 4.6 4.8 0 0 0.1 0.2 0.2 0.1 0.0 0.0 0.0 15.3 13.7 15.1 Lowfreq. RMS 6.98 10.32 8.42 2.50 2.84 3.04 0.04 0.03 0.07 0.16 0.15 0.38 0.02 0.02 0.04 2.74 2.57 3.86 Wavefreq. RMS 0.49 0.44 0.44 0.49 0.45 0.46 3.36 3.46 3.37 0.98 1.26 1.22 1.33 1.39 1.34 0.28 0.31 0.28
(unit: m , deg.)
Total RMS 7.0 10.3 8.4 2.5 2.8 3.1 3.4 3.5 3.4 1.0 1.3 1.3 1.3 1.4 1.3 2.6 2.7 3.9 Max 34.6 46.7 39.5 13.4 13.8 16.9 10.9 12.1 11.1 3.5 4.3 5.5 4.3 4.7 4.5 22.7 22.3 24.3
Surge (m)
Sway (m)
Heave (m)
Roll (deg.)
Pitch (deg.)
From this result, it is clearly seen that slowly varying components are dominant in horizontalplane motions (surge, sway, yaw), while wavefrequency responses are more important in verticalplane motions (heave, roll, pitch). It is also found that the effect of riser damping is very important in the surge, particularly its slowly varying component. When riser damping is absent, the surge rms and maximum values are overestimated by about 47% and 35%, respectively. For the other modes, the effect of riser damping is less significant. If riser damping is not accounted for, the total rms
(with risers) Full QTF (with risers) Newman’s Approx. (w/o risers) Full QTF (with risers) Newman’s Approx. The simulation results for mooring lines and risers are summarized in Table 5. respectively. There also exist significant differences in rms and maximum tension of individual risers. (with risers) Newman’s Approx. (with risers) Full QTF (with risers) Newman’s Approx. (with risers) Full QTF (with risers) Newman’s Approx. which indirectly shows the importance of fully coupled analysis. (with risers) Newman’s Approx.11. (with risers) Full QTF (with risers) Mean 2160 2157 2201 903 943 901 2345 2343 1253 1254 4284 4383 2744 2746 960 961 Total RMS 424 583 479 249 349 296 272 262 278 265 403 391 234 227 166 166 Max 3529 4252 3639 1860 2319 2077 4941 5393 3509 3213 7629 6923 4082 4054 1804 1781 Mooring Line #2 Mooring Line #8 Liquid production riser #13 Gas production riser #20 Water injection riser #22 Gas injection riser #23 Gas export riser #25 .105 tension values on taut(#2) and slack(#8) mooring lines are overestimated by 38% and 40%. (with risers) Full QTF (with risers) Newman’s Approx.11 The results of tensions on the mooring lines and risers (unit: kN) Condition Newman’s Approx. Table 5. (w/o risers) Full QTF (with risers) Newman’s Approx.
slowly varying components are dominant. and therefore. the effects of riser coupling/damping and the validity of Newman’s approximation. The coupling effects are also minimal in verticalplane motions. we particularly addressed two points. In the present study. The error caused by massless riser modeling appears to be much more serious than that caused by Newman’s approximation in this example. only horizontalplane motions are appreciably affected. The use of Newman’s . particularly in surge.106 In Table 3 and 4. the horizontalplane motion amplitudes (slowly varying parts) are underestimated by using Newman’s approximation. but the differences are not large. As expected. This case is similar to the relevant study in DEEPSTAR Offshore Industry Consortium and the overall comparison looks reasonable. the comparison between Newman’s approximation and the full QTF is also shown. In horizontalplane motions.7 Summary and Conclusions The global motions of a turretmoored FPSO with 12 chainpolyesterchain mooring lines and 13 steel catenary risers in a nonparallel windwavecurrent environment are investigated in the time domain using a fully coupled hull/mooring/riser dynamic analysis program. wavefrequency responses are dominant and even the firstorder potentialbased theory can do a good job in heave and pitch. In general. 5. the reliable estimation of the secondorder mean and slowly varying wave forces and the magnitude of total system damping is very important. For verticalplane motions. The riser damping is found to be important in surge/sway modes.
and mooring lines can be directly seen on the screen through graphicsanimation software. In a fully coupled simulation in the time domain. when an input wave spectrum is not narrowbanded or doublepeaked. However. . risers. care should be taken. the behaviors of vessel.107 approximation slightly underestimates the actual horizontalplane motions but seems to be adequate in practical applications. which will greatly enhance the understanding of the relevant physics and the overallperformance assessment of the system.
the OCIMF data is used. the draft is changed to 15. The numerical model is made based on the experimental model conducted in the OTRC basin. This FPSO is also a tanker–based and turretmoored vessel. The principle data is the same as the FPSO introduced in the previous chapter. but the loading condition is different. wind and current force are used in the analysis as what the OTRC used in the experiment. The force coefficients for 80 % loading are interpolated automatically in the program using both data. which corresponds to 80 % loading of full load. . and the turret position is moved forward to the bow.108 CHAPTER VI CASE STUDY 2: DYNAMIC ANALYSIS OF A TANKER BASED FPSO COMPARED WITH THE OTRC EXPERIMENT 6. The force coefficients are taken for the full load and ballast loading. So. which is positioned at 38. The GoM environmental conditions for wave. The wave loads in the consideration of the different loading with the previous vessel are calculated by using WAMIT.734 meters aft of the forward perpendicular of the vessel. the tanker based FPSO designed for the water depth of 6.000 ft and tested in the OTRC basin is adopted for the verification of the WINPOSTFPSO program.121 m. For the wind and current forces.27 m along the ship’s center line. The x coordinate of the turret position is 116.1 Introduction In this study.
and 13 steel wire risers. The top tension of mooring lines is assumed to be the same as that of the original FPSO.2 OTRC Experimental Results and Design Premise Data Here the OTRC experimental results in the published paper in ISOPE 2001 will be used for comparison with the analysis results by WINPOSTFPSO. With the given draft. The details of the design premise data are shown in Table 6. Due to the change of draft for the different loading conditions. On the basis of this starting point.1. the attempt to find the model data and the experimental condition data is tried. There are 4 groups of mooring lines. The design premise data is basically the same as this in the previous chapter. the vessel is toward the East (the bow is heading the East). except for the draft and turret position.625 MT.1. The paper contains the experimental results of the static offset test. the draft is given as 15. the weight balance is checked. the principle data of vessel and mooring line are estimated by some hand calculations and rechecked by some numerical calculations. As shown in the above Figure. In the original design data there are 12 combined mooring lines with chain. The displacement can be expected to be 80 % of that of full load. Using this basic design data and the OTRC experimental results. The displacement can be evaluated with the different loading condition data and corresponding draft. The mooring lines and risers are spread from the turret. each of which is normal to other group. many design premise data should be changed. the free decay test and some time simulation. The general arrangement and body plan of the vessel are shown in Figure 6.109 6. In this loading condition. Each group . wire and chain.121 meters. so it will be 192.
L. 20 0 1 19 2 3 5 610 4 18 17 16 1115 Figure 6.000 ft .P.110 is composed of 3 mooring lines 5 degrees apart from each mooring line in the group. Station#0 Station#10 Station#20 A.P. F. Each mooring line has a studless chain anchor of Grade K4. C. The center of the first group is heading the true East. and so the second group is toward the true North.1 General arrangement and body plan of FPSO 6.
85 Lpp B H T m m m m MT m m 2 6.111 Table 6.878 0.85 m m m .440.6 12.869 6.52 15.32 5.57 3.0 47.17 28.73 1.1 Main particulars of the turret moored for the OTRC FPSO Description Production level Storage Vessel size Length between perpendicular Breadth Depth Draft (in full load) Diaplacement (in full load) Lengthbeam ratio Beamdraft ratio Block coefficient Center of buoyancy forward section 10 Water plane area Water plane coefficient Center of water plane area forward section 10 Center of gravity above keel Transverse metacentric height Longitudinal metacentric height Roll raius of gyration in air Pitch raius of gyration in air Yaw radius of gyration in air Frontal wind area Transverse wind area Turret in center line behind Fpp (12.121 240.12 0.04 15.9164 1.000 200 310.000 1.78 403.5 % Lpp) Turret elevation below tanker base Turret diameter L/B B/T Cb FB A Cw FA KG MGt MGl Rxx Ryy Rζζ Af Ab Xtur Ztur m m m m m m m m m 2 2 Symbol Unit bpd bbls kDWT Quantity 120.0 13.83 38.
MBL Segment 2: Polyester Length Diameter Weight in air Weight in water Stiffness.841 6.4 88.7 690. AE Mean breaking load.00 0.9 164. AE Mean breaking load.421 m mm kg/m kg/m kN kN 914. rope and wire for the OTRC FPSO Hydrodynamic Coefficients Normal drag Tangential drag Normal added inertia coefficient Tangential added inertia coefficient Coulomb friction over seabed Symbol Cdn Cdt Cin Cit F Chain 2.7 88.50 1.841 6.9 7.112 Table 6.515 m mm kg/m kg/m kN kN 1127.0 Rope/Poly 1.168 6.9 164.087.0 35.8 107. MBL Segment 3 (hangoff position): chain Length Diameter Weight in air Weight in water Stiffness.515 deg m m Unit kN Quantity 1.3 Hydrodynamic coefficients of the chain.4 794.15 0.65 2.9 143.3 1. AE Mean breaking load.0 Table 6.2 Main particulars of mooring systems for the OTRC FPSO Description Pretension Number of lines Degrees between 3 lines Length of mooring line Radius of location of chain stoppers on turn table Segment 1 (ground position): chain Length at anchor point Diameter Weight in air Weight in water Stiffness.9 42.2 0.201 4*3 5 2.9 143.4 794.2 0.6 . MBL m mm kg/m kg/m kN kN 45.45 0.
The equivalent mooring lines are spread 90 degrees apart from the adjacent mooring lines. only four equivalent mooring lines were used without risers.2. Table 6. the mooring lines are arranged symmetrically.113 However. in ORTC model. One equivalent mooring line is combined with 3 mooring lines. #2 equivalent mooring line is spread toward 135 degrees apart from the true East.3 gives the hydrodynamic coefficients for mooring lines. The schematic plot of the arrangement for mooring lines is shown in Figure 6. the equivalent mooring system is used.2 Arrangement of mooring lines for turretmoored FPSO . NORTH NORTH #6 #5 #4 Mooring Line #2 Mooring Line #1 45 0 #7 #8 #9 #3 #2 #1 EAST EAST Incident Wave Mooring Line #3 Mooring Line #4 #10#11 #12 (a) Mooring system of the original FPSO (b) Mooring system of the OTRC experiment Figure 6. #1 equivalent mooring line goes to 45 degrees apart from the true East. With respect to the x.2 shows the main particulars of equivalent mooring lines.and yaxis (the xaxis toward the East and the yaxis toward the North). Table 6. In the numerical model for this study. So.
The wave condition is composed of the significant wave height of 12 m.0914 1) 210 Remark: 1) The angle is measured counterclockwise from the xaxis (the East). From 60. the peak period of 14 sec. .5) 1) deg 180 m/s deg 41.44 m under the sea surface. the current speed is varied from 0. The wind direction is applied differently with the original FPSO case in Chapter V. and the overshooting parameter of 2. Hs Peak period. Table 6.9144 m/s. The velocity of current at the sea surface is 0.12 m/s.9144 0.09144 m/s. The mean wind velocity at the reference height of 10 m for one hour sustained is 41.19 14 JONSWAP ( γ =2.96 m under the sea surface.9144 m/s to 0. the 100year extreme hurricane condition at the Gulf of Mexico (GoM) is used as the same as in the previous case.96 m to 91. and it keeps until 60. which spectrum is shown in Figure 6.96 m at 91.5.114 6. Tp Wave spectrum Direction Wind Velocity Spectrum Direction Current Profile at free surface (0 m) at 60.0914 0.44 m on the sea bottom Direction Unit m sec Quantity 12. The wind spectrum of NPD formulae is taken as the design condition.3 Environmental Data For the loading condition for the analysis. The current is mainly induced by the storm.12 m/s @ 10m API RP 2AWSD 1) 150 m/s m/s m/s m/s deg 0.4 Environmental loading condition for the OTRC FPSO Description Wave Significant wave height.9144 0.3.
But.06 0.1 F (Hz) Figure 6. the loop current in the GoM should be considered as a design loading condition.44 m. in the OTRC . The current speed is uniformly kept 0.3 NPD wind spectrum curve For the intermediate region between 60.04 0.115 NPD Wind Spectrum. the current profile is determined by the linear interpolation. But. In the original data. the current speed at the free surface is 1. when the GoM environmental condition is applied to the platform design. The current speed and direction in the OTRC experiment were set up differently with the original FPSO case. and the direction is 150 o from the xaxis (true East). the current is assumed as one directional current.08 0. In this study.44 m under the surface to the sea bottom.96 m to 91.02 0. since the hurricane condition is severer than the loop current case.09144 m/s from 91. The summary of the environmental conditions for this study is shown in Table 6. S(F) 4000 3500 3000 2500 S(F) 2000 1500 1000 500 0 0 0.07 m/s.4. the loopcurrent condition will not be applied. While the storm wave and wind arise. however.
and the direction of 210 o .5. From the WAMIT output. Using the experimental model data and results. The obtained data from the WAMIT output is summarized in Table 6. The given data are L × B × D . T and the body plan are given. The numerical modeling for WAMIT is very similar to the FPSO model in the previous chapter except the draft. the turret position. KG . and the top tension of mooring lines as shown in Table 6. That’s the reason why the top tension is called the net buoyancy: Static equilibrium: B =W + T g (6. the hydrodynamic coefficients can be calculated by making the hydrodynamic modeling and by using WAMIT (the fluid interaction software to get the hydrodynamic coefficients). since the data of L × B × D . First.9144 m/s.116 experiment. T . the center of buoyancy and the restoring coefficients can be obtained. the required data should be newly estimated. the current speed was applied at the free surface of 0. The natural frequencies obtained from the free decay test in the OTRC experiment are known in a published paper (2001).1. Based on these data the weight of the model can be derived from the static equilibrium condition that the sum of the line top tensions and the weight is to be equal to the buoyancy. 6. the displacement volume.1) .4 Regeneration of the Experimental Model The design data are reestimated to match the experimental model condition.
ton kN m.5 WAMIT output and handcalculation Description Displaced volumn Buoyancy Total top tension Weight in mass Center of gravity Symbol Unit m3 m. respectively.ton m m m m Quantity 182. and so T and g mean the top tension of mooring lines and the gravitational constant.tonm 2 m.3251 4688. W denotes the weight of the body in mass unit. j = 1. Table 6.2.27 m.9566E+05 1.1018E+07 3.086 7.5189E+09 WAMIT WAMIT WAMIT Reference WAMIT ∀ × ρw ∀ B T Given data Static equilibrium Given data W xg zg Center of buoyancy xb zb Restoring coefficients C 33 C 44 C 55 M a 33 M a 44 Added mass/moment M a 55 The relations between the natural frequency.499 187. and the restoring coefficients and the masses are defined as follows: f = 1 2π Cij M V ij (1/sec or Hz) (i.801 89.L. T is the g mass tension or the net buoyancy.401 56.3226 22.ton m.tonm2 1.870 109.670 1.060 11.649 185.117 where B is the buoyancy.6) (6.2) .
ρ w and Aw are the water density and the water plane area.4) m55 = W ( R yy + z g + x g ) 2 2 2 (6. y g . These data are also summarized in Table 6.5) where ( x g .7) C55 = ρ w g ∫∫ x 2 n3 ds + ρ w g∀zb − mgz g = ρ w g∀M Gl . C 44 and C 55 are the nondimensionalized restoring coefficients. Aw C 33 = C33 2 ρ w gLR C44 4 ρ w gLR (6. From the WAMIT output. Cij is the restoring coefficients in which i and j can be any combination of six DOF. .6) C 44 = (6.3) 2 2 2 m44 = W ( Rxx + z g + y g ) (6. Aw C 55 = C55 4 ρ w gLR (6.118 where f is the natural frequency. M V ij can be obtained. The restoring coefficients are defined by: C33 = ρ w gAw . z g ) is the center of the gravity. and LR is the referenced length that is taken as the depth or the breadth of the vessel. m is the mass of the body to be the same as W . C44 = ρ w g ∫∫ y 2 n3 ds + ρ w g∀zb − mgz g = ρ w g∀M Gt . The relationship between mij and W are as follows: m33 = W (6. z b is the zcoordinate of the center of buoyancy.8) where C 33 . R yy are the radii of gyrations for roll and pitch motions. ∀ is the displaced volume.5. and Rxx . and M V ij (= M a ij + mij ) is the virtual mass in which M a ij is the added mass and mij is the mass of the body in the i and j direction.
The acquired data will be used as the analysis model data.3) to (6.0 R xx R yy R zz M Gt M Gl 6.400 11.2) and the experimental results in Table 6. Table 6. It is the process to clarify whether the data obtained from the above equations are acceptable for the numerical calculation on behalf of the experimental model.674 81. the data are verified. M Gt and M Gl denotes the transverse and longitudinal metacentric heights and n3 represents the directional cosine in zdirection.950 1349. if the data in Table 6. and are summarized in Table 6. Next. using the equation (6.6 and the equation (6.4. the radii of gyrations. They show the stiffness of the re .6 Reestimated data from WAMIT output and handcalculation Description Water plane area Radius of roll gyration Radius of pitch gyration Radius of yaw gyration Transverse metacentric height Longitudinal metacentric height Symbol Aw Unit m2 m m m m m Quantity 12.878 14. and metacentric heights can be derived.6. restoring coefficients. Therefore.119 Here.7.5 Results and Discussion 6.8) are taken advantage of.036 79.5.1 Static Offset Test with Regenerated Model Data The static offset tests are performed with the data obtained above by WINPOSTFPSO. The test results are depicted in Figure 6.
With the regenerated data. But. It results from the fact that the OTRC experiment started with the initial setting of the experimental instruments after a standing position in the calm water at a certain moment was set as the static equilibrium state. The reason to adjust the restoring coefficients for matching with the experimental is why the mooring line stiffness may contribute to the restoring forces of the system. Only a small difference is shown in the initial point.5. 6. . it is hard to say that moment is the same instant as the time when the model reached static equilibrium position.7. it is impossible and cannot be expected to get the same results once in the numerical calculation. the compatible results for the natural periods are obtained as in Table 6. It can make the difference in surge motion. very similar results were obtained.120 estimated model is well matched with that of the OTRC model. Fortunately. The line tensions at #1 mooring line and #3 mooring line show a slight difference from the experiments. After small modification of the restoring coefficients.2 Free Decay Test with Regenerated Model Data The proportional hull damping coefficients can be obtained from the free decay tests and the results are compared with the OTRC experiments.
0E+06 4.0E+06 2.0E+06 4.0E+05 6.0E+06 WINPOST(OTRC) MARIN(Experim ent) Mooring line tension [N] 7.0E+07 8.Surge Motion WINPOST(Full Load) MARIN(Experim ent) 1.0E+05 4.0E+06 3.0E+05 0.Mooring Line #1 WINPOST(Full Load) 1.2E+07 1.Mooring Line #2 WINPOST(Full Load) 8.0E+06 0.4 Comparison of the static offset test results .0E+06 6.0E+06 2.0E+05 2.121 Static Offset Curve of FPSO 6000 ft Polyester .2E+06 1.0E+06 1.0E+06 6.4E+07 WINPOST(OTRC) OTRC(Experim ent) Surge force [N] 1.6E+06 1.4E+06 WINPOST(OTRC) MARIN(Experiment) Mooring line tension [N] 1.0E+06 5.0E+00 0 10 20 30 40 50 60 70 80 90 100 Offset [m] (b) Static offset test result of #2 mooring line in the surge direction Static Offset Curve of FPSO 6000 ft Polyester .0E+06 8.6E+07 1.0E+06 0.0E+00 0 10 20 30 40 50 60 70 80 90 100 Offset [m ] (a) Static offset curves for surge motion obtained by experiments and WINPOSTFPSO Static Offset Curve of FPSO 6000 ft Polyester .0E+00 0 10 20 30 40 50 60 70 80 90 100 Offset [m] (c) Static offset test result of #1 mooring line in the surge direction Figure 6.8E+07 1.
0 13.7 10.8 5.5 182.8 10.48 0.5 193.122 2.9 5. moorings + 1 riser period(sec) damping(%) 4 equiv.8 6.23 #0 #2 #4 #18 1.1 0.4 16.9 5.8 10. mooring lines w/o riser period(sec) damping(%) surge (m) heave (m) roll (deg) pitch (deg) 206.8 8.2 13.9 6.7 Comparison of the free decay test results OTRC Experiment (4 equiv.64 1.6 10.0 0.72 0.7 10.40 1. Mooring lines) period(sec) damping(%) 12 mooring lines +13 risers period(sec) damping(%) WINPOST 4 equiv.9 4.36 1.9 12.7 12.19 #20 (b) Hull drag coefficients in consideration of the current effect Figure 6.5 3.9 4.38 0.5 8.5 .1 1.5 5.5 Hull drag coefficients proposed by Wichers (1998 & 2001) Table 6.32 1.4 12.5 10.1 8.4 13.00 #0 #2 #4 #18 #20 (a) Hull drag coefficients not in consideration of the current effect 2.46 1.13 Full Load Ballast 0.0 181.
When the drag coefficients considering the current effect are used. The frontal wind area is 20 % larger. But. the linear steel springs are used for the implementation of the steel wiring mooring lines in the experiments. and the lateral area is 30 % larger than that of the full load case. the spring has no static and dynamic mass. As is well known. It can be caused by taking the mooring line truncation in the experiment due to the depth limitation of the OTRC basin and the difference of the mooring lines between the experimental model and the real vessel.3 Time Simulation Results The comparison of the OTRC experiment and the WINPOSTFPSO analysis is shown in Table 6. Normally. the hull drag coefficients proposed by Wichers (1998. The first column in the table is the case to use the hull drag coefficients without considering the current. 2001) are used in this study as shown in Figure 6.8. The difference in the projected wind areas can results in the difference of statistically calculated values of motions. and the drag coefficients in surge are multiplied by 2.5. the analysis results have the trend to follow the experiment in sway and roll. In cases illustrated in the second and third column of the table. In the table.5.123 6. For the last test among four different cases. the frontal areas in surge and sway direction are used as the same as those in full load condition. in surge and yaw motion. there are still rather big differences between the experiment and the numerical simulation results. . the hull drag coefficients considering the current in sway and/or surge direction are used.5 for reviewing the drag force effect.
10 3.31 10.506 3.14 31.041 12.333 2. mean min.020 20. mean min. surge (m) 22.55 5.26 2.457 511 9.22 83.16 22.45 1. #3 (kN) max.73 0.38 4. rms.036 2.18 5.218 14.66 12.08 4.537 1. mean Mooring line min.50 0.29 1.84 3.64 0.40 3.16 0.803 2.38 3.18 2.02 1.67 3.893 2.60 3. sway (m) max.75 5.04 2.487 1.72 11.70 8. mean min.511 23.80 5.047 19.72 0. rms. The difference in the line tension as .883 2.870 20.643 827 13. rms.84 6.600 2.06 2.447 16.60 3.59 11.08 0. rms.5 hrs) New Sway and Surge Cd*2.403 1.333 204 7.566 14.230 14.44 0.657 1.543 1.600 2.046 2. They can make the difference in the surge and the yaw motions.25 20.23 1.29 9.38 5.927 8.49 4.53 16.931 2.96 6.57 0.96 0.562 782 7.67 5.94 21.163 9.67 23.05 4. heave (m) max.17 22.067 1.27 0.99 4.95 9.03 2.898 7.893 7.56 11.697 3.52 0.09 21.10 21.360 827 25.51 0.59 10.124 Table 6.5 hrs) New Sway Cd and Surge Cd (1.233 1. mean Mooring line min.01 4.907 3.735 2.91 9.90 7.91 3. pitch (deg) max.600 2.72 24.07 5.15 1.81 22.11 3. line model) Old Sway Cd (1.33 22. mean min.01 1. rms.8 Comparison of time simulation results OTRC Experiment Motion mean min.48 7.31 0.89 88.679 10.560 WINPOST (with 4equiv.11 6.127 801 In addition.853 2. rms.800 1.38 14.31 24. rms.53 14.28 1.26 83. yaw (deg) max. rms. Mooring Tension mean Mooring line min.32 0. #4 (kN) max.47 3.537 1. rms.58 3. mean min.09 1.440 1.68 0.05 2.35 0.359 3.565 2.590 1.35 0.796 7.42 0.688 2.58 5.757 2.79 2.379 202 7.5+old wind area (3 hrs) 5.346 8.18 6.644 630 7.39 5.41 8.5 hrs) New Sway Cd (1.76 8.46 0.61 7. #1 (kN) max.20 3.173 2.95 6.400 197 7.49 18.02 24. rms.31 16.597 802 13.39 78.21 21.02 2.99 8.15 21. it has no lateral stiffness.90 0.13 4.92 61.28 1.48 0. mean Mooring line min. #2 (kN) max.14 11.54 11.26 1.89 3.00 24.43 13. so it can react only in line.540 1.593 530 7.783 2.91 3. roll (deg) max. max.64 18.
the current profile control. There are still many uncertainties for the reasons for the differences between the experiment and the numerical analysis results. some analyses are conducted with the WINPOST program. Some reasons for these differences can be imagined. the surge drag force is newly considered (Cd=1. The new sway hull drag coefficients are used as shown in Table 6. With the numerical model to be matched to the experimental model.0). but those in surge and yaw motion show no good agreement. the trends in sway and roll motion may well follow the experimental results.5. Newman’s approximation scheme is used for evaluating the wave forces applied to the single body model and also to the twobody model. But. For the consistency. 6. some efforts are exerted to regenerate the experimental results by the OTRC. For example. the yaw and surge motion still has a little large difference compared to the experiment. The analysis results are rather close to the experiments in viewpoint of overall trend. the mooring line truncation and the usage of springs for the steel wiring mooring lines. the experimental static offset curve and the free decay test results are used.6 Summary and Conclusions In this study. . To find the model parameters.3 (b) and (c) may be the reason for the discrepancy. Furthermore. the investigation of the wind and current generated in the basin might give some clues.125 shown in the static offset tests in Table 6. When the hull drag coefficients are applied in consideration of the current effect. such as the wind force generation.
126 CHAPTER VII CASE STUDY 3: CALCULATION OF HYDRODYNAMIC COEFFICIENTS FOR TWO BODY SYSTEM OF FPSO AND SHUTTLE TANKER 7. the hydrodynamic coefficients for the twobody system are performed and compared with the experimental results of other institutes (KRISO. In this study. 2002). It has resulted in conservative estimates for the behaviors of two bodies. The multiple body system is composed of an LNG FPSO and a shuttle tanker.1 Introduction In this study. developed by MIT using Boundary Element Method) program and experiments. the interaction characteristics for the tandem and sidebyside moored vessels are investigated and compared with the experiments carried out for a twobody tanker model with different arrangements in regular waves. In many cases of the conventional tandem mooring of the FPSO and shuttle tanker. Motions and drift forces are mainly reviewed with the numerical calculations by the WAMIT (Wave Analysis program. the hydrodynamic interaction between the two bodies has been ignored since the interaction is not considered large enough to be taken account of. The changes of the distances between two vessels and the mooring types are used as parameters for investigation of the interaction characteristics. This program has the module to solve the interaction problem based on the multiple body interaction theory. .
the tandem mooring is taken into account since this type of mooring system has been used for the offloading operation in the way that the shuttle tanker is located behind FPSO. In such a case. Dijk and Wilde(2001) developed the numerical time simulation solver to predict the hydrodynamic response of alongside moored vessels. as the conventional mooring pattern. For the same topic. parametric studies of the interaction effects on the drift forces and vessel behaviors are being performed in this study. Inuoue and Islam(2001) investigated the roll motion effect on wave drift force for the sidebyside moored vessels. Buchner. As another mooring system. an LNG FPSO and a shuttle tanker are taken. Here. Pinkster and Wilde(2001) tried to obtain the numerical approach to solve the diffraction and the radiation potential problem for a very close multibody system. the distance between the two is very close. For the test models. .127 There are several research works on this matter. the distances are kept between 1 4 to 1 3 of the ship’s length. sidebyside mooring is being considered since the offloading operations are sometimes preferred under the parallel position in relatively calm seas. and so the hydrodynamic interaction and mooring design are very important. On the situation. Huijsmans. Garrison (2000) developed the numerical tool for the timedomain analysis of the hydrodynamic loads and motions for a very large multibody floating structure(VLFS) using the panel method based on the timedependent Green’s function. For two types of moorings and two different distances between the LNG FPSO and the shuttle tanker.
which is made for a sensitivity study. the roll natural period of the LNG FPSO is 15.8 for LNG FPSO. the natural frequencies are compared with each other.1 sec for shuttle tanker. In Figure 7. the distances for sidebyside mooring are determined as 4 m and 10 m.2.1. The mooring lines modeled as springs are posted at the posts located at the end of the mooring lines. The arrangements of tandem and sidebyside mooring are shown in Figure 7. Table 7.e. the finemeshed numerical models are shown. the number of elements was not very sensitive to the results. it was proved that the model size.2 shows the free decay test results. The free decay tests are conducted with the numerical models. The main particulars. It has 4 times number of elements of the roughmeshed model. On the other hand. and that of the shuttle tanker 9. the springs are not considered since the stiffness of the spring is too small and so their hydrodynamic effects can be negligible.97 sec. Steel springs for the mooring systems are used. of which the FPSO is fully loaded and the other is ballast loaded. and according to the test results. In Figure 7. the roll natural period of 15.1. the numerical models are shown.2 Particulars of Models and Arrangements for the Tests Both models are tanker type vessels. For the calculation of the hydrodynamic coefficients. are listed in Table 7.128 7. For the validity of the numerical modeling for the two vessels.. In . i. including the principle data of the vessels. and of 10. The test reveals that the numerical model is good enough to use for the numerical calculation. The distances between the two vessels in tandem mooring are taken as 30 m and 50 m.3. Consequently.7 sec. and the stiffness of the springs is set to 320 kN/m. According to the experiment by KRISO (2002).
888 14.3.65 8.129 Table 7. Table 7.8 7.743.5 53.7 55.75 59.028 16.577 12.54 6. the comparison of the hydrodynamic coefficients obtained from the rough model and the fine model is shown.75 55.75 ∀ LCG KG GM Kxx Kyy Kzz .82 15.20 8.152 9.82 139.585 9.1 Main particulars of two vessels Description Length b/w perpendiculars Bredth Draft at FP Draft at midship Draft at AP Displacement Longitudinal center of gravity Vertical center of gravity Metacentric height Radius of roll gyration Radius of pitch gyration Radius of yaw gyration Symbol Lpp B TFP TMID TAP Unit m m m m m m3 m m m m m m LNG FPSO 239 45.82 15.75 Shuttle Tanker 223 42 6.04 59.82 15.636 14.
152 0.2 50.786 1.14% 0.777 1.459 0.51 7.10% LNG FPSO 4 3 Shuttle Tanker Heave Motion [m] 2 1 0 1 2 3 4 0 20 40 60 80 100 120 Time [sec] LNG FPSO Time(s) Period(s) Roll(deg) ln(x1/x2) Damp.04% 0.0 10.13% 0.2 40.258 0.42% 7.08% 0.802 0.01 0.4 10.4 Average 10.0 3.8 15.55 0.183 0.0 10.6 11.657 ln(x1/x2) Damp.747 1.8 0.05% 0.717 1.4 11.19 0.6 11.0 12.0 10.13% 3.18 0.19 0.01 0. Time(s) Period(s) SHUTTLE TANKER Roll(deg) ln(x1/x2) Damp.4 10.4 10.13% 0.6 73.434 1.8 31.0 23. Time(s) 0.4 10.2 92.703 1.0 94.01 0.2 10.06% 3.608 2.14% 0.4 Average 0.36 0.11% 0.8 81.0 10.50 0.784 1.54% 8.0 2.0 10.8 11.91% 9. Ratio 1st 3 Ave.0 58.8 110.689 1.0 15.0 10.130 Table 7.8 15.50 0.0 15.0 Shuttle Tanker Roll Motion [deg] 1.14% 0.808 1.732 1.20 0.01 0.0 0 10 20 30 40 50 60 70 80 90 100 Time [sec] . Ratio 1st 3 Ave.68% 8.13% 0.739 1.042 ln(x1/x2) Damp.07% 8.01 0.80% 5.4 47.74 0.44 0.399 1.756 1. 0.2 Heave(m) 2.46 0.0 83.6 35.01 0.4 SHUTTLE TANKER Period(s) Heave(m) 3.13% 0.2 41.28% 3.0 10.779 1.02% 0.76% 31.07% LNG FPSO 3.8 15.8 1.808 1.4 20.01 0.2 60.42 0.8 15.4 80.09% 0.2 10.6 10.83% 9.00 0.798 1.0 10.43 1.14% 0.8 52.00 0.4 Average Period(s) 0.55 0.8 16.53 0.08% 0.0 10.84% 3.14% 0.06 0. Ratio 1st 3 Ave.01 0.4 10.20 0.1 1.01 0.976 0.118 2.801 0.0 0.765 1.2 62.2 Freedecay test results for a LNG FPSO and a shuttle tanker (heave and roll) LNG FPSO Time(s) 0.2 70.20 0. 11.0 10.4 11.13% 0.58 0.0 2.8 15.13% 3.01 0.00 0.0 11.4 11.21 0.0 20.77 0.4 69.43 0.8 15.122 1.762 1.06% 3.6 47.092 0.17% 3.01 0.0 12.6 Average 0.00 0.738 1.19 0.07% 3.0 15.60 0.0 30.19 2. Ratio 1st 3 Ave.0 1.75 1.17% 3.13% 0.01 0.01 0.01 0.792 1.4 10.6 126.11% 0.20% 8.4 63.99% 6.2 79.
2% 4.4176E+01 1.2937E+09 1.6% 2.3147E+05 2.3208E+06 2.3179E+05 2.4537E+07 4.8971E+04 1.1139E+05 9.0976E+03 8.1 Configuration of the mooring system .0029E+05 9.8210E+08 4.6910E+00 4.2265E+11 5.9999E+05 1.3145E+06 2.2282E+11 5.9963E+05 1.8058E+00 2.8218E+08 4.8793E+02 5. difference in radiation dampings LNG FPSO Extended Simple model model 2.0734E+06 1.2623E+08 1.4776E+06 1.9093E+04 1.1563E+02 1.8811E+02 5.9242E+06 3.9999E+06 1.1766E+02 1.1462E+03 9.7447E+01 4.4782E+06 1.9253E+06 3.4379E+01 1.4468E+07 4.7570E+07 1.0% 1.2205E+10 2.2727E+01 8. difference in added masses Fd11 Fd22 Radiation damping Fd33 Fd44 Fd55 Fd66 Max.1447E+06 9.4853E+07 1.131 Table 7.3% SHUTTLE TANKER SHUTTLE TANKER LNG FPSO LNG FPSO (a) Tandem arrangement (b) Sidebyside arrangement Figure 7.1748E+06 9.7754E+07 1.2794E+09 1.2637E+08 1.4606E+07 7.7347E+01 4.3 Comparison of the hydrodynamic coefficients obtained from the rough model and the fine models Hydrodynamic coefficients Symbol Ma11 Ma22 Added mass Ma33 Ma44 Ma55 Ma66 Max.2050E+10 Shuttle tanker Extended Simple model model 7.2678E+01 8.
3 Finemeshed numerical modeling for a LNG FPSO and a shuttle tanker 7. Only head sea conditions are considered for the tandem moored case. for the sidebyside moored vessels. both beam sea and head sea conditions are .2 Roughmeshed numerical modeling for a LNG FPSO and a shuttle tanker (a) the sidebyside mooring (b) the tandem mooring Figure 7.132 (a) the sidebyside mooring arrangement (b) the tandem mooring arrangement Figure 7.3 Environmental Conditions Regular waves are taken for the calculation of the beam sea and head sea conditions. On the contrary.
6 and 7. As is well known. those for the tandem mooring are selected as 30 m and 50 m.5. The distances for the sidebyside mooring are taken as 4 and 10 meter as the parameters.133 considered.4 to 7. . For more clear comparison.2 rad/s with 50 intermediate intervals.8 for the head sea condition. The range of the wave frequencies is from 0. the longitudinal drift forces are compared as shown in Figures 7. the effects are large enough to pay attention to the matter for solving the interaction problem more accurately.9 and 7.5 for heave and roll motions in beam sea state.4 and 7.10 in different heading condition. The distance effect on the longitudinal drift force is shown in Figure 7. The whole trends show good agreement to the experiments. 7.4 rad/s to 1. the calculated RAOs and drift forces for a single body of the FPSO and a single body of the shuttle tanker in the same condition are depicted in the above figures. They are very clear over the whole frequency range. For the different mooring systems. and on the contrary.4 Results and Discussion The analysis results and the experiments can now be compared. Motion RAOs as varying the distance apart from each other for the sidebyside mooring are compared as shown in Figures 7. The shielding effects on heave and roll motion RAO are well investigated in the lee side vessel of the sidebyside mooring vessels as shown in Figures 7. The drift forces in the lateral direction for the sidebyside moored vessels are shown in Figures 7.7 for the head sea condition.
6 Frequency (rad/s) Heave RAO for a sidebyside mooring.0 1.6 0.0 0.0 0.5 0.4 1.2 0.8 1.0 0.) Heave RAO (Z/A) 1.6 Frequency (rad/s) Figure 7.5 1. Distance=10m 2.4 Heave response operators of sidebyside moored vessels in the beam Sea .0 1.4 0.4 1.8 0.5 FPSOTw o Body ShuttleTw o Body FPSOSingle Body ShuttleSingle Body FPSOTw o Body (Exp.0 0.) 2. Head=90 deg.0 1.2 0.4 FPSOTw o Body ShuttleTw o Body FPSOSingle Body ShuttleSingle Body FPSOTw o Body (Exp.2 1.2 0.) ShuttleTw o Body (Exp.2 1.4 0.8 1. Head=90 deg.0 0.4 0.2 1.6 0.8 1.0 Heave RAO (Z/A) 1.6 1.) ShuttleTw o Body (Exp. Distance=4 m 2.134 Heave RAO for a sidebyside mooring.0 0.6 0.
Head=90 deg.0 0. Head=90 deg.4 0.0 1.) Roll RAO (phi/kA) 8 6 4 2 0 0.4 1.5 Roll response operators of sidebyside moored vessels in the beam sea . Distance=4 m 14 12 10 FPSOTw o Body ShuttleTw o Body FPSOSingle Body ShuttleSingle Body FPSOTw o Body (Exp.) Roll RAO (phi/kA) 8 6 4 2 0 0.2 1.6 0.0 1.2 1.0 0. Distance=10m 14 12 10 FPSOTw o Body ShuttleTw o Body FPSOSingle Body ShuttleSingle Body FPSOTw o Body (Exp.) ShuttleTw o Body (Exp.8 1.8 1.6 0.135 Roll RAO for a sidebyside mooring.6 Frequency (rad/s) Roll RAO for a sidebyside mooring.2 0.4 1.4 0.2 0.) ShuttleTw o Body (Exp.6 Frequency (rad/s) Figure 7.
2 0.) XDIR.6 Frequency (rad/s) Figure 7.4 1.6 Frequency (rad/s) Drift force: Tandem mooring.8 1.136 Drift force: Tandem mooring.8 1.) 40 XDIR. Head=180 deg.0 1.4 1. Head=180 deg.2 1.6 Longitudinal wave drift force of tandem moored vessels in the head sea . Drift Force (kN/m2) 0 40 80 120 160 0.0 1.0 0.6 0. Distance=50m 80 FPSOTwo Body ShuttleTwo Body FPSOSingle Body 40 ShuttleSingle Body FPSOTwo Body (Exp.4 0.6 0.) ShuttleTwo Body (Exp.) ShuttleTwo Body (Exp.2 0.4 0.0 0. Drift Force (kN/m2) 0 40 80 120 160 0.2 1. Distance=30m 80 FPSOTwo Body ShuttleTwo Body FPSOSingle Body ShuttleSingle Body FPSOTwo Body (Exp.
0 0.0 0.2 0.4 1. Head=180 deg.2 0. Drift Force (kN/m2) 80 160 FPSOTw o Body ShuttleTw o Body 240 FPSOSingle Body ShuttleSingle Body Exp(FPSO)KRISO EXP(Shuttle)KRISO 320 0.) 0. Distance=4m 80 0 XDIR.4 0.) ShuttleTw o Body (Exp. Drift Force (kN/m2) 80 160 FPSOTw o Body ShuttleTw o Body FPSOSingle Body ShuttleSingle Body FPSOTw o Body (Exp.4 0.6 0.2 1.8 1.2 1.7 Longitudinal wave drift force of sidebyside moored vessels in the head sea . Head=180 deg. Distance=10m 80 0 XDIR.0 1.6 Frequency (rad/s) Figure 7.6 240 320 Frequency (rad/s) Drift force: SideBySide.137 Drift force: SideBySide.0 1.6 0.8 1.4 1.
SideBySide mooring.4 1. Drift Force (kN/m2) 40 80 120 Two Body (50m) 160 Two Body (30m) Single Body Two Body (Exp.) (4m) 240 0.0 0.2 0.2 1.4 0.6 Frequency (rad/s) Figure 7. Drift Force (kN/m2) 40 80 120 160 Two Body (10m) Two Body (4m) 200 Single Body Two Body (Exp.6 0. Head=180 deg 40 0 XDIR. Tandem mooring.6 Frequency (rad/s) Longitudinal Drift force: FPSO.0 1.) (50m) 200 0.0 Two Body (Exp.6 0. Head=180 deg 40 0 XDIR.) (10m) Two Body (Exp.4 0.138 Longitudinal Drift force: FPSO.8 1.) (30m) 1.2 1.0 0.8 1.4 1.8 The distance effect on the longitudinal wave drift force for a twobody and a single body model in the head sea .2 0.
) ShuttleTw o Body (Exp.2 0.) YDIR. Head=180 deg.6 FPSOTw o Body ShuttleTw o Body FPSOSingle Body ShuttleSingle Body FPSOTw o Body (Exp.4 0. Head=180 deg.2 1. Distance=10m 1200 1000 800 FPSOTw o Body ShuttleTw o Body FPSOSingle Body ShuttleSingle Body FPSOTw o Body (exp.) ShuttleTw o Body (Exp.4 0.0 0.4 1.139 Drift force: SideBySide.9 Lateral wave drift force of sidebyside moored vessels in the head sea .0 1.6 0.0 1.4 1.6 0.8 1.6 Frequency (rad/s) Drift force: SideBySide. Distance=4m 2000 1500 1000 YDIR.) Frequency (rad/s) Figure 7.8 1. Drift Force (kN/m2) 600 400 200 0 200 400 600 800 0.2 1. Drift Force (kN/m2) 500 0 500 1000 1500 2000 2500 3000 0.2 0.0 0.
140 Drift force: SideBySide. Drift Force (kN/m2) Frequency (rad/s) Figure 7.4 1.6 0. Distance=4m 8000 6000 4000 2000 0 2000 4000 6000 8000 0.6 Frequency (rad/s) Drift force: SideBySide.8 1.0 1.4 0.) YDIR.) ShuttleTw o Body (Exp. Drift Force (kN/m2) 500 0 500 1000 1500 2000 0.) ShuttleTw o Body (Exp. Distance=10m 1500 FPSOTw o Body ShuttleTw o Body FPSOSingle Body ShuttleSingle Body FPSOTw o Body (Exp.0 0.0 0.0 1.2 1.6 FPSOTw o Body ShuttleTw o Body FPSOSingle Body ShuttleSingle Body FPSOTw o Body (Exp.10 Lateral wave drift force of sidebyside moored vessels in the beam sea . Head=90 deg.8 1.6 0.2 0. Head=90 deg.2 0.) 1000 YDIR.4 0.2 1.4 1.
the shielding effect is noticeable on the drift force.141 As shown in Figures 7. and are also remarkable in the tandem moored vessel. In sidebyside mooring. the shielding effect of the lee side vessel is significant on the drift force and motion RAO.6 and 7. so that the conclusion is drawn that the program can be applied to that kind of interaction problem. The distance effect on the drift force is not significant. but are not clear in the sidebyside moored vessels. 7. The lateral drift force of sidebyside moored vessels in head sea and in beam sea are quite different.7.9 and 7. the magnitude of the lateral drift seems to be reciprocally amplified against the distance. The distance has no great effect on the longitudinal force. The distance effects on the motions and drift forces of the two vessels are also reviewed.10.5 Summary and Conclusions The hydrodynamic interaction effects for the multibody system are investigated by a comparative study for the numerical calculations and experiments. as shown in Figure 7. and the sidebyside and tandem mooring are considered. the lee side ship acts as a block to disturb the flow pattern of the wave. As the distance gets closer. Furthermore. the shielding effects on the longitudinal drift forces for the head sea conditions are investigated. the WAMIT gives the fairly reasonable results. the blockage effect on the lateral drift force increases. . It causes the force to be magnified as the lee side vessel approaches the weather side vessel. With comparing the experiment. The LNG FPSO and a shuttle tanker are taken as the multibody system. when the distance between both vessels gets closer. In lateral. In tandem mooring.
and yaxis. the dynamic coupled analysis for twobody structures is performed to verify the program (WINPOSTMULT) for the dynamic coupled analysis of the multiplebody floating platforms and the results are compared with the analysis results using the idealized model of a twomassspring model. The simplified massspring model will give a compatible result to judge the validity of the multiplebody program. The WAMIT program has the module to solve the fluid interaction problem based on multiple body interaction theory. the body motions and line tensions are mainly reviewed with the numerical calculations performed by WINPOSTMULT. the dynamic coupled analysis program for multiplebody platforms. since the structure is symmetric about the x. the sidebyside mooring and the tandem mooring have no difference. The twobody interaction problem of the fluid was studied in the previous chapter. The hydrodynamic coefficients in consideration of the multiplebody interaction are calculated by the WAMIT. The conventional tandem moorings have been taken for the multiplebody connection in many cases. For the multiplebody model of spar structures. as explained before. In this study. The analysis results by the program are compared with the analysis .1 Introduction In this study. The multiple body system is composed of two identity spars.142 CHAPTER VIII CASE STUDY 4: DYNAMIC COUPLED ANALYSIS FOR A TWOBODY SYSTEM COMPOSED OF SPAR AND SPAR 8.
The distance between the two spars in tandem mooring is taken as 30 m.1. It can be said that the sidebyside mooring should be identical to the tandem mooring due to the symmetry of the structure. the distance is determined to be 30 meter to allow the maximum surge or sway motion.143 results of the twobody spar model connected by a hawser with and without the hydrodynamic interaction effect. Thus. and also compared with the results by the linear spring model replaced for the hawser.2 Particulars of Models and Arrangements for the Analyses The main particulars including the principle data of spar are listed in Table 8. For the mooring system. . since the expected maximum surge motion is about 30 meters and the maximum sway motion about 10 meters.1. For the validity of the numerical modeling. 8. for the linear spring modeling. From this study. For this verification. For the calculation of the hydrodynamic coefficients. the effect of the hawser to connect the two structures can also be clarified. the tandem mooring is taken into account since this type of mooring system has been used for many years for offloading operations to transfer the oil from one platform to other structures. The mooring lines are fixed at the sea floor. the program is modified slightly. the WAMIT program is. Static offset test and free decay tests are performed and compared with the target values. The distance is kept as close as possible. which are given from experiments conducted by other institute. the models with a hawser and without a hawser are made and analyzed. Especially. The arrangement of the tandem is shown in Figure 8.
69 1.=37.1856 m d=30 m Wind Hawser SPAR #2 SPAR #1 Wave Current Figure 8.68 x 17.88 198.84 95.06 17.12 67.000 72 214.07 Dia.59 129.600 220.68 164.740 67.36 8.15 2 T H m m m m m m mT mT m m N/(m/s) m KG Rxx Rζζ Cd Cw 2671.6 220.4 55.144 Table 8.1 Main particulars of moored spar Description Water depth Production level of oil Production level of gas Length Draft Hard tank depth Well bay dimension (25 slots) Center of buoyancy center above base line Center of gravity above base line KG (based on total displacement) Displacement Total displacement Pitch radius of gyration in air Yaw radius of gyration in air Drag force coefficient Wind force coefficient Center of pressure above base line KB KG ∀ ∀  Symbol Unit m bpd mmscfd m Quantity 914.1 Configuration of the mooring system and the environmental loads (Tandem arrangement.71 53. d=30m) .
2 Particulars of the mooring systems Description Pretension Number of lines Scope ratio Length of mooring line Firlead location above base line Segment 1 (ground position): chain Length at anchor point Diameter Weight in air Weight in water Stiffness.44 .3.18E+05 1. MBL Added mass Current force coefficient Segment 2: wire Length Diameter Weight in air Weight in water Stiffness. Table 8. MBL Added mass Current force coefficient Segment 3 (hangoff position): chain Length m Other parameters are the same as those of segment 1.08 91.18E+04 37.2. the numerical models are shown.145 In Figures 8. AE Mean breaking load. the particulars of the mooring systems are tabulated.41 1.0 36.77 3.5 287. AE Minimum breaking load.357 14 1.8 250.8 1. In Table 8.2 and 8.45 m m Unit kN Quantity 2. 91.44 m mm kg/m kg/m kN kN kg/m 2347.92 24.3 1.20 m mm kg/m kg/m kN kN kg/m 121.402.28E+04 28.4 2.44 21.03E+06 1.52 7.
and it is kept 60. .5 rad/s to 1.12 m/s @ 10m API RP 2AWSD 1) deg 210 m/s m/s m/s m/s m/s deg 1. The current velocity is 1.0914 0.146 8.0668 m/s at the free surface. Hs Peak period. and the overshooting parameter of 2.3 Environmental conditions Description Wave Significant wave height.2 rad/s with 50 intermediate intervals. In Table 8.192 meters.3 Environmental Conditions The environmental conditions to be used in this analysis correspond to the 100year storm conditions in Gulf of Mexico.0668 1. The wind velocity is 41.19 sec 14 JONSWAP ( γ =2.3. After that. The range of the wave frequencies is from 0.0914 m/s. the peak period of 14 seconds.12 m/s at 10 m of reference height for 1 minute sustained. the current speed becomes uniform as 0. The wave spectrum used here is the JONSWAP spectrum.44 m on the sea bottom Direction Unit Quantity m 12. Under the water depth of 91.0914 m/s from 60.96 m under the water surface. Tp Wave spectrum Direction Wind Velocity Spectrum Direction Current Profile at free surface (0 m) at 60. API RP2T is used.44 m. which has the significant wave height of 12. the environmental conditions are summarized.3. Table 8. as shown in Figure 8.0914 1) 150 Remarks: 1) The angle is measured from xaxis (the East) in the counterclockwise. it varies from 1.96 m to 91.0668 0. For wind force calculation.5.5) 1) deg 180 41.44 m under the water surface.0668 m/s to 0. For wave. irregular waves are taken for the calculation of the head sea condition.96 m at 91.
so that for both analyses Newman’s Approximation Method is adopted for conforming the full QFT when the wave force coefficients are considered.147 8. the 1st order model is used for the single body analysis and also for the twobody analysis. The hydrodynamic interaction effect is calculated with the 1st order model.3. All coupling terms are considered for the twobody analysis.3.and differencefrequency quadratic transfer function (QTF) of diffraction potential force are calculated by the WAMIT 1st order module and the 2nd order module. the twobody model for the 1st order wave force coefficients is shown. The hydrodynamic coefficients of added mass. The body has 1024 elements. The program WINPOSTMULT can treat the numerical calculation with the fully coupled system matrices composed by multiple bodies.2 and 8. wave damping. if the hawser or the fluid transfer lines . the system stiffness and the resorting coefficient matrix are fully coupled with each other due to the interaction effects of both structures. and the free surface has 576 panel elements. the 1st order wave force coefficients and wave drift force coefficients are calculated.2. The hawser connecting each spars to the other is taken to have 1/100 of the mooring stiffness and 1/10 of the mooring pretension. Especially. Here. For the twobody analysis. the numerical models are shown. The added mass. linear transfer function (LTF) of diffraction potential force and the sum. In Figure 8. the 2nd order wave force coefficients are calculated with free surface modeling. the linear wave damping. The hydrodynamic coefficients are calculated by WAMIT. For the single body analysis. In Figure 8. the model for the 2nd order wave force coefficients is shown. for the purpose of comparison.4 Calculation of Hydrodynamic Coefficients Using WAMIT 1st and 2nd Order In Figures 8.
Figure 8. the coupling terms of the hydrodynamic coefficients are set as zero. the analysis of the twobody system is performed using the 1st order model with and without interaction effects.148 are connected.2 Configuration of the modeling of a single spar Figure 8. In the case of no interaction effects.3 Configuration of the modeling of a twobody spar . As mentioned above. they will cause to make the stiffness matrix coupled so that the whole system stiffness matrix composed by the body and line stiffness and restoring coefficients comes to a huge sparse matrix.
Furthermore. the effect makes the sway motion of the lee side structure amplified a little. In the table. At every time step. The corresponding case to this is the sparhawserspar model with no interaction effect.6 Results and Discussion The analysis results using the twobody spar model with a hawser connection and a linear spring model between two spars are compared with the results of a single spar as shown in Table 8. It means that the weather side structure acts as a protector for the leeside structure.149 8. The results of the interaction case and the nointeraction case with no cable reveal that the fluid interaction effect makes the rear side structure move a little less in all directional motion except the sway motion. 8. For verifying the numerical analysis results by the full numerical model. the distance between both spars is checked in the modified program. . the WINPOSTMULT program is modified slightly since the replaced spring can work only when two bodies move in the opposite direction against each other out of phase. However. and then the spring works only when spars are moving over 30 m in surge direction. the sparspringspar model is considered an ideal case so that the responses of both spars are identical.4.5 Linear Spring Modeling The hawser for connecting the two spars can be replaced by a linear spring. a linear spring for the hawser is considered by putting the linear spring constant as a restoring coefficient in surge direction into the body system matrix of the restoring force coefficients inside the program. These models show a good agreement to each other.
wind and current. It can make the differences between the analysis results for the singlebody model and the twobody model in surge.c.4. To get some clues for the reason of the sudden increases in surge and yaw motion RMS in the case of interaction effect with one hawser.5. the surge motion RAO is illustrated in Figure 8.b and 8. it also forces the second body to move in a more restricted way and less than the first body in the front side of the wave.4. As shown in Figure 8. The heave motion RAO and the roll motion RAO are shown in Figures 8.b and 8.4.150 When one hawser is used for the connection. In Figure 8. the twobody interaction effect can be seen. In the figure. The cable can be imagined to limit the motion of the second body. As shown in Figures 8.c. the surge drift force for the twobody model has twice large than that for a single body model. the heave and roll motion RAOs for the twobody model have similar trends to those for the singlebody model. since the hawser has the rigidity in the surge direction and so it will go to the opposite direction against the second body movement when they are in an outofphase state. But.a. the surge motion RAO for the twobody model has a similar trend to that for the singlebody model.a. The magnitude of the compensating reaction will vary according to the stiffness of the hawser. the surge mean drift forces for a single body and those for twobody by the pressure integration method are shown for comparison purpose.4. heave and roll dynamic motions. .4.4.
2 1.1 0.6 Frequency (rad/s) Figure 8.5 0.a Comparison of the surge motion RAOs Heave Motion RAOs 1.2 0.0 1.4 0.2 0.8 1.4 0.9 0.0 0.2 0.6 0.b Comparison of the heave motion RAOs .4 1.0 0.0 1.0 0.6 Freqiency (rad/s) Figure 8.4 0.4.4.3 0.2 1.8 Heave RAO (Z/A) Single SPAR TwoBody (SPAR #1) TwoBody (SPAR #2) 0.6 0.4 1.8 1.7 0.0 0.6 0.151 Surge Motion RAOs 5 Single SPAR 4 TwoBody (SPAR #1) TwoBody (SPAR #2) Surge RAO (X/A) 3 2 1 0 0.
0 2.2 0.0 1.0 1.4.5 2.152 Roll Motion RAOs 0.5 4.0 0.2 1.5 3.00 0.0E+04 0.0E+05 Single SPAR 8.4 1.04 0.c Comparison of the roll motion RAOs Wave Drift Force in Xdirection 1.4 0.0E+04 4.10 Single SPAR TwoBody (SPAR #1) TwoBody (SPAR #2) Roll RAO (theta/kA) 0.5 1.06 0.0 3.02 0.6 0.12 0.0 0.0E+00 0.0E+04 TwoBody(at Body #1) TwoBody(at Body #2) Drift Froce (N) 6.0 Frequency (rad/s) Figure 8.0E+04 2.08 0.8 1.6 Freqiency (rad/s) Figure 8.5 Comparison of the surge drift force .
73 18.119 174 45.91 10.67 0.89 3.45 33.44 2.22 0.23 0.73 1.19 0.629 292 7.63 2.650 115 45.97 0.382 215 7.170 24.374 29.393 4 SPAR 1 24.382 7.54 2.978 166 45.17 0.45 2.090 9.26 0.214 7.93 0.63 0.629 29. mean Riser min.33 1.033 9.82 5.045 2.67 0.48 0.75 29.380 3 SPAR 2 24.391 215 7.15 0.403 8.360 45.784 12.16 1.207 6.09 16.792 2.73 0.57 5.98 0.60 0.46 9.421 9.040 2.46 9.56 10.96 0.83 0.395 8.05 2.821 10.00 1.807 9.360 45.348 8.04 1.54 2.23 0.113 7.117 23.91 1.304 10. rms.37 10.04 5.441 10.73 5.03 15.20 2. rms. mean sway min.46 0.27 6.368 45.13 0.82 0.49 2.80 9. Line Tension mean Mooring min.18 0.377 2.38 1.05 0.743 8.38 2.351 8.49 0.22 1.19 0.90 1.4 The analysis results for twobody model composed of two spars SPAR+SPAR1) Single SPAR SPAR+SPRING+SPAR2) w/o interaction w/o hawser with hawser with interaction w/o hawser with hawser w/o interaction with a linear spring SPAR 1 Body Motion mean surge min.658 115 45.937 10.78 0.232 6.14 0.04 0.426 8.85 0.49 1.823 9.13 6. rms.24 0.021 8. (deg) max.152 7.678 25.26 0.05 0. (kN) rms.64 0.132 10.04 0. mean pitch min.095 8.39 2.723 9.05 0.31 1.05 0.672 11. (kN) rms.69 17.33 6.11 0.356 8.996 8.361 45.095 9.59 1.67 0.40 30.53 0.392 4 SPAR 2 24.368 45.15 0.27 0.93 0.49 1.72 3.660 164 8.70 0.40 0.710 9.29 2.022 9.67 0.079 2.26 6.361 45.368 45.89 2.24 0.91 1.05 15.39 2.34 0.88 2.246 6. (kN) max.071 10.74 0.368 45.774 128 45.368 45.55 2.66 5.36 1. 24.65 0.16 0.217 7.47 1.678 116 45.83 0.22 0. line #1 max. mean Mooring min.871 224 8.86 11.769 7.393 237 7.339 10. 2) A linear spring of the same stiffness as the hawser is put directly in the system stiffness matrix.18 0.44 9.95 1.423 24.96 1.68 0.431 10.24 0.17 6.36 2. (deg) max.67 0.73 18.43 2.40 2.368 45.46 0.09 0.41 33. rms.04 0.377 228 7.553 7.91 1.04 0.05 0.07 6.60 1.251 6.80 0.04 0.16 0.04 17.360 45. mean Mooring min.52 1.41 2.368 45.389 4 SPAR 2 25.14 6.774 137 45.96 0.876 1.25 6.27 0.48 16.964 193 8.579 121 45.568 236 7. mean roll min.01 5.958 9.57 31.388 3 SPAR 2 25.04 0.381 3 66 29 207 25 SPAR 1 23.15 0.04 0.839 9.78 1.50 0. line #2 max.702 8.16 0.833 229 8.23 0.56 1.160 10.43 1.375 245 7.088 8.610 45.12 0. rms.222 6.60 1.57 3. (kN) rms.003 9. (m) max. rms.16 0.60 31.16 4.093 7.086 8.64 0.24 0.360 45.49 0.04 0.092 25.597 290 7.04 16.34 2.60 0.18 6.39 2.368 45.759 235 8.392 4 24.46 31.17 0.66 18.360 45.717 2.36 9. line #4 max.987 1.48 1.168 6.380 3 SPAR 1 23.44 2.403 7.350 12. mean yaw min.59 18.71 18.085 8.348 8.369 45. mean Hawser min.32 30.260 9.882 9.15 0.369 45.790 133 45.304 10.49 0.64 0.578 236 7. .094 8.28 0.07 0.40 6.280 10.87 1.587 27.19 0.03 16.360 45.31 0.070 10.57 2.69 0.152 11. rms.04 17.17 6.072 8.831 9. (m) max.815 9.356 8.11 0.361 45.823 203 8.04 0.63 18.57 0.62 3.83 0.05 5.412 8.73 1.45 17.817 9.85 2.368 45.22 0.05 0. (deg) max.393 4 66 31 171 21 SPAR 1 23.244 6. mean heave min.153 Table 8.54 2.669 161 8.958 9.259 9.823 9.778 7.72 2.09 3.95 19.81 30.686 26.36 1.573 120 45.49 0.04 0.18 6. (kN) rms. rms.53 0.54 0.94 7. line #3 max.859 215 8.850 4 Remarks: 1) Both SPARs have 4 equivalent mooring lines and 1 equivalent central riser.745 9.04 0.47 10. (kN) max.223 2.85 15.64 0.658 125 45.711 2.389 215 7.31 2.03 16.863 218 8.00 1.40 0.162 10.63 0.54 0.19 0.196 24.78 1.87 2.26 0. mean Mooring min.225 2.67 0.333 7.85 1.633 136 8.17 4.16 0.19 5.216 6.03 0.16 0. (m) max.19 0.165 6.50 0.173 23.78 0.18 0.018 8.385 4 SPAR 2 24.638 7.322 10.16 30.436 238 7.16 6.15 0.40 0.619 132 8.89 0.65 1.06 2.40 0.768 11.354 8.220 6.04 1.54 18.23 1.48 18.024 12.03 0.081 8.222 6.45 2.84 30.79 0.71 9.18 1.17 0.131 10.03 16.78 0.19 1.19 0.01 0.279 12.16 0.360 45.
7 Summary and Conclusions The multiple body interaction effects on the twobody model of two spars due to the hawser connection and the hydrodynamic interaction effects are investigated by comparative study using two numerical models.154 8. When the second body tends to move out of phase against the first body motion. the statistical results of the motions of two bodies are shown to be identical. . When a linear spring is used. It shows that the hawser acts as a compensator for the second body in the lee side. and the front structure acts as a protector for the rear structure when the environmental loads are applied to the first structure collinearly with the direction of the body connection. However. for the sway motion. It is why the flow route of the external forces of wind. the second body will be able to move within a certain range. On the whole point of view. Therefore. With comparing this. the results of the hawser connection model make the two bodies move a little differently. The hydrodynamic interaction effect is exhibited well in the six DOF motions as the motions of the second body. except the sway motions are a little bit smaller than those of the other. wave and current is restricted by the protection effect of the front structure. the fluid interaction effect is clearly illustrated in the leeside structure. it makes the second body move to the opposite direction. the results must be an ideal case. it is hard to say that the second body will move less that the first body. So.
The other twobody model is composed of an FPSO and a shuttle tanker. an FPSOFPSO and an FPSOShuttle tanker are taken as the multiplebody models for the verification of the program (WINPOSTMULT) for the dynamic coupled analysis of the multiplebody floating platforms. . and the results are compared with the exact solution using a twomassspring model. For the multiplebody model of the FPSOshuttle tanker. The conventional tandem moorings have been used for the multiplebody connections in many cases of the operation of offloading in the sea. The simplified massspring model will give a compatible result to judge the validity of the multiplebody program. the interaction characteristics for the tandemmoored vessels are calculated in regular waves at several frequencies by using WAMIT. the tandem mooring is considered to investigate the interaction effect. An FPSOFPSO model consists of two identical FPSOs.155 CHAPTER IX CASE STUDY 5: DYNAMIC COUPLED ANALYSES FOR TWOBODY SYSTEM COMPOSED OF AN FPSOFPSO AND AN FPSOSHUTTLE TANKER 9. In this study.1 Introduction In this chapter. The body motions and line tensions are mainly reviewed with the numerical calculations performed by WINPOSTMULT. The coupled analysis results for the model of two identical FPSOs by the WINPOSTMULT program are compared with the exact solution for the twomassspring model. the dynamic coupled analysis program for multiplebody platforms.
they are equivalently combined as 4 groups for mooring lines and 1 group for risers. The main particulars and dimensions of the shuttle tanker are taken as the same as the FPSO’s. models both with a hawser and without a hawser are made and analyzed. The numerical models and the particulars of the mooring systems are the same as the FPSO’s reviewed in . The WAMIT program is used for the calculation of the hydrodynamic coefficients of the vessels. The interaction effect is studied as well for this model. The water depth is 6. The distance between the two FPSOs in the tandem mooring is taken as 30 meters. 9.8 m).1. The arrangement of the tandem is shown in Figure 9. For this verification. which is the same as in the previous chapter.000 ft (1828. including the principle data of spar. the effect of the hawser to connect two structures is also specified. The distance of the tandem mooring system is taken as 30 meters. for simplification. and one riser group is composed of all (13) risers.2. For the mooring system. The original FPSO studied in Chapter V has 12 taut mooring lines and 13 steel catenary risers(SCR). a tandem mooring is taken into account.156 From this study. are listed in Table 9. The tandem mooring has been used for many years. The validity of the numerical modeling was already proven in the previous chapters by the static offset test and free decay tests. Here.2 Particulars of Models and Mooring Arrangements The main particulars. The mooring lines are fixed at the sea floor. The configuration for the mooring of the equivalent mooring lines is shown in Figure 9. Each mooring line group has 3 legs.1. The riser group is centralized on the geometrical center of the turret.
5 0.000 200 310.6 13.32 5.000 1.57 2.772 63.440.0 13.78 403. The hawser connecting the two FPSOs and the FPSOShuttle tanker has the stiffness of 1/100 of the mooring stiffness and the pretension of 1/10 of the mooring pretension.52 15.85 . Main particulars of the mooring systems are summarized in Table 9.47 79.55 1.1 Main particulars of the turret moored FPSO Description Production level Storage Vessel size Length between perpendicular Breadth Depth Draft (in full load) Diaplacement (in full load) Lengthbeam ratio Beamdraft ratio Block coefficient Center of buoyancy forward section 10 Water plane area Water plane coefficient Center of water plane area forward section 10 Center of gravity above keel Transverse metacentric height Longitudinal metacentric height Roll raius of gyration in air Pitch raius of gyration in air Yaw radius of gyration in air Frontal wind area Transverse wind area Turret in center line behind Fpp (20.2.83 14.869 6.157 Chapter V.400 0.85 6.77 77.5 % Lpp) Turret elevation below tanker base Turret diameter L/B B/T Cb FB A Cw FA KG MGt MGl Rxx Ryy Rζζ Af Ab Xtur Ztur m m m m m m m m2 m2 m m m m m2 Lpp B H T Symbol Unit bpd bbls kDWT m m m m MT Quantity 120.0 47.30 1.9164 1.09 240.012 3.17 28. Table 9.04 18.
515 deg m m Unit kN Quantity 1. MBL m mm kg/m kg/m kN kN 45.0 m Wind FPSO 2 or Shuttle Tanker FPSO 1 Wave Current Figure 9.201 4*3 5 2.515 m mm kg/m kg/m kN kN 1127.9 42.9 7.4 88.087.7 88.158 310.168 6.0 m 30. AE Mean breaking load.0 .421 m mm kg/m kg/m kN kN 914.841 6. AE Mean breaking load.7 690.2 Main particulars of the mooring systems Description Pretension Number of lines Degrees between 3 lines Length of mooring line Radius of location of chain stoppers on turn table Segment 1 (ground position): chain Length at anchor point Diameter Weight in air Weight in water Stiffness.4 794.1 Configuration of the mooring systems (Tandem mooring system) Table 9. MBL Segment 2: chain Length Diameter Weight in air Weight in water Stiffness.8 107.4 794.9 164. AE Mean breaking load. MBL Segment 3 (hangoff position): chain Length Diameter Weight in air Weight in water Stiffness.841 6.9 143.0 35.9 143.9 164.
2 Configuration of the arrangement of the mooring line groups . #2 Equiv.159 NORTH #6 #5 #4 #7 #8 #9 #3 #2 #1 EAST #10#11 #12 Equiv. #4 Figure 9. #3 Equiv. #1 Equiv.
2 rad/s. and it is reduced as 0. The API wind velocity spectrum is also used. The wave frequencies are taken account of the range from 0. and it is summed up with a random phase at every time.a and 9. API RP 2T is referred to obtain the wind velocity spectrum. but the wind speed is slower than that in the 100yr.0914 m/s at the sea floor. storm condition.3 Environmental Conditions The environmental conditions correspond to the 100year storm conditions in GoM and the sea condition of West Africa. For the wave force. JONSWAP spectrum is used. The wave is calculated at every frequency. The environmental conditions at GOM and at the west Africa sea are summarized in Tables 9. The reason that the mild condition is taken for the FPSOShuttle tanker model is that the tandem mooring system for transferring oil or gas from the FPSO to the shuttle tanker in the real open sea has been tried in a rather mild sea condition for the safety. The current velocity is 1. respectively.160 9. The incident wave heading in hurricane conditions is 180 o when the xcoordinate is set to the East and yaxis is set to the North.3. . The current speed in the West Africa is less than that in GoM.5 rad/s to 1. The wave heading of this condition is 180 o when the xcoordinate is set to the East and yaxis is set to the North. dividing the range by 100 intervals. The west Africa sea conditions are used for the twobody model of an FPSO and a shuttle tanker. For the wind force. It varies linearly to the sea floor.3.b.0668 m/s at the free surface. The 100year storm conditions are used in the case of tandem moored vessels of the two body model of an FPSO and an FPSO.
96 m at 91.0 m/s @ 10m API RP 2AWSD deg 210 1) m/s m/s m/s m/s m/s deg 0.5) deg 180 1) m/s deg 41.150 0.050 0. Hs Peak period.0914 150 1) Remark: 1) The angle is measured counterclockwise from the xaxis (the East).161 Table 9.12 m/s @ 10m API RP 2T 210 1) m/s m/s m/s m/s deg 1. .70 sec 16.150 0.44 m on the sea bottom Direction Unit Quantity m 12.0 JONSWAP ( γ =2. Hs Peak period.19 sec 14.0668 1. Table 9.3.96 m at 91. Tp Wave spectrum Direction Wind Velocity Spectrum Direction Current Profile at free surface (0 m) at 60.a Environmental conditions (100year storm condition at GOM) Description Wave Significant wave height.0914 0.5 JONSWAP ( γ =6.0668 0.050 150 1) Remark: 1) The angle is measured counterclockwise from the xaxis (the East).0) deg 180 1) 5. Tp Wave spectrum Direction Wind Velocity Spectrum Direction Current Profile at free surface (0 m) at 60.b Environmental conditions (west Africa sea condition) Description Wave Significant wave height.3.44 m on the sea bottom Direction Unit Quantity m 2.
it is not practical to calculate the coefficients at .3. Under the circumstances of applying the environmental conditions associated with wave. The hydrodynamic coefficients of added mass. she will move and rotate freely. the mooring lines and risers are installed only for FPSO. it will pursue the dynamical equilibrium position corresponding to the neutral location for the sum of the environmental loads to be zero and trace the path by itself. For a twobody model composed of FPSO and FPSO. In Figure 9. the hydrodynamic coefficients at every angle should be calculated for the dynamic analysis.4 Calculation of Hydrodynamic Coefficients Using WAMIT The hydrodynamic coefficients are calculated by WAMIT. the model for the wave force coefficients is shown. and the shuttle tanker has no mooring line and riser. The modeling is made only for the port side. For the twobody analysis. and the number of elements is 1684. Accordingly. the wave force coefficients and wave drift force coefficients are calculated. in the timedomain simulation. For FPSO and shuttle tanker model. the West Africa sea condition is applied. It is well known that the range of yaw angle in which she may move in the 100year storm condition will be about 10~20 degrees. A turretmoored FPSO has been designed to weathervane in the sea so that the mooring lines and risers are only connected at the bottom of turret. FPSO and the shuttle tanker are connected with one hawser. After that. the mooring lines and risers are connected as what they are. However. But. for a twobody model composed of FPSO and a shuttle tanker. wind and current load. wave damping and linear transfer function (LTF) of diffraction potential force are calculated by WAMT. and the 100year storm conditions at GoM are applied.162 9.
the coefficients are calculated prior to the coupled analysis. If the yaw angle is beyond 5 degrees from the starting position. when the coupled analysis of the body and the mooring system is performed. the other coefficients are read from the premade files.163 every time step. twobody models and the mooring system . In this study. (a) A single–body FPSO model (b) Twobody model of FPSO and FPSO ( or Shuttle tanker) in tandem arrangement Figure 9. at every time step the yaw angle is checked. at every 5degree interval.3 Configuration of singlebody. So.
The masses are .3 Continued 9. The idealized model is shown in Figure 9. The environmental loads are calculated using Morison’s equation for the wind and current forces and the JONSWAP spectrum formula for the wave force.164 FPSO #2 FPSO #1 SEA BED (Tandem Arrangement) (c) Configuration of moorings for twobody model of FPSO and FPSO Shuttle Tanker FPSO SEA BED (Tandem Arrangement) (d) Configuration of moorings for twobody model of FPSO and Shuttle tanker Figure 9.4.5 TwoMassSpring Modeling The twomassspring model is devised to get an exact solution for the idealized twobody FPSO model and is used for verifying the numerical analysis results by the WINPOSTMULT program.
165 determined to add the FPSO body mass and the added mass at around surge natural frequency. Spring constants are calculated by considering the total top tension of the mooring lines and risers in the horizontal direction. The hawser stiffness can be directly converted to the linear spring in the middle of the idealized model.
F1 X1 K1 M1 K2 M2 F2 X2 K3
Figure 9.4 Twomassspring model
The wind force in xdirection, Fxw , is obtained from Morison’s formula and OCIMF wind coefficient as:
1 Fxw = C xw ρ w AT Vw 2 2
(9.1)
where C xw is the wind force coefficient that can be read from the OCIMF document, ρ w is the water density, AT denotes the projected area in the lateral direction of the vessel against wind, and Vw is the wind velocity. The wind force by API RP 2T, Fww (1) ,
represents the force per unit area in the normal direction to the wind blowing, and is given by:
Fww (1) =
1 ρ wVw 2 2
(9.2)
166 Here, in this study, the unit wind force, Fww (1) , is calculated by a separate program, and the resultant wind force is computed in the WINPOST program, since the force varies according to the wind blowing direction. In WINPOST, the yaw angle of the body at every time step is checked, and the wind force coefficient is interpolated by using the reading data from the OCIMF document. AT is given by a user as an input data. In ydirection, the wind force is obtained in the same way by the following formula:
Fyw = C yw AL Fww (1) (9.3)
where C yw is the wind force coefficient in ydirection obtained from the OCIMF document, and AL denotes the projected area in the longitudinal direction to be normal to the wind. As the initial wind direction is considered to be 210 o counterclockwise
from the xaxis (true East), the coefficients of C xw and C yw are evaluated as 0.73 and 0.30, respectively, in the full load condition. The current forces, Fxc in xdirection and Fxc in ydirection, are also calculated from Morison’s formula as follows: In xdirection:
Fxc =
1 C xc ρ cVc 2 L ppT 2
(9.4)
In ydirection:
1 Fyc = C yc ρ cVc 2 L ppT 2
(9.5)
Where L pp and T are the same as in Table 9.1, ρ c is the water density, and Vc is the current velocity, and here current speed is used at the free surface. The current
167 coefficients, C xc and C yc are evaluated as 0.024 and 0.922, respectively, by considering the initial current direction of 150o from the xaxis counterclockwise. The formula of the JONSWAP wave spectrum was written in Chapter V (equation (5.1)). If the significant wave height, H s , the peak period, T p , and overshooting parameter, γ , are taken in Tables 9.3.a and 9.3.b, the wave can be estimated at any time with random phases.
Fφ (ti ) = ∑ A(ω j ) cos ω j ti + φ j
j
(
)
(9.6)
where i and j are the indices for representing the time instant and the frequency of any wave component, ω j is the frequency of the incident wave component j , A(ω j ) is the wave amplitude, and φ j is the random phase between wave components. The total force is determined as the linear sum of the equation (9.2) ~ (9.6) as: F1 (t ) = F2 (t ) = Fw + Fc + Fφ (9.7)
where F1 (t ) and F1 (t ) are the applied forces to the mass M 1 and M 2 in the idealized model, and M 1 and M 2 represent the virtual masses made of the mass weights and the added masses of the FPSOs. The body mass and stiffness are obtained by considering the mass weight of FPSO, m , the added mass, ma , and the line top tension as follows:
M 1 = M 2 = m + ma K1 = K 3 = stiffness of mooring lines and risers K 2 = stiffness of the hawser
(9.8) (9.9) (9.10)
168
Table 9.4 The system parameters for twomassspring model
ITEM Added mass FPSO weight in mass Mass of FPSO #1 Mass of FPSO #2 Stiffness of mooring #1 Stiffness of hawser Stiffness of mooring #2 Natural period (Mode #1) (Mode #2) Symbol ma m M1 M2 K1 K2 K3 Unit kg kg kg kg N/m N/m N/m sec sec Magnitude 1.466E+07 2.397E+08 2.543E+08 2.543E+08 2.389E+05 1.868E+03 2.389E+05 16.34 205.02
forces1
Mux Mux current Mux1 Mux2 Mux wave
Mux3
f(u) Current 1 f1_wave Wave Force 1 t Clock To Workspace1 1 Gain [time, wf] KF1 Mux x' = Ax+Bu y = Cx+Du StateSpace disp Kres Demux emu velo Wind Force 1 Gain3 Demux1 Demux
f2_wave Wave Force 2
1 Gain1 [time, wf] Wind Force2 Gain2 f(u) Current 2 F2 Mux
forces2
Mux
Mux wind
Mux4
Mux5
Figure 9.5 The diagram of the time simulation in SIMULINK of MATLAB
The time simulation for the massspring model is performed using MATLAB.0 0.0 10.6 The surge motion of the FPSO and FPSO model by MATLAB for massspring model and by WINPOSTMULT for twobody model .5.0 M as s #2 Displacement (m) 0.0 40.0 20.0 50.169 The calculated results to get the idealized twomassspring model are summarized in Table 9.0 0 500 1000 1500 2000 2500 3000 3500 4000 Time (sec) (b) The surge motion of FPSO+FPSO model by WINPOSTMULT (without the interaction effect) Figure 9.0 30.0 20.0 FPSO #1 FPSO #2 Surge motion (m) 20. The calculation diagram in MATLAB is depicted in Figure 9.0 40.0 30.0 0 500 1000 1500 2000 2500 3000 3500 4000 4500 Time (sec) (a) The displacements at mass #1 and #2 of the massspring model by MATLAB Timesimulation results for FPSO+FPSO model (without the interaction effect) 30. For the validity of the model data. TimeSimulation Result Using MassSpring Model M as s #1 10.0 10.0 10.4. the eigenvalues are checked using MATLAB.
0 40.97 11.0 20.99 42.08 .46 14.0 40.0 0.45 38. Max.0 10.0 30.71 8.0 10.47 15.0 0.0 50.0 0 500 1000 1500 2000 2500 3000 3500 4000 FPSO #1 FPSO #2 Surge motion (m) Time (sec) (c) The surge motion of FPSO+FPSO model by WINPOSTMULT (with the interaction effect by iteration method) Timesimulation results for FPSO+FPSO model (with the interaction effect) 20.55 14.0 20.6 Continued Table 9.5 Analysis results of massspring model: displacement at mass #1 and #2 (unit: m) Mean Min. RMS Mass #1 Mass #2 15.0 0 500 1000 1500 2000 2500 3000 3500 4000 FPSO #1 FPSO #2 Surge motion (m) Time (sec) (d) The surge motion of FPSO+FPSO model by WINPOSTMULT (with the interaction effect by combined method) Figure 9.0 50.0 10.0 30.170 Timesimulation results for FPSO+FPSO model (with the interaction effect by iteration method) 20.0 10.
72 34.73 26.697 2.55 0.231 5.08 1.34 9.711 3.78 3.328 3. (deg) max.857 4.968 4.44 5.30 6.61 2.37 2.19 3.83 1.57 3.634 7.79 2.601 FPSO 1 14.734 4.516 10.31 16. max.100 102.55 5.81 4. (deg) max.76 0.455 866 878 913 920 677 417 766 669 709 431 815 484 4.05 1.41 2.291 701 3.68 5.50 9.27 4.83 0.00 3.23 3.480 FPSO 2 13.49 1.68 5.65 1.691 4.82 2. mean pitch min.754 668 1.92 3.74 2.39 2. mean heave min.83 2.34 3.20 1.87 4.24 0.900 663.100 45.884 2.69 3.040 4.73 12.70 1.82 6.300 127.208 2.404 2.58 5.52 0.700 671.23 3.44 3.900 75.20 4.800 108.26 12.809 4.02 1.44 9.55 4.40 11.55 7.828 5.42 0.84 13.43 5.09 11.48 10.349 3.929 2. rms.677 7.765 3.43 3.728 3.100 130.70 37.704 6.194 3.50 2.01 4.87 1.11 7.46 2.30 9.72 0.32 9.48 1.11 1.23 2.06 3.724 5.43 2.700 132.500 73.45 0.78 1.66 0.32 0. mean roll min.61 16. mean min.193 3.86 33.073 6.02 1.56 10.19 34.796 4.780 4.44 5.12 4.53 2.18 6.03 8.784 4.51 3.167 654 1.41 5.780 4.430 FPSO 2 13.271 3.47 3.506 3.26 5.171 Table 9.01 0.937 3.098 5.350 5.53 10.93 1.01 0.704 2.490 FPSO 1 13.34 0.41 0.100 638.72 2.330 FPSO 2 10.65 5. mean min.57 0.37 7.46 4.900 7.89 4.500 316.09 4.44 0.77 3.53 1.56 2.30 6.87 5.39 6.41 1.91 12.25 3.847 2.45 9.19 0. rms.06 1.868 622 1.617 3.07 5.312 8. (kN) Mooring line #2 (kN) Mooring line #3 (kN) Mooring line #4 (kN) rms. max.48 5.35 11.29 18.29 6.93 2.18 8.297 927 1.080 FPSO 2 7.300 49.98 1. Line Tension mean Mooring min.41 3.26 2.873 4.42 8.47 1.44 3.87 1.792 496 460 383 409 473 440 435 500 500 405 427 405 2.360 103.93 6.768 488 2.00 5.00 4. (m) max.887 3. line #1 max.37 0.17 1.63 35.989 2. rms.805 2.57 2.67 0.019 535 608 558 530 785 1.68 1.35 5.42 6.326 3.47 16.37 0.710 1.700 FPSO 2 9.060 131.53 11.29 9.900 254.01 4.33 0.90 0.51 0.28 0.59 14.54 4.81 1.609 4.43 0.63 8.800 703.631 1. max.685 4.270 110.37 5. mean sway min.77 1.01 5.56 6.06 4.585 570 4.52 0.45 5.41 1.565 3.621 3.82 0.006 120.000 128. 14.679 3.82 4.33 0.88 1.97 1.38 0.51 3.42 3.52 20.818 Riser (kN) Hawser (kN) 1.59 0.870 106.360 0 0 0 0 0 0 0 0 0 0 0 0 724.853 767 4.6 Summary of the analysis results for two body FPSO+FPSO FPSO+FPSO Single FPSO 1) w/o interaction w/o hawser with hawser with interaction (by iteration method) w/o hawser with hawser with interaction (by combined method) w/o hawser with hawser FPSO 1 Body Motion mean surge min.64 2.83 21.20 0.91 0.56 2.29 1.18 4. rms.759 4.345 3.23 3.306 3.50 7.72 1.67 3.543 10.001 10.803 4.25 0.15 2.369 7.14 3. rms.86 2.556 3.07 8.700 73.700 107.15 0. rms.80 17.62 0.45 13.65 21.91 2.53 4.41 8.13 8. (m) max.59 2.50 0.61 0.193 3.45 3.537 1.72 0.642 3.38 0. rms.783 4.373 9.07 1.75 10.200 274.672 5.04 1.32 21.89 9.65 0.04 2. (deg) max.45 3.470 1.67 8.051 5.38 1.377 1.01 5.300 14.97 2.82 2.000 48.34 0.757 FPSO 1 13.23 0.94 3.95 20.604 1.43 6.70 3.184 673 1.53 14.57 0.40 3.23 0.59 4.09 3.102 2.496 3.15 5.553 3. .512 3.20 2.26 6.400 120. max. (m) max.45 3.43 5.995 5.81 3.98 13.98 4.300 734.704 4.400 74.634 2.613 6.31 10.24 36.466 1.79 1.800 0 676.416 3.598 591 542 430 483 563 513 462 534 619 452 492 462 109. rms.36 5.85 4.62 16.31 18.18 0.216 3. mean yaw min.60 0.85 0.28 12. rms.12 2.994 3. max. mean min.66 1.932 FPSO 1 14.600 721.67 1.24 10.662 2.49 1.419 4.619 3. mean min.00 4.40 5.88 1.45 7. mean min.24 8.28 1.237 4.580 6.284 4.400 131.15 1.04 4.879 4.03 7.24 0.859 10.639 2.75 5.098 2.52 2.300 255.98 33.923 4.08 1.025 10.18 1.28 9.55 8.477 3.22 1.42 1.47 0.47 1.49 2.00 2.872 3.349 6.24 23.110 FPSO 2 10.455 2.224 6.07 3.798 693 1.570 1.95 0. rms.13 9.956 6.728 4.554 2.413 3.373 1.384 3.25 8.286 2.25 1.900 655.19 8.99 6.000 5.34 24.89 2.300 108. rms.40 2.751 4.085 5.79 17.850 101 101 102 100 100 100 103 104 106 0 1 1 Remarks: 1) Both FPSOs have 4 equivalent mooring lines and 1 equivalent central riser.730 2.625 3.285 4.330 10.330 125.76 2.71 3.411 4.08 6.83 4.89 22.97 34.09 1.480 10.550 5.61 6.033 2.399 3.313 3.85 3.701 4.36 3.19 4.562 3.07 5.96 0.54 0.335 5.93 1.54 0.700 3.041 10.667 4.95 6.36 1.25 0.263 FPSO 1 13.889 4.788 2.27 9.747 6.87 1.802 7.906 561 109.53 0.36 37.
0 Shuttle Tanker Surge motion (m) 60.0 0 500 1000 1500 Time (sec) 2000 2500 3000 (b) The time simulation results of FPSO+shuttle tanker model by the iteration method (with the interaction effect) Timesimulation of surge motion for FPSO+Shuttle Tanker FPSO Shuttle Tanker 40.0 20.0 0.0 0 500 1000 1500 2000 2500 3000 Time (sec) (a) The time simulation results of FPSO+shuttle tanker model (without the interaction effect) Time simulation of surge motion for FPSO+Shuttle Tanker FPSO 40.0 20.0 10.0 40.0 20.0 20.0 0 500 1000 1500 2000 2500 3000 3500 4000 Time (sec) (c) The time simulation results of FPSO+shuttle tanker model by the combined method (with the interaction effect) Figure 9.0 Surge motion (m) 30.0 10.7 The time simulation results of the FPSO and shuttle tanker model .0 10.172 Time simulation of surge motion for FPSO+Shuttle Tanker FPSO 80.0 0.0 Surge motion (m) Shuttle Tanker 30.0 10.0 0.
99 2.33 0.21 0.36 0.51 0. max.449 78 69. (deg) max.14 24.84 0.81 9.10 2.80 0. mean min.69 FPSO 0. mean sway min.967 4.28 0.39 1.21 0.397 75 4.01 0.51 0.43 0.74 3.01 0.69 0.84 0.900 24. rms.27 0.60 1.31 0.51 8.018 4.03 0.26 5. mean roll min.26 0.23 0.57 5.48 3.560 0 151.944 5.72 6.16 4.15 m/s at free surface.48 0.54 3.28 0.268 4.40 0.50 2.974 4.28 0. max.0.353 74 4.79 0.34 0.35 10.60 1.600 24. 0.81 0.44 0. mean heave min.779 4.01 1.19 1.34 0.86 6.375 104 4.189 3.70 0.35 0.77 1.508 225 4. the current speed is 0.367 93 4.374 4.56 2. Line Tension mean Mooring min.47 0.51 0.11 FPSO 0. rms.086 4.17 2. (deg) max.77 1.13 54.811 4.81 1.300 23.20 0.16 8.122 4.67 0. rms.62 0.38 2.48 0.991 4.66 0.490 0 146.11 0.34 1. (kN) surge (m) Mooring line #2 (kN) Mooring line #3 (kN) Mooring line #4 (kN) rms.094 4.15 0.26 0.12 8. mean min. rms. max.433 78 69.69 0.041 4.87 0. Tp of 16. line #1 max.16 0.27 0.7 Summary of the analysis results for the twobody FPSO+shuttle tanker w/o interaction with hawser Single FPSO FPSO+Shuttle Tanker2) with interaction with interaction by the iteration by the combined method method with hawser with hawser FPSO Body Motion mean min.197 4.550 0 164.195 3.75 2.051 3.19 5.73 3.21 2.05 0.428 51 4.41 0.05 1.62 4.509 89 4.174 3.184 3.07 1.02 0.27 1.39 0.65 0.00 0. (m) max.66 1.257 4.97 0. mean min.32 0.27 1.47 0.59 8.487 74 4.5 sec.13 4. rms.339 3.12 0. rms.38 0.65 0.23 0.25 4.62 10.20 2.408 78 4.350 57 4.09 3. rms.85 0.08 0.965 4.73 0.730 0.68 4.16 1.170 254 5 844 254 Shuttle 21.52 0.50 14.00 0.91 5. max.28 0.298 4.530 0 150.39 0.46 2.11 0. max.410 119 6 296 86 Shuttle 16.86 0.41 0.15 17.54 0.72 0.375 67 4.21 0.21 0. rms.26 1.11 33.23 0.086 3. (deg) max.58 0. mean min.77 2. rms.11 4. max.79 0.60 1.210 4. which is intended to investigate the difference with the results by three methods in a mild loading condition (West Africa sea condition).00 0. and the wave has Hs of 2.66 4.7 m.67 2.98 1.50 2. rms.74 1.126 3. .00 0.946 4.26 0. rms.41 0.84 Riser (kN) Hawser (kN) 2) The loading condition is changed for this calculation.00 0.71 7.10 0.00 0.27 0.187 4.34 0.11 0.353 67 69.422 54 69.13 0. mean min. mean yaw min.918 4.00 0.52 1.173 Table 9.44 3.60 1.600 24.200 3.06 0.27 1.14 0.62 0.730 79 6 252 77 Shuttle 17.20 0.115 3.48 2.33 0.20 0.The wind velocity is 10 m/s at 10 m height.23 0.62 3.019 4. mean pitch min. and gamma of 6.050 232 4.81 0.46 7.34 5. (m) max.
the body stiffness matrix and mass matrix have only the terms for the single body. The analysis results for the FPSO and FPSO model are summarized in Table 9. The surge motion amplitude for each case is very similar.6(a)~(d). the displacements in xdirection (surge motion) by the time simulation analyses for the massspring model and the FPSO and FPSO model when the mooring is in tandem arrangement are depicted. The time simulation results are shown for the purpose of comparison in Figure 9.6 The two tables show that the statistical results are well matched with each other. the hydrodynamic coefficients induced by wave. That means. However.6 Results and Discussion In Table 9. In Figure 9. the analysis cases for the twobody model of an FPSO and a shuttle tanker are summarized for three different cases. and the interaction terms are set to zero. In the case of the “with the interaction effect by iteration .174 9. The hawser stiffness used for this analysis was 1/100th of the mooring stiffness. In the case of “no interaction”. In Table 9. The hawser stiffness used for this analysis was 1/1000th of the mooring stiffness. in this case.5. the statistics of the analysis results for the massspring model is shown. so that the validity of the program WINPOSTMULT for the twobody analysis with one hawser is proved.7. and the top tension of the hawser was taken as 1/10th of the mooring line tension.7. the interaction effect between two vessels of the fluid and the structures is not considered. whether the interaction effect is considered or not affects the shape and the phase difference between surge motions of two bodies in the time simulation. and the top tension of the hawser was taken as 1/10th of the mooring line tension.
the analysis results for the case of “with interaction by the iteration method” give medium values among the results for the cases of “with no interaction” and “with interaction by the combined method”. Thus.10d.7. It means that it is significant to consider the fully coupled interaction effect for the twobody analysis. In the twobody model of the FPSO and shuttle tanker. In Table 9. the selfcoupling terms in the hydrodynamic coefficients. The program WINPOSTMULT has the kind function of performing the above three cases by handling the system matrix or the hydrodynamic coefficient matrices. the time histories and the motion amplitude spectra are shown for all analysis cases.175 method” for the twobody model. the fully coupled matrices are used for the analysis. To review the motion amplitude spectrum for each case.8a through 9. In all motions at the rear side vessel. the vessels have almost the same characteristics in their dynamic behaviors. to review the results of all cases can make some clues drawn about the hawser connection effect and the hydrodynamic interaction effect between two bodies. The purpose of this study is to compare the analyzed results by the developed program with the results produced by the methods used in the industry. . the twobody stiffness matrix and the twobody mass matrix are only considered. the interaction and hawser effects are clearly illustrated. From Figures 9. In the case of the “with the interaction effect by the combined method”. the interaction terms between two bodies are set to zero.
197 10 500 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 4.595×10 3 ti time (sec) 0. without interaction effect) .423 1 0 Heave (m) heave1 i 1 − 1.574 2 500 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 4.932 5 0 Surge (m) surge1i 5 − 7. tandem.8.546 2 Sway (m) sway1i 0 − 1.176 1.595 10 × 3 ti time (sec) Figure 9.595×10 3 ti time (sec) 1.446 2 500 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 4.a Time simulation for the two body model of the FPSO and shuttle tanker (at body #1=FPSO.
434 0.a Continued .8.177 0.508 1 500 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 4.595×10 3 ti time (sec) Figure 9.595×10 3 ti time (sec) 3.5 Roll (deg) roll1 i 0 − 0.968 1 Pitch (deg) pitch1i 0 − 0.167 5 0 Yaw (deg) yaw1i 5 − 5.5 500 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 4.852 10 500 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 4.595×10 3 ti time (sec) 0.485 0.
8. without interaction effect) . tandem.595 10 × 3 ti time (sec) Figure 9.921 50 500 500 1000 1500 2000 2500 ti 3000 3500 4000 4500 5000 4.648 20 10 Sway (m) sway2i 0 − 8.b Time simulation for the two body model of the FPSO and shuttle tanker (at body #2=shuttle tanker.348 100 50 Surge (m) surge2i 0 − 13.666 5 500 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 4.595×10 3 time (sec) 11.595 10 × 3 ti time (sec) 4.194 5 Heave (m) heave2 i 0 − 2.595 10 500 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 4.178 64.
69 2 500 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 4.b Continued .159 2 Pitch (deg) pitch2 i 0 − 1.813 1 500 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 4.81 1 Roll (deg) roll2 i 0 − 0.595×10 3 ti time (sec) 1.595 10 × 3 ti time (sec) 11.595 10 × 3 ti time (sec) Figure 9.843 20 10 Yaw (deg) yaw2 i 0 − 4.179 0.103 10 500 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 4.8.
180
Sureg Amplitude (m)
0.663
1
Asp j 0.5
9.4 ×10
−4
0
0 0
0.2
0.4 freq j frequency(rad/s)
0.6
0.8
1 1.01
Sway Amplitude (m)
0.234
0.3 0.2
Asp j 0.1 2.432 ×10
−4
0
0 0
0.2
0.4 freq j frequency(rad/s)
0.6
0.8
1 1.01
Heave Amplitude (m)
0.056
0.1
Asp j 0.05
4.134×10
−5
0
0 0
0.2
0.4
0.6 freq j frequency(rad/s)
0.8
1 1.01
Figure 9.8.c Amplitude spectrum density curve of the motion responses for the two body model of the FPSO and shuttle tanker (at body #1=FPSO; tandem; without interaction effect)
181
Roll Amplitude (deg)
0.022
0.03 0.02
Asp j 0.01 3.977×10
−5
0
0 0
0.2
0.4
0.6 freq j frequency(rad/s)
0.8
1 1.01
Pitch Amplitude (deg)
0.041
0.06 0.04
Asp j 0.02 2.322×10
−5
0
0 0
0.2
0.4
0.6 freq j frequency(rad/s)
0.8
1 1.01
Yaw Amplitude (deg)
0.921
1
Asp j 0.5
6.074×10
−4
0
0 0
0.2
0.4
0.6 freq j frequency(rad/s)
0.8
1 1.01
Figure 9.8.c Continued
182
Surge Amplitude (m)
12.615
15 10
Asp j 5 0.015 0
0 0
0.2
0.4 freq j frequency(rad/s)
0.6
0.8
1 1.01
Sway Amplitude (m)
2.163
3 2
Asp j 1 1.61×10
−3
0
0 0
0.2
0.4
0.6 freq j frequency(rad/s)
0.8
1 1.01
Heave Amplitude (m)
0.254
0.3 0.2
Asp j 0.1 1.669×10
−4
0
0 0
0.2
0.4
0.6 freqj frequency(rad/s)
0.8
1 1.01
Figure 9.8.d Amplitude spectrum density curve of the motion responses for the two body model of the FPSO and shuttle tanker (at body #2=shuttle tanker; tandem; without interaction effect)
183
Roll Amplitude (deg)
0.049
0.06 0.04
Asp j 0.02 2.585 ×10
−5
0
0 0
0.2
0.4 freq j frequency(rad/s)
0.6
0.8
1 1.01
Pitch Amplitude (deg)
0.109
0.15 0.1
Asp j 0.05 6.176×10
−5
0
0 0
0.2
0.4
0.6 freq j frequency(rad/s)
0.8
1 1.01
Yaw Amplitude (deg)
1.69
2
Asp j
1
1.423×10
−3
0
0 0
0.2
0.4
0.6 freq j frequency(rad/s)
0.8
1 1.01
Figure 9.8.d Continued
a Time simulation the for two body model of the FPSO and shuttle tanker (at body #1=FPSO.184 0. tandem.377 2 Sway (m) sway1i 0 − 1.595×10 3 ti time (sec) 1. with interaction effect by iteration method) .162 2 500 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 4.798 2 0 Surge (m) surge1i 2 − 2.229 4 500 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 4.9.337 1 0 Heave (m) heave1i 1 − 1.499 2 500 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 4.595×10 3 ti time (sec) Figure 9.595×10 3 ti time (sec) 0.
185 0.462 1 500 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 4.502 5 500 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 4.595×10 3 ti time (sec) Figure 9.595×10 3 ti time (sec) 0.392 0.338 5 Yaw (deg) yaw1i 0 − 4.5 Roll (deg) roll1 i 0 − 0.5 500 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 4.423 0.9.922 1 Pitch (deg) pitch1i 0 − 0.595×10 3 ti time (sec) 2.a Continued .
808 10 500 500 11.595 ×10 3 time (sec) 3.466 5 Heave (m) heave2i 0 − 1.843 5 500 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 4.155 40 Surge (m) surge2i 20 3.031 0 500 500 1000 1500 2000 2500 ti 3000 3500 4000 4500 5000 4.595×10 3 time (sec) 20 10 sway2 i 0 − 2.186 33. tandem.595×10 3 ti time (sec) Figure 9. with interaction effect by iteration method) .883 Sway (m) 1000 1500 2000 2500 ti 3000 3500 4000 4500 5000 4.b Time simulation for the two body model of the FPSO and shuttle tanker (at body #2=shuttle tanker.9.
617 0 500 500 1000 1500 2000 2500 ti 3000 3500 4000 4500 5000 4.b Continued .285 0.595×10 3 time (sec) Figure 9.595×10 3 time (sec) 10.763 Pitch (deg) 1000 1500 2000 2500 ti 3000 3500 4000 4500 5000 4.595×10 3 time (sec) 1 0 pitch2 i 1 − 1.329 2 500 500 0.5 500 500 1000 1500 2000 2500 ti 3000 3500 4000 4500 5000 4.307 0.187 0.9.5 Roll (deg) roll2i 0 − 0.172 15 10 Yaw (deg) yaw2 i 5 0.
4 freq j frequency(rad/s) 0.8 1 1. with interaction effect by iteration method) .4 freq j frequency(rad/s) 0.691 ×10 −5 0 0 0 0.3 0.8 1 1.8 1 1.1 1.9.2 0.01 Figure 9.2 Asp j 0.185 0.01 Sway Amplitude (m) 0.1 1.1 Asp j 0.c Amplitude spectrum density curve of the motion responses for the two body model of the FPSO and shuttle tanker (at body #1=FPSO.255 ×10 −5 0 0 0 0.055 0.6 0.6 0.4 freq j frequency(rad/s) 0.6 0. tandem.2 Asp j 0.2 0.285 ×10 −4 0 0 0 0.01 Heave Amplitude (m) 0.227 0.2 0.05 4.188 Sureg Amplitude (m) 0.
03 0.411 ×10 −5 0 0 0 0.02 Asp j 0.6 0.97 ×10 −5 0 0 0 0.8 1 1.189 Roll Amplitude (deg) 0.2 0.4 freq j frequency(rad/s) 0.6 0.4 freq j frequency(rad/s) 0.039 0.6 0.8 1 1.024 0.8 1 1.01 Figure 9.9.02 2.419 ×10 −3 0 0 0 0.c Continued .01 2.5 1.01 Pitch Amplitude (deg) 0.2 0.2 0.584 1 Asp j 0.4 freq j frequency(rad/s) 0.04 Asp j 0.01 Yaw Amplitude (deg) 0.
190 Surge Amplitude (m) 3.8 1 1.2 0.8 1 1.6 0.9.4 Asp j 0.2 2.2 0.212 3 2 Asp j 1 2.8 1 1.148 ×10 −3 0 0 0 0.4 freq j frequency(rad/s) 0.4 freq j frequency(rad/s) 0. tandem.272 0.01 Heave Amplitude (m) 0.d Amplitude spectrum density curve of the motion responses for the two body model of the FPSO and shuttle tanker (at body #2=shuttle tanker.976 ×10 −4 0 0 0 0.01 Sway Amplitude (m) 2.01 Figure 9.6 0.069 ×10 −4 0 0 0 0.2 0. with interaction effect by iteration method) .4 freq j frequency(rad/s) 0.6 0.876 4 Asp j 2 1.
436 ×10 −6 0 0 0 0.6 0.15 0.4 freq j frequency(rad/s) 0.8 1 1.8 1 1.191 Roll Amplitude (deg) 0.8 1 1.05 1.6 0.1 Asp j 0.01 Yaw Amplitude (deg) 1.851 ×10 −4 0 0 0 0.01 Pitch Amplitude (deg) 0.02 Asp j 0.438 2 Asp j 1 6.03 0.2 0.4 freq j frequency(rad/s) 0.6 0.106 0.2 0.021 0.9.d Continued .2 0.01 5.01 Figure 9.4 freq j frequency(rad/s) 0.16 ×10 −4 0 0 0 0.
tandem.41 1 0 Surge (m) surge1 i 1 − 1.595×10 3 ti time (sec) Figure 9.522 2 500 500 1000 1500 2000 2500 ti 3000 3500 4000 4500 5000 4.192 0.a Time simulation for the two body model of the FPSO and shuttle tanker (at body #1=FPSO.408 2 Sway (m) sway1 i 0 − 1.595 ×10 3 time (sec) 0.10. with interaction effect by combined method) .347 1 0 Heave (m) heave1i 1 − 1.487 2 500 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 4.595×10 3 time (sec) 1.256 2 500 500 1000 1500 2000 2500 ti 3000 3500 4000 4500 5000 4.
674 10 500 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 4.5 Roll (deg) roll1i 0 − 0.595×10 3 ti time (sec) 0.5 500 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 4.193 0.a Continued .938 1 Pitch (deg) pitch1 i 0 − 0.635 5 0 Yaw (deg) yaw1 i 5 − 7.429 0.595 10 × 3 ti time (sec) Figure 9.10.42 0.595 10 × 3 ti time (sec) 1.449 1 500 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 4.
137 0 500 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 4.595 10 × 3 ti time (sec) 3. tandem.934 5 500 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 4.027 5 500 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 4. with interaction effect by combined method) .656 10 5 Sway (m) sway2i 0 − 2.b Time simulation for the two body model of the FPSO and shuttle tanker (at body #2=shuttle tanker.257 30 20 Surge (m) surge2 i 10 8.10.535 5 Heave (m) heave2 i 0 − 1.194 24.595×10 3 ti time (sec) 9.595 10 × 3 ti time (sec) Figure 9.
b Continued .59510 × 3 time (sec) 14.5 Roll (deg) roll2 i 0 − 0.745 20 Yaw (deg) yaw2 10 i 0.195 0.356 2 500 500 1000 1500 2000 2500 ti 3000 3500 4000 4500 5000 4.336 0.10.59510 × 3 ti time (sec) 0.595 10 × 3 ti time (sec) Figure 9.5 500 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 4.797 1 0 Pitch (deg) pitch2 i 1 − 1.813 0 500 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 4.344 0.
2 Asp j 0.c Amplitude spectrum density curve of the motion responses for the two body model of the FPSO and shuttle tanker (at body #1=FPSO.6 0.6 freqj frequency(rad/s) 0.1 9.01 Sway Amplitude (m) 0.05 1. with interaction effect by combined method) .6 freqj frequency(rad/s) 0.05 3.196 Sureg Amplitude (m) 0.01 Figure 9.10.2 0.4 0.056 0. tandem.4 0.8 1 1.1 Asp j 0.2 0.095 0.8 1 1.2 0.1 Asp j 0.01 Heave Amplitude (m) 0.8 1 1.183 0.657×10 −6 0 0 0 0.336×10 −5 0 0 0 0.4 freq j frequency(rad/s) 0.389 ×10 −4 0 0 0 0.
02 2.197 Roll Amplitude (deg) 0.8 1 1.5 1.6 freqj frequency(rad/s) 0.c Continued .038 0.10.6 freq j frequency(rad/s) 0.01 Figure 9.8 1 1.2 0.01 Yaw Amplitude (deg) 0.4 0.033 0.507×10 −3 0 0 0 0.769×10 −5 0 0 0 0.04 Asp j 0.8 1 1.2 0.878 1 Asp j 0.02 4.6 freqj frequency(rad/s) 0.4 0.608×10 −5 0 0 0 0.01 Pitch Amplitude (deg) 0.04 Asp j 0.4 0.2 0.
tandem.01 Sway Amplitude (m) 1.2 0.2 0.8 1 1. with interaction effect by combined method) .038 1.01×10 −3 0 0 0 0.5 1 Asp j 0.d Amplitude spectrum density curve of the motion responses for the two body model of the FPSO and shuttle tanker (at body #2=shuttle tanker.438×10 −4 0 0 0 0.01 Figure 9.8 1 1.5 1.6 freq j frequency(rad/s) 0.4 Asp j 0.2 0.4 0.10.6 freqj frequency(rad/s) 0.4 0.94×10 −3 0 0 0 0.01 Heave Amplitude (m) 0.2 3.6 freq j frequency(rad/s) 0.4 0.8 1 1.284 0.198 Surge Amplitude (m) 2.599 3 2 Asp j 1 1.
6 0.6 freqj frequency(rad/s) 0.1 Asp j 0.4 0.6 freqj frequency(rad/s) 0.48×10 −6 0 0 0 0.15 0.8 1 1.342×10 −4 0 0 0 0.01 2.10.2 0.374 2 Aspj 1 3.2 0.589 10 × −4 0 0 0 0.016 0.199 Roll Amplitude (deg) 0.01 freqj frequency(rad/s) Figure 9.4 0.05 1.02 Asp j 0.8 1 1.8 1 1.2 0.01 Yaw Amplitude (deg) 1.4 0.111 0.d Continued .01 Pitch Amplitude (deg) 0.
The distance effects on motions and drift forces of two vessels are already reviewed in Chapter VII. The comparison of the analysis results for the FPSO and FPSO model and the massspring model has the validity of the program WINPOSTMULT. and the tandem mooring is considered. A simplification by the massspring model is also considered.200 9. The comparative study of an FPSO and a shuttle tanker illustrates the importance of including the interaction effect between multiple bodies. .7 Summary and Conclusions The hydrodynamic interaction effects and the hull/mooring/riser/hawser coupling for the multiple body system are investigated by numerical simulations. The coupling and interaction effects are studied using the twobody model of an FPSO and a shuttle tanker. An LNG FPSO and a shuttle tanker are taken as a multiple body system.
201 CHAPTER X CONCLUSIONS FOR ALL CASE STUDIES WINPOST program was developed for the hull/mooring/riser coupled dynamic analysis of floating structures. They are moored in a tandem arrangement and a sidebyside arrangement. there exist several uncertainties to be clarified. Despite the uncertainties mentioned. Thus. and FPSO. several platform parameters are not clearly identified. including hydrodynamic interactions. TLP. the trend of the numerical simulations follows that of experimental results. current force and the truncated mooring lines with buoys and springs may well not match with our numerical modeling. such as SPAR. The interaction effect is much stronger in the sidebyside mooring system than in the tandem mooring system. the motion RAOs double. The third case is to review the hydrodynamic characteristics of twobody interaction. In this study. Both mooring systems are considered for this study. In the second case. For example. For example. if the distance closes to a half of the original distance. the wind force. the program was extended to multiple body problems. . 5 case studies are presented for the verification of the developed program WINPOSTMULT. Even though the adjustment is made. In the OTRC experiment. the missing parameters are deduced from the free decay test. For the twobody model. the intermediate loading conditions and the simulated results are compared with OTRC experiment. an FPSO and a shuttle tanker are selected. The first two cases are for single FPSOs. The first one is a turretmoored FPSO in full load or ballast condition.
The spring is programmed to work in taut state. For the validity of this analysis. To verify the results. which includes the full hydrodynamic interactions. From the analysis results. either completely neglecting or partially including the hydrodynamic interaction effects. and an approximate solution is obtained. The WINPOSTMULT program is proved to be a useful tool for solving multiplebody interaction problems.e. The environmental loads are calculated in a simplified form to apply to the massspring model. . the conclusion is drawn that the interaction effects of the twobody problem can be very important. but not to work in slack state.. the connecting hawser is modeled as a spring. i. The spring stiffness is directly input in the system matrix in the program. When multiple floated dynamics are solved. FPSOs with and without hawsers and an FPSO and a shuttle tanker with and without hawser. several cases are considered. mooring lines and two FPSOs are modeled as a simple twomassspring system. These analyses are conducted for the tandem mooring system. the connecting hawser. The existing methods used in the industry are reviewed with the more sophisticated WINPOSTMULT program.202 The fourth case is for the twobody analysis with two identical SPARs. For the verification of the twobody module of the program WINPOSTMULT. The analysis results using the simplified massspring model and twospar model show a reasonable agreement with each other. a typical approach in offshore industry is one of them.
. 11th Int. OTC 12147 [CDROM]. 2001 Hull/Mooring/Riser Coupled Dynamic Analysis of a Deepwater Floating Platform with Polyester Lines. B. American Petroleum Institute. Dissertation. P. & de Wilde. 3.C. Texas. Conference. 2nd Edition. Designing. 303308. ISOPE. Baar.J. Offshore and Polar Eng. Ph.N. S. Texas A&M University.R. N.H.M. & Kim. 2000 Extreme Responses of Turret Moored Tankers. SNAME. van Dijk. T. 2002 An Analysis of Hydrodynamic Interaction of Floating MultiBody Using HigherOrder Boundary Element Method. J. 11th Offshore Symposium The Texas Section of the Society of Naval Architects and Marine Engineers. Proc. Arcandra. and Consulating Tension Leg Platforms.J. M.203 REFERENCES API RP 2T 1997 Recommended Practice for Planning. Offshore Technology Conference. T. Proc. 343353. Offshore and Polar Eng. & Hong. Proc. Washington D. 2001 Numerical MultipleBody Simulations of SidebySide Moored to an FPSO. Choi. A. Proc. C. 2002 Hull/Mooring/Riser Coupled Analysis of a TurretMoored FPSO 6000 ft: Comparison between Polyester and BuoysSteel Mooring Lines. 18.Y. 12th Int. Heyl. Nurtjahyo. Arcandra. G. & Rodenbusch.. Conference. Y.W. Buchner. J.. ISOPE.D. Houston. 1.
J. M. Y. 6571. Huijsmans. & Choi. Offshore and Polar Eng. 10th Int. Conference.M.E. ISOPE. of Energy Resources Technology.H. Dover.. Conference.J. Garrison. J.. 841847.A.R. J.R. Proc.G. 302307. Kim. Proc. Proc. 2001 Effect of Viscous Roll Damping on Drift Forces of MultiBody Floating System in Waves.L. O. Advanced Series on Ocean Engineering. 2000 An Efficient TimeDomain Analysis of Very Large MultiBody Floating Structures. 11th Int. R. 1992 Water Wave Mechanics for Engineers and Scientists. ISOPE. ISOPE. 12th Int.A. 1. Trans. Garrett. & Dalrymple. Hong. 2001 Diffraction and Radiation of Waves Around SidebySide Moored Vessels. Cambridge. of ASME. Y.Y. Conference. C.J. R. 1982 Dynamic Analysis of Slender Rods. ISOPE. 104. & de Wilde. The Cambridge University Press. D. 279285.. Offshore and Polar Eng. Pinkster. H. Kim. Inoue.M. & Islam. J. 1. 11th Int. 2. 3. Offshore and Polar Eng. World Scientific Press. Conference. . S.204 Dean. 2002 Experimental Study on Behavior of Tandem and Sideby Side Moored Vessels. D. J. 406412. 1. R. Faltinsen. 1998 Sea Loads on Ships and Offshore Structures. Offshore and Polar Eng. Proc.H.
301308. France. M. & Ran. M. Arcandra. 1994 Response of an Articulated Tower in Waves and Currents. 1989a The Complete SecondOrder Diffraction Solution for an Axisymmetric Body. Kim.H.K. & Zheng. Dept. 4. & Kim. 1999 Hull/Mooring/Riser Coupled Dynamic Analysis of a Truss Spar in TimeDomain. 465473. W. 557593.. ISOPE. & Yue. & Kim. 2001a Validability of Spar Motion Analysis against Various Design Methodologies/Parameters.H. of Ocean Engineering. 1. J. 235264. L. Offshore and Polar Eng. M.H. Texas A&M University. 1992 WINPOST V3. J. ISOPE. 20th Offshore Mechanics and Arctic Eng. M. Proc. Y. & Yue. Y. Part 2. 298231. Brest. M. Ran. 12th Int. Kim. Proc. (4). Califonia. 211.A. Offshore and Polar Eng. Bichromatic Incident Waves. M.H. of Fluid Mechanics.P.B. Kim.205 Kim. International Journal of Offshore and Polar Engineering.. Conference. Z. Conference. Arcandra.0 Users Manual. 3. D. Z. OMAE01OFT1063 [CDROM].H. D. 2001b Validability of TLP Motion Analysis against Various Design Methodologies/Parameters.K. 1989b The Complete SecondOrder Diffraction Solution for an Axisymmetric Body.B.P. Conference. Monochromatic Incident Waves. Kim. T. 9th Int. Part 1. T. 200. of Fluid Mechanics.. . Proc. Kim. Kim..H.H. M.
E. C. W. 11th Int. 3. 7.H. Nordgen. P. 1974 On Computation of the Motions of Elastic Rods. (2). & Kim. 5th OMAE Conf. Proc. ASME Journal of Applied Mechanics. J. M. 2001 Numerical Investigation of Nonlinear Wave Effects Around Multiple Cylinders. Houston. Proc. 777780. Cambridge.206 Lee. Zou. & Zheng. OTC 12085 [CDROM]. 3. 1997 Nonlinear Coupled Responses of a Tethered Spar Platform in Waves. Z. Proc. Teigen. J. W... & Huang. Ltd. OCIMF 1994 Prediction of Wind and Current Loads on VLCCs. (2).. Massachusetts Institute of Technology. International Journal of Offshore and Polar Engineering. 126133. England. Conference. London. Ran.Y. Tokyo. ISOPE. K. 2734. Pauling. 1999 Coupled Dynamic Analysis of a Moored Spar in Random Waves and Currents (TimeDomain versus FrequencyDomain Analysis). Journal of Offshore Mechanics and Arctic Engineering. Dept. Offshore Technology Conference.P. M. 2nd Edition. W.C.H. of Ocean Engineering. 1986 A Consistent LargeAmplitude Analysis of the Coupled Response of TLP and Tendon System.A.R. & Trulsen. Lee. Ma. Z. Ran. 194200.H. 2000 Deep Water Nonlinear Coupled Analysis Tool. Kim. 121. R. & Webster. Witherby & Co. 369378.. . M. M. Texas. 1999 WAMIT User Manual. Offshore and Polar Eng.
Dissertation.B. 4. Roelofs.W. ISOPE.M. 2001 The Behavior of a TankerBased FPSO in Hurricane Waves. The Netherlands.V...207 Ward. MARIN. The Netherlands. Offshore and Polar Eng. Wichers. 164171OB.E. The Netherlands. 156291OE. MARIN. Offshore and Polar Eng.E. Technical Rep. Wichers. Technical Rep. Delft. Wageningen. & Johnson.W. No. 1988 A Simulation Model for a Single Point Moored Tanker. 480487. Texas.W & Ji. Winds. C. No. and Currents. J. Conference. P. Proc. 2000a On the Coupling Term in the LowFrquency Viscous Reaction Forces of Moored Tankers in Deep Water.E. H. Wageningen. H. E.W.W. J. J. M.W. 3.P. J. P.Theme Structure Benchmark Analysis for Tanker Based FPSOGoM.G. 2001 DeepStarCTR 4401. Irani. .C. ISOPE. Wichers. & Ji. Proc. J. 11th Int. Wichers. Voogt. 11th Int.E. 650653. 2000b DeepStarCTR 4401. OTC 12086 [CDROM]. Proc. Delft University of Technology.J.Benchmark Model Test. R. Ph. Offshore Technology Conference.E. & Driessen. Houston.D. Conference. 2001 Effect of Coupling of Mooring Lines and Risers on the Design Values for a Turret Moored FPSO in Deep Water of the Gulf of Mexico. & Develin. Wichers. C.
After working for seven years for DWSH. (DWSH) on Keoje Island. 305810. Hayong and Harin. he entered the graduate school of Seoul National University in 1992. After he served in the Korean Army about for 10 months. vibration analyses and measurements for newly built ships. JeonminDong. He majored in naval architecture and ocean engineering. He married DeockSeung Seo in 1983 and has two sons. He was involved in ship design. There he worked as a structural engineer and also as a ship vibration analysis engineer. 1958. he received his Ph. After that. His permanent address is: 4596. While he worked at KR. he was employed by the Daewoo Ship Building and Heavy Industry Co. . he moved to the Korean Register of Shipping (KR) in Seoul. Korea.D. Taejon. Ltd.208 VITA YoungBok Kim was born in Incheon in the Republic of Korea on September 9. and two years later he received his Master of Science degree in February 1994. He graduated from Inha University with a Bachelor of Science degree in naval architecture and ocean engineering in February 1981. Chowon Villa 102. he went abroad to pursue the doctoral degree at Texas A&M University in January 1999. Korea. YusungKu. In May 2003. in the field of ocean engineering. Republic of Korea.