You are on page 1of 11

Proceedings of OMAE2004

23rd International Conference on Offshore Mechanics


and Arctic Engineering
20-25 June 2004, Vancouver, Canada

OMAE2004-51370

COUPLED MOORING ANALYSIS FOR A DEEP WATER CALM BUOY

Ir. J.L. Cozijn Dr. Ir. T.H.J. Bunnik


Maritime Research Institute Netherlands (MARIN) Maritime Research Institute Netherlands (MARIN)
P.O. Box 28 P.O. Box 28
6700 AA Wageningen 6700 AA Wageningen
The Netherlands The Netherlands
J.L.Cozijn@MARIN.nl T.Bunnik@MARIN.nl

ABSTRACT In this paper, a CALM buoy moored in a water depth of


The effect of the mooring loads on floater motions can be 1,200 m is considered. The purpose of this CALM buoy is to
significant for small water plane area floaters like CALM provide an offloading point for transferring oil from an FPSO,
buoys. Not only does the mooring system contribute to the which is moored at a distance of approximately 1 nautical mile,
static restoring force components, but the dynamic behaviour of to a shuttle tanker. The buoy is moored to the sea bed using a
the mooring lines also affects the inertia and damping of the semi-taut mooring system with mooring legs consisting of a top
moored CALM buoy. chain, a steel wire section and a bottom chain. The same
CALM buoy was also considered in Reference [2].
The results from model tests with a moored CALM buoy
were compared with the results from two series of time-domain The results of model tests carried out at a scale of 1:20 in
computer simulations. First, fully dynamic coupled simulations MARIN's Offshore Basin in July 2001 were made available by
were carried out, in which the interaction between the floater Bluewater Energy Services B.V. for comparison with
motions and the dynamic mooring line loads was modelled for simulation results. The test results considered here include
all 6 modes of motion. Second, quasi-static simulations were static load tests, motion decay tests and mooring tests in
carried out, in which only the (non-linear) static restoring force irregular waves. The wave conditions in the model tests were
characteristics of the mooring system were taken into account. relatively mild. The purpose of the tests was to investigate the
buoy first order motions, as well as the loads in the mooring
The comparison of results from the simulations and the lines and in the export risers suspended between the CALM
model tests clearly indicates that the fully dynamic coupled buoy and the FPSO. A relatively large buoy model was used, so
simulations show a much better correspondence with the model that the short wave periods considered could be accurately
test results than the quasi-static simulations. It is concluded that modelled in the basin. Another advantage of the large model
for the simulation of the behaviour of a moored CALM buoy in was that possible scale effects in the viscous hydrodynamic
waves a fully dynamic coupled mooring analysis is essential. loads on the buoy skirt were kept to a minimum.

INTRODUCTION With the selected model scale it was not possible to model
The dynamic behaviour of mooring lines has been the complete mooring system in the model test basin.
investigated in model tests and numerical simulations for many Therefore, instead of the full length mooring system and water
years. The coupling effects between the mooring line dynamics depth, a truncated mooring system was used in an equivalent
and the behaviour of the moored floater are in many cases not water depth of 208 m. This truncated mooring system was
negligible. For large water plane area floaters with a large designed such, that the behaviour of the full length mooring
displacement, like FPSOs, these coupling effects contribute system was represented as accurately as possible. For the same
only to the low frequency damping, generally without affecting reason, the FPSO was not included in the model tests. Instead,
the wave frequency motions. See e.g. Reference [1]. For other only part of the total length of the export risers was considered
floater types, like CALM buoys, semi-submersibles and spar in the tests. Anchor points for the export riser models were
buoys, however, the mooring line dynamics may have a placed at a fixed point on the basin wall. General considerations
significant effect on the wave frequency motions of the floater. for the model testing and analysis approach of deep water
floaters can be found in Reference [3].

1 Copyright © 2004 by ASME


The results of the model tests for the CALM buoy with the K = matrix of retardation functions kNs2/m, kNms2/rad
truncated mooring system are used to validate a numerical L = length of line element m
model of the moored CALM buoy. In this model combined low M = mass matrix tonnes, tonnes.m2
frequency and wave frequency motions of the floater are σ = standard deviation m, kN
described using linear equations of motion. The wave exciting t, τ = time s
forces on the CALM buoy and the linear hydrodynamic T0 = natural period s
reaction forces (added mass and damping) are based on the Tp = wave spectrum peak period s
results of frequency domain radiation and diffraction X = motion vector m, rad
calculations, which have been transferred to the time-domain. x& = velocity vector
Additional (non-linear) viscous damping coefficients for the m/s, rad/s
CALM buoy have also been included. The dynamic behaviour &&
x = acceleration vector m/s2, rad/s2
of the mooring lines, including hydrodynamic loads, is x& n =
described using non-linear equations of motion. A lumped mass normal velocity of line element m/s
method is used. The coupling between the floater motions and x& t = tangential velocity of line element m/s
the dynamic mooring line loads is taken into account in the
simulation model. MODEL TESTS FOR A CALM BUOY
WITH A TRUNCATED MOORING SYSTEM
The simulation model for the moored CALM buoy is the Model tests with the moored CALM buoy were carried out
same as the model used in Reference [2], except that some in MARIN's Offshore Basin at a model scale of 1:20 on behalf
improvements were made in the simulation input for the of Bluewater Energy Services B.V. in July 2001. Tests were
mooring system. The mooring system was specified in more carried out in irregular waves, without wind or current. The
detail, which resulted in a better correspondence between the water depth modelled in the basin was 208 m. A description of
model test and simulation results. Furthermore, the the Offshore Basin is given in Reference [4].
communication between floater model and dynamic mooring
line model in the simulation program has been improved. Description of the Test Set-up
The model of the CALM buoy was constructed of PVC
The content of this paper is as follows. First a description pipe and plate material and consisted of a cylindrical buoy body
of the model tests for the deepwater CALM buoy is given, as with a circular skirt. A centre well was not present. The buoy
well as an overview of the available model test results that are model was fitted with fairleads for connection of the mooring
used in this paper. Second, the time-domain simulation model lines and export risers. The main particulars and weight
for the moored CALM is presented. The fully dynamic coupled distribution of the CALM buoy model are given in Table 1.
simulation model, including all coupling effects between the Photographs of the buoy model are shown in Figures 1 and 2.
mooring lines and the floater, is described. A quasi-static
simulation model, modelling the mooring system only as a non- Based on the specified full length mooring system and
linear spring, is also discussed. Third, a comparison is made export risers, a truncated mooring system was designed for use
between the model test results and the results from simulations in the model tests. The design of the truncated mooring system
for the CALM buoy moored in the truncated mooring system. has already been discussed in Reference [2]. It was designed
Finally, some conclusions and recommendations are presented. such, that the load-displacement characteristics of the full
length mooring system were represented as accurately as
NOMENCLATURE possible by the truncated mooring system. Furthermore, the
mooring line and export riser pretensions and pretension angles
ρ = density of water tonnes/m3 were chosen such, that the correct CALM buoy draft and trim
A = added mass matrix tonnes, tonnes.m2 were achieved.
an = normal added mass tonnes
at = tangential added mass tonnes The resulting truncated mooring system consisted of 9
b(1) = linear damping coefficient kNs/m, kNms/rad mooring lines and 2 export risers. The mooring lines included 6
b(2) = quadratic damping coefficient kNs2/m2, kNms2/rad2 "short legs" (Line No. 1 - 3 and 7 - 9) and 3 "long legs" (Line
B = potential damping matrix kNs/m, kNms/rad No. 4 - 6). The mooring lines were manufactured of model
C = hydrostatic spring matrix kN/m, kNm/rad chain and steel wire sections, with a linear axial spring included
CDn = normal drag coefficient - to obtain the correct total axial stiffness. The export risers were
CDt = tangential drag coefficient - manufactured of sections of PVC pipe and steel wire. In the
CIn = normal inertia coefficient - export risers and the long legs clump weights were included to
CIt = tangential inertia coefficient - obtain the correct line pretensions at the fairleads. The
D = diameter of line element m properties of the truncated mooring lines and export risers can
EA = line axial stiffness kN be found in Tables 2 through 4. All lines are defined from
EI = line bending stiffness kNm2 anchor to fairlead. Figure 3 shows a top-view of the truncated
F = external force vector kN, kNm CALM buoy mooring system in the Offshore Basin. In Figure 4
Fi = nodal force vector kN, kNm a side-view of the mooring system is shown (excluding the
FDn = normal drag force kN "short" mooring legs).
FDt = tangential drag force kN
Hs = significant wave height m

2 Copyright © 2004 by ASME


Purpose of the Model Tests and a numerical model for the dynamic behaviour of the
The purpose of the model tests was to measure the motions mooring lines (at the bottom right corner). The interaction
of the moored CALM buoy, as well as the loads in the mooring between the two models is also shown.
lines and export risers, in irregular waves. The model test
results were to be used as input to the design of the CALM In the mathematical model of the CALM buoy the
buoy and for validation and calibration of the simulation model. hydrodynamic reaction loads, as well as the first and second
A range of wave heights and wave periods was considered. An order wave loads, are calculated based on the results of linear
overview of the irregular wave conditions is shown in Table 5. diffraction calculations. Additional viscous damping
contributions are modelled as external loads, using linear and
Available Model Test Results quadratic damping coefficients. The CALM buoy equations of
The model test results that are used in the present paper are motion are solved in the time-domain. The resulting buoy
listed below : motions, velocities and accelerations are used as input for the
calculation of the dynamic mooring line loads. The effects of
1. Static load tests were carried out to measure the load- mooring line dynamics and current loads (if any) on the
displacement characteristics of the truncated CALM buoy mooring lines are included in this calculation. The resulting
mooring system. In these tests the buoy was given a non-linear dynamic mooring line loads are used as external
number of horizontal displacements (in the x-direction), loads in the simulation of the CALM buoy motions. The
after which the resulting mooring loads were measured. numerical models for the CALM buoy and for the mooring line
2. Motion decay tests for the moored CALM buoy were dynamics each operate at their own time step. Communication
carried out for all 6 modes of motion. Natural periods and between both models takes place at a fixed time step of
still water damping values were determined from the typically 0.1 seconds.
decay test results. In these tests, the decaying oscillating
motions of the CALM buoy were recorded, after it had Equations of motion for the CALM buoy
been given an offset and was released. The CALM buoy motions are described using linearised
3. Mooring tests were carried out with the moored CALM equations of motion in the time domain. The wave exciting
buoy in various long crested irregular wave conditions. forces are calculated for the CALM buoy at its mean position
The test results included the buoy motions and mooring and the hydrodynamic reaction loads due to the buoy motions
loads. are calculated separately.

TIME-DOMAIN SIMULATIONS FOR THE CALM BUOY Prior to the actual time domain simulations for the moored
Two types of time-domain simulations were performed for CALM buoy, linear diffraction and radiation calculations are
the moored CALM buoy, being fully dynamic coupled carried out to determine the buoy hydrodynamic reaction loads
simulations and quasi-static simulations. Below, first a (added mass and damping), as well as the first and second order
description of the applied numerical model is given, which is wave exciting loads. The frequency domain results are
used for both types of simulations. Second, the differences in transformed to the time domain, using a retardation function
approach between the fully coupled and the quasi-static approach for the hydrodynamic reaction forces, see Reference
simulations are discussed. [5], and Fourier transforms for the wave exciting loads.

Description of the Numerical Model The following time domain equations of motion are used
The numerical model used in the time-domain simulations for the moored CALM buoy.
for the moored CALM buoy consists of two separate models,
which are linked as shown in the figure below. t

(M + A ) ⋅ &&x + ∫ K ( t-τ ) ⋅ x& (τ ) ⋅ dτ + C ⋅ x = F(t,X)


−∞

FLOATER In which M is the mass matrix, A is the frequency-


Waves independent added mass matrix, K(t) is the retardation function
Combined Low Frequency
and Wave Frequency Model matrix and C is the matrix with the hydrostatic spring terms.
Note that the mooring system restoring forces are not included
in the spring matrix C.

The retardation functions K(t) are the Fourier transform of


the frequency dependent damping coefficients (wave radiation).
Fmooring These are formulated as follows.
MOORING
. .. 2

x, x, x K (t) =
Dynamic
Mooring Model π ∫ B (ω ) ⋅ cos (ω t ) ⋅ dω
0

The two main components of the simulation model are a The load function F(t) at the right hand side of the equation
numerical model for the CALM buoy (at the upper left corner) of motion contains all loads on the CALM buoy, other than the
buoy inertia, the added mass, the wave radiation damping and

3 Copyright © 2004 by ASME


the hydrostatic loads. The load vector F(t) can be formulated as The system of equations of motion for the mooring line
follows. dynamics is formulated as follows.

F(t) = F(1)wave(t) + F(2)wave(t) + Fdamping(t) + Fmooring(t) (  A  + a (τ ) ) ⋅ &&x(τ ) = F (τ )


j j j

In which :
F(1)wave(t) = first order wave loads The inertia matrix A contains the mass properties of all
F(2)wave(t) = second order wave loads nodes, while the time-dependent inertia matrix aj(τ) contains
Fdamping(t) = viscous CALM buoy damping loads the added mass contributions. The nodal force vector Fj(τ)
Fmooring(t) = dynamic mooring loads contains the contributions due to line segment axial tensions,
buoyancy and weight, hydrodynamic loads and sea floor
Time records of the first and second order wave loads are reaction forces (if any).
calculated prior to the start of the time domain simulations. The
frequency domain results in combination with a wave spectrum Current loads and hydrodynamic reaction forces resulting
(with random phase model) or with a wave elevation record from the line motions (drag and inertia) are calculated using the
(e.g. undisturbed waves measured in the model basin during formulas below. The loads are applied as external forces to the
wave calibrations) are transformed to the time domain. discrete nodes.

Viscous damping loads are not included in the diffraction Normal drag force : FDn = −
1
ρCDnD L x& n x& n
and radiation calculation results, since these are potential theory 2
calculations. Therefore, the viscous damping contributions are Tangential drag force : FDt = 1
− ρCDt D L x& t x& t
modelled as external loads. During the time domain simulations 2
these are calculated using linear and quadratic damping an = ρ(CIn π 2
Normal added mass : − 1) D L
coefficients and the CALM buoy velocity. The following 4
formulation is used for each mode of motion of the buoy. at = π
Tangential added mass : ρ(CIt − 1) D 2L
4
Fdamping (t) = -b(1) ⋅ x& - b(2) ⋅ x& ⋅ x&
The drag and inertia coefficients used in the dynamic
mooring line simulations can be found in Table 7. In the
The values of the viscous damping coefficients have been calculation of the relative fluid velocities, the current velocity
selected based on the results of the motion decay tests. The (if any) and the velocity of the discrete nodes themselves are
values used in the present simulations are shown in Table 6. taken into account, but the wave orbital velocities are not.
The dynamic loads from the mooring lines and export The export risers in the present series of model tests were
risers are also modelled as external loads to the CALM buoy. not connected to the CALM buoy through a uni-joint, but
The motions, velocities and accelerations at the CALM buoy through a clamped connection. Furthermore, the bending
fairleads are used as input in the dynamic mooring line stiffness of the PVC riser section models was not negligible.
simulation model. The exact modeling of the mooring line The effect of the clamped connection and the bending stiffness
dynamic behaviour is described below. on the CALM buoy total pitch stiffness was calculated in an
approximate manner and included in the simulation model as an
Equations of motion for the mooring lines and export risers additional linear spring coefficient for pitch.
The fully dynamic three dimensional behaviour of the
mooring lines and export risers is simulated in the time domain
Fully Dynamic Coupled Simulations
using a lumped mass method, see Reference [6]. In this method In the fully coupled time domain mooring simulations for
each mooring line and export riser is discretised by the moored CALM buoy, the buoy motions and the mooring
concentrating the mass in a finite number of nodes along the line dynamics are modeled as described in the sections above.
line. A schematic example is shown in Figure 5. The CALM buoy motions and the dynamic behaviour of the
mooring lines are fully coupled. In other words, the buoy
The axial stiffness of the lines is modelled by springs motions are used as input for the dynamic mooring line
connecting the nodes. Bending stiffness is not modelled, since simulations, while the dynamic mooring line loads are used as
it is assumed that the behaviour of the mooring lines and export input for the simulation of the buoy motions. In this manner,
risers is governed by catenary effects. The lumped mass the effects of the mooring line dynamics on the first and second
approach results in a system of coupled equations of motion for order CALM buoy motions are taken into account. These
the discrete nodes. This system of coupled equations is solved effects can be summarised as follows :
in time with the CALM buoy motions at the fairleads
prescribed as boundary conditions. The results include the 1. Dynamic behaviour of the mooring lines
positions of all nodes in time, as well as the dynamic tension in 2. Damping contributions of the dynamic mooring line loads
the mooring lines and export risers. to the buoy motions
3. Inertia contributions of the dynamic mooring lines to the
buoy motions

4 Copyright © 2004 by ASME


Due to the dynamic behavior of the mooring lines the line together. From these motion signals the natural periods of the
loads cannot be determined by just using their static load- moored CALM buoy were derived, as well as the linear and
displacement characteristics, but are usually higher (dynamic quadratic damping coefficients. Their values have been
amplification). The drag loads on the mooring lines result in summarised in Table 8.
additional low frequency and wave frequency damping for the
moored CALM buoy. For moored FPSOs only the low The results for the CALM buoy surge decay show an
frequency damping contributions from the mooring system are accurate correspondence between the natural periods and
important, while the effect on the wave frequency damping is damping values from the model tests and the fully dynamic
negligible. The mass of the mooring lines, as well as the added simulation. The results of the quasi-static simulation, however,
mass, contribute to the total inertia of the moored buoy. For show a shorter natural period and a lower (quadratic) damping.
moored FPSOs this effect is negligible, due to the large mass of The CALM buoy mooring system contributes to the inertia and
the floater itself. Direct current loads on the mooring lines the damping of the surge motion. These effects are included in
mainly affect the mean displacement of the floater. the fully dynamic simulations, but not in the quasi-static
simulations. The inertia effects are a result of the mass of the
Quasi-static Simulations mooring lines and export risers, which move with the CALM
The simulation model for the quasi-static simulations is buoy, as well as the added mass. The damping effects are a
based on the fully dynamic coupled model, with the difference result of the drag loads on the mooring lines and export risers.
that in the quasi-static simulations only the restoring force
characteristics of the mooring lines and export risers are taken The results for the CALM buoy heave decay show the
into account. The line mass is not taken into account and same behaviour as was found in the surge decay results. The
neither are the drag and inertia loads on the lines. In this correspondence between the fully dynamic simulations and
manner, the mooring system acts as a (non-linear) spring only. model tests is accurate, while the quasi-static simulation results
In the quasi-static simulations the mooring lines do not show show a shorter natural period and less damping.
dynamic behaviour and neither do they experience any
hydrodynamic loads. A reason for using this approach could be The results for the CALM buoy pitch decay show again a
that less calculation power is required, because only the floater shorter natural period for the quasi-static simulation than in the
equations of motion are solved and not the system of coupled fully dynamic simulation. This is a result of the contribution of
equations of the mooring system lumped mass model. the mooring lines and export risers to the pitch inertia. A
comparison with the model test results, however, reveals that
COMPARISSON OF THE MODEL TEST AND the natural pitch period is too long in the fully dynamic coupled
SIMULATION RESULTS simulations. The reason for this may be that the pitch added
In the following sections the results of the fully dynamic mass is overestimated in the simulations, because the
coupled simulations and the quasi-static simulations are perforations in the CALM buoy model skirt were not modelled
compared with the results of the model tests. Comparisons are in the diffraction calculations. Therefore, the pitch added mass
made for the mooring system load-displacement characteristics, of the CALM buoy in the model tests may have been smaller
the motion decay response and the behaviour of the CALM than in the simulations. The pitch damping values from the
buoy in irregular waves. model test and both simulations appear to be similar.
Apparently, the contribution from the mooring system to the
Static Load-Displacement Characteristics total pitch damping is relatively small.
The results of the static load tests and calculations for the
moored CALM buoy with export risers are shown in Figure 6. Moored CALM Buoy in Irregular Seas
The load-displacement characteristics for displacements in the The results from the model tests and simulations for the
x-direction are shown. The calculated results (line) and the moored CALM buoy in irregular waves are summarised in
results measured in the basin (points) are plotted together. The Tables 9 through 12 and Figures 9 through 19. In general, the
calculated static (non-linear) load-displacement characteristics results from the fully dynamic coupled simulations clearly
for the fully dynamic simulations and the quasi-static show a more accurate correspondence with the results from
simulations are identical, because only the static behavior of the model test results than the quasi-static simulation results.
mooring system is considered here. The results show that the
correspondence between the calculated and measured results is CALM Buoy Motions
very accurate. This means that the static restoring force In Table 9 a summary of the CALM buoy motion statistics
characteristics of the mooring system are correctly represented is given. The motion standard deviations shown here are
in the simulations. considered to be a measure for the buoy motion amplitudes in
irregular waves. In Figures 10 through 13 some samples are
Natural Periods and Still Water Damping shown of plotted time traces of the moored CALM buoy surge,
In Figures 7 through 9 the results of the motion decay tests heave and pitch motions. In Figures 17 through 19 some
and simulations for the moored CALM buoy with export risers examples of the CALM buoy surge, heave and pitch motion
are shown. The recorded motion signals from the surge, heave RAOs are shown. In all these plots, the results from the fully
and pitch motion decay tests and simulations are shown. In dynamic simulation (dashed lines), the quasi-static simulations
each graph the motion signals from the fully dynamic (dash-dotted lines) and the model tests (solid lines) are plotted
simulations (dashed line), the quasi-static simulations (dash- together.
dotted line) and the model tests (solid line) have been plotted

5 Copyright © 2004 by ASME


The statistical results show a close correspondence The statistics and plotted time traces show that short leg 1
between the CALM buoy surge motions in the dynamic does not show a large variation in line tensions. Apparently, the
simulations and the model tests, except for the cases with the line does not show strong dynamic behaviour in this particular
shortest wave period of 5 s. The quasi-static simulations condition. As a consequence, the differences between the fully
generally over predict the surge motions. These observations dynamic simulation, the quasi-static simulation and the model
are confirmed by the surge motion RAOs shown in Figure 17. test results are small. Good correspondence is achieved for both
The plotted time-traces, however, show a more complex types of simulations.
situation. In Figure 10 the CALM buoy surge motions are
shown. At first sight, the correspondence between simulations For long leg 5 and the upper export riser the results from
and model tests is not accurate. In Figure 11, however, only the the model tests show much larger variations in the line loads,
wave frequency content of the surge signals is plotted. Here, due to the dynamic behaviour of the lines. The plotted time
the correspondence is very accurate for the dynamic traces and the statistical results show that in the fully dynamic
simulations and reasonably accurate for the quasi-static simulations a good correspondence with the model test results
simulations. The above observation indicates that the CALM is achieved, except for a small difference in mean value. The
buoy wave frequency surge motions are accurately predicted, quasi-static simulations, however, strongly under predict the
but its low frequency surge motions are not. This may be variations in the line loads. Here, only the load variation due to
caused by inaccuracies in the low frequency wave drift forces restoring force characteristics of the lines is included, while the
for surge. These second order loads depend on the first order mooring line dynamic behaviour is not included. These results
vessel motions. It is shown below that the pitch motions are not indicate, that the load variations in the mooring lines and export
accurately represented in the simulations. This may in turn risers can only be accurately predicted through computer
affect the accuracy of the wave drift forces and thus the low simulations if a fully dynamic coupled simulation model is
frequency surge motions of the moored CALM buoy. used. In quasi-static simulations, the tension variations are
under predicted.
The CALM buoy heave motion statistics show a good
agreement between model tests and dynamic simulations. The CONCLUSIONS AND RECOMMENDATIONS
quasi-static simulations generally under predict the heave Based on the comparisons made between the model test
motions. This can also be observed in the sample plotted time results, the fully dynamic coupled simulations and the quasi-
trace in Figure 12. In addition, the heave motion RAO in Figure static simulations, the following conclusions may be drawn.
18, shows that the correspondence between the dynamic
simulations and the model tests is accurate. The RAO from the 1. The load-displacement characteristics of the CALM buoy
quasi-static simulations, however, shows a peak response at a mooring lines and export risers are identical in the fully
shorter wave period. This can be explained by the difference in dynamic and in the quasi-static simulations. An accurate
natural heave period, as discussed before. correspondence with model test results is observed.
2. The natural periods and damping values of the moored
The pitch motions are over predicted by both the dynamic CALM buoy are affected by the mooring line dynamics.
simulations and the quasi-static simulations. This can be In the fully dynamic coupled simulations the mooring
observed in the pitch statistics, the plotted time traces and the system contributions to the system inertia (through the
pitch motion RAOs. The over prediction may be a result of the mass of the mooring legs and export risers) and the system
observed difference in natural period, but may also be caused damping (through the drag loads on the mooring lines and
by non-linearities in the wave exciting pitch moment on the export risers) are taken into account.
CALM buoy, as discussed in Reference [2]. Viscous and non- 3. In general, the results from the fully dynamic coupled
linear components in the pitch wave exciting moment are not simulations for the moored CALM buoy clearly show a
taken into account in the linear diffraction calculations made better correspondence with the model test results than the
for the CALM buoy. Also possible contributions from the wave results from the quasi-static simulations.
orbital motions on the mooring line and export riser loads are 4. The (wave frequency) surge and heave motions of the
not considered here. In order to be able to improve the accuracy moored CALM buoy in irregular waves are accurately
of the pitch motions in the simulation results, further research predicted in the fully dynamic simulations. The
to improve the modelling of the exciting pitch moments on the correspondence between model tests and simulation
CALM buoy is necessary. results is less accurate for the quasi-static simulations.
5. The pitch motions in irregular waves are over predicted in
Mooring Line and Export Riser Loads the quasi-static simulations and in the fully dynamic
In Tables 10 through 12 a summary of the CALM buoy simulations. This may be the result of differences in the
mooring load statistics is given. The mean values, standard pitch added mass and non-linear viscous effects in the
deviation and maximum values of the tensions in "short leg 1" wave exciting pitch moment.
and "long leg 5" are shown, as well as the longitudinal load FX 6. The low frequency surge motions are not accurately
of the "upper export riser". In Figures 13 through 15 some represented in the CALM buoy simulations. This may be
samples are shown of plotted time traces of the CALM buoy caused by inaccuracies in the surge wave drift forces.
line loads at the fairleads. In these plots, the results from the 7. The tension variations in the mooring lines and export
fully dynamic simulation (dashed lines), the quasi-static risers of a moored CALM buoy in irregular waves can
simulations (dash-dotted lines) and the model tests (solid lines) only be accurately predicted through time-domain
are again plotted together. computer simulations if a fully dynamic coupled
simulation model is used.

6 Copyright © 2004 by ASME


8. Within the range of wave conditions considered, the
accuracy of the simulation results for the moored CALM Table 5 - Review of tests in irregular waves
buoy in irregular waves appears to be independent of the Hs Tp Wave direction
wave height and period. An exception is made for the Test No.
[m] [s] [deg]
shortest wave period considered (5s), for which the results 1 0.75 5 180
were less accurate. 2 0.75 7 180
3 0.75 10 180
Finally, some possibilities for further improvement of the 4 1.25 5 180
simulation model are suggested. First of all, the pitch added 5 1.25 7 180
mass and the wave exciting pitch moment of the CALM buoy
6 1.25 10 180
are difficult to predict accurately through linear diffraction and
7 1.25 14 180
radiation calculations. An improvement in the accuracy of the
pitch exciting moment could improve the accuracy of the 8 2.25 10 180
CALM buoy pitch motions in the time-domain simulations. 9 2.25 14 180
Second, the low frequency motions of the moored CALM buoy
are not accurately predicted. Possibly this is the result of Table 6 - Applied linear and quadratic CALM buoy damping
inaccuracies in the surge wave drift forces. Further research is Mode of motion b(1) b(2)
required to determine whether this is the case. Surge 5 120
Sway 5 120
TABLES AND FIGURES Heave 13.9 488
Roll 0 9.5E5
Table 1 - Main particulars of the CALM buoy Pitch 0 9.5E5
Particular Unit Value Yaw 0 1.0E6
Diameter buoy body m 20
Diameter buoy skirt m 24 Table 7 - Applied line inertia and drag coefficients
Position skirt above keel m 1.25 Line type CDn CDt CIn CIt
Mass tons 1,299 chain 2.4 0.8 3.1 1.7
Draft m 5.9 riser 1.3 0.2 2.0 1.2
Pitch radius of inertia kyy m 5.7 steel wire 1.3 0.4 1.6 1.2
Vertical position CoG above keel m 5.1
Longitudinal position CoG m 0.96 Table 8 - Summary of CALM buoy motion decay results
Test / Simulation Motion T0 b(1) b(2)
Table 2 - Properties of the truncated mooring lines (short legs) Model test 76 7.2 229
Particular Unit Value Coupled simulation Surge 74 9.6 217
Length (wire - chain) m 174 / 77 Quasi-static simulation 66 4.2 122
Diameter m 0.073 / 0.102 Model test 7.0 440 493
Axial stiffness EA kN 1,378 / 8.7E5 Coupled simulation Heave 7.0 480 313
Under water weight N 217 / 1,775 Quasi-static simulation 6.5 466 154
Model test 5.9 5.6E3 1.1E6
Table 3 - Properties of the truncated mooring lines (long legs) Coupled simulation Pitch 6.7 5.6E3 8.0E5
Particular Unit Value Quasi-static simulation 5.9 1.7E3 8.0E5
Length (wire-wire-wire-chain) * m 327/140/141/80
Table 9 - CALM buoy motion statistics in irregular waves
Diameter m 0.02/0.02/0.073/0.102
Axial stiffness EA kN 1.3E6/1.3E6/4.7E5/8.7E5 Test Result σx [m] σz [m] σθ [m]
Under water weight N/m 21 / 21 / 217 / 1,775 Model test 0.34 0.06 0.16
* Clump weight properties are not included in above table 1 Dynamic 0.19 0.06 0.31
Quasi-static 0.31 0.07 0.47
Table 4 - Properties of the truncated export risers Model test 0.33 0.24 0.17
Particular Unit Value 2 Dynamic 0.35 0.24 0.46
224 / 177 / 200 (up) Quasi-static 0.39 0.22 0.62
Length (wire - wire - riser) * m Model test 0.20 0.22 0.16
224 / 220 / 200 (low)
Diameter m 0.02 / 0.02 / 0.50 3 Dynamic 0.20 0.21 0.25
Axial stiffness EA kN 1.3E6 / 1.3E6 / 4.4E6 Quasi-static 0.25 0.18 0.41
Bending stiffness EI kNm2 ---- / ---- / 1.1E5 Model test 0.72 0.09 0.25
Under water weight N/m 21 / 21 / 1,854 4 Dynamic 0.42 0.09 0.48
63.0 (upper) Quasi-static 0.65 0.11 0.64
Top angle w.r.t. horizontal deg Model test 0.68 0.34 0.23
68.5 (lower)
* Clump weight properties are not included in above table 5 Dynamic 0.71 0.34 0.52
Quasi-static 0.85 0.33 0.84

7 Copyright © 2004 by ASME


Model test 0.38 0.37 0.18 5 Model test 771 15.1 818
6 Dynamic 0.39 0.35 0.33 Dynamic 800 17.8 858
Quasi-static 0.48 0.31 0.55 Quasi-static 801 0.8 803
Model test 0.36 0.34 0.15 Model test 771 13.0 819
7 Dynamic 0.36 0.32 0.26 6 Dynamic 801 14.0 849
Quasi-static 0.41 0.30 0.41 Quasi-static 801 0.5 803
Model test 0.81 0.64 0.26 Model test 771 9.3 809
8 Dynamic 0.83 0.58 0.44 7 Dynamic 801 9.9 833
Quasi-static 1.03 0.52 0.77 Quasi-static 801 0.5 803
Model test 0.69 0.62 0.19 Model test 771 22.6 855
9 Dynamic 0.69 0.58 0.33 8 Dynamic 800 21.9 871
Quasi-static 0.78 0.54 0.58 Quasi-static 801 1.1 804
Model test 771 15.7 833
Table 10 - Line tension (short leg 1) statistics in irregular waves 9 Dynamic 801 15.4 849
Test Result Mean [kN] σ [kN] Max. [kN] Quasi-static 801 0.9 804
Model test 290 1.3 294
1 Dynamic 297 1.3 302 Table 12 - Line load FX (upper export riser) statistics in waves
Quasi-static 297 0.8 300 Test Result Mean [kN] σ [kN] Max. [kN]
Model test 290 2.5 300 Model test 429 10.3 465
2 Dynamic 297 2.7 308 1 Dynamic 460 6.0 483
Quasi-static 297 1.6 305 Quasi-static 460 0.2 461
Model test 289 1.6 296 Model test 429 13.7 476
3 Dynamic 297 1.4 304 2 Dynamic 460 14.2 509
Quasi-static 297 1.1 302 Quasi-static 460 0.2 461
Model test 291 2.3 303 Model test 429 7.7 459
4 Dynamic 298 2.2 308 3 Dynamic 460 8.4 495
Quasi-static 298 1.6 306 Quasi-static 460 0.2 461
Model test 290 3.9 306 Model test 429 15.0 485
5 Dynamic 298 4.2 316 4 Dynamic 460 9.8 497
Quasi-static 298 2.8 312 Quasi-static 460 0.3 462
Model test 289 2.5 299 Model test 429 21.4 498
6 Dynamic 297 2.2 309 5 Dynamic 460 21.6 526
Quasi-static 297 2.0 307 Quasi-static 460 0.5 463
Model test 289 1.6 295 Model test 428 12.4 478
7 Dynamic 297 1.4 303 6 Dynamic 460 13.6 514
Quasi-static 297 1.8 303 Quasi-static 460 0.3 462
Model test 290 4.4 308 Model test 428 8.2 457
8 Dynamic 298 4.2 323 7 Dynamic 460 9.2 491
Quasi-static 298 3.7 319 Quasi-static 460 0.3 461
Model test 290 2.8 302 Model test 428 22.0 514
9 Dynamic 297 2.6 306 8 Dynamic 460 23.6 545
Quasi-static 297 3.3 310 Quasi-static 460 0.6 464
Model test 428 14.4 481
Table 11 - Line tension (long leg 5) statistics in irregular waves 9 Dynamic 460 15.8 511
Test Result Mean [kN] σ [kN] Max. [kN] Quasi-static 460 0.5 462
Model test 771 4.2 787
1 Dynamic 801 8.6 842
Quasi-static 801 0.3 802
Model test 771 11.1 803
2 Dynamic 801 12.4 848
Quasi-static 801 0.4 802
Model test 771 8.3 802
3 Dynamic 801 9.3 834
Quasi-static 801 0.3 802
Model test 771 5.9 792
4 Dynamic 801 12.9 866
Quasi-static 801 0.6 803

8 Copyright © 2004 by ASME


Figure 1 - Photograph of the CALM buoy model Figure 5 - Example of mooring line lumped mass method

Figure 2 - Photograph of the CALM buoy model in the basin


Figure 6 - Static load-displacement characteristics
of the CALM buoy truncated mooring system
1,500

computed
1,000 measured

Total Mooring Force FX [kN]


500

0
-50 -40 -30 -20 -10 0 10 20 30 40 50

-500

-1,000

-1,500
Buoy Displacement X [m]

Figure 3 - Top view of the truncated mooring system


Figure 7 - CALM buoy surge decay plotted time traces
short legs 10

CALM buoy 8

4
Surge Motion [m]

0
long legs 0 100 200 300 400 500
-2
export risers
-4 Model Test
-6 Fully Dynamic Simulation
short legs -8 Quasi-static Simulation

-10
Time [s]

Figure 4 - Side view of the truncated mooring system


(only long legs 4 / 5 / 6 and export risers are shown)

export risers long legs 4, 5, 6

9 Copyright © 2004 by ASME


Figure 8 - CALM buoy heave decay plotted time traces Figure 12 - Sample plot of heave motions from Test No. 5
1.00 Test No. 5 - Hs = 1.75 m, Tp = 7.0 s
1.50
0.75 Model Test
Fully Dynamic Simulation
0.50 1.00
Quasi-static Simulation
Heave Motion [m]

Heave Motion [m]


0.25
0.50
0.00
0 10 20 30 40 50
-0.25 0.00
1675 1700 1725 1750
-0.50 Model Test
Fully Dynamic Simulation -0.50
-0.75
Quasi-static Simulation

-1.00 -1.00
Time [s] Time [s]

Figure 9 - CALM buoy pitch decay plotted time traces Figure 13 - Sample plot of pitch motions from Test No. 5
0.10 Test No. 5 - Hs = 1.75 m, Tp = 7.0 s
3.00
0.08 Model Test
Fully Dynamic Simulation
0.05 2.00 Quasi-static Simulation
Pitch Motion [deg]

Pitch Motion [deg]


0.03
1.00
0.00
0 10 20 30
-0.03 0.00
1700 1725 1750 1775
-0.05 Model Test
Fully Dynamic Simulation -1.00
-0.08 Quasi-static Simulation

-0.10 -2.00
Time [s] Time [s]

Figure 10 - Sample plot of surge motions from Test No. 5


Test No. 5 - Hs = 1.75 m, Tp = 7.0 s Figure 14 - Sample plot line loads (short leg 1) from Test No. 5
1.50
Model Test Test No. 5 - Hs = 1.75 m, Tp = 7.0 s
500
1.00 Fully Dynamic Simulation Model Test
Quasi-static Simulation Fully Dynamic Simulation
0.50 400 Quasi-static Simulation
Surge Motion [m]

Load Short Leg 1 [kN]

0.00
300
1450 1475 1500 1525 1550
-0.50
200
-1.00

-1.50 100

-2.00 0
Time [s]
1450 1475 1500 1525 1550
Time [s]

Figure 11 - Wave frequency surge motions from Test No. 5


Test No. 5 - Hs = 1.75 m, Tp = 7.0 s Figure 15 - Sample plot line loads (long leg 5) from Test No. 5
0.75
Test No. 5 - Hs = 1.75 m, Tp = 7.0 s
1000
0.50
WF Surge Motions [m]

800
Load Long Leg 5 [kN]

0.25

600
0.00
1450 1475 1500 1525 1550
-0.25 400
Model Test
-0.50 Model Test 200 Fully Dynamic Simulation
Fully Dynamic Simulation Quasi-static Simulation
Quasi-static Simulation
-0.75 0
Time [s]
1450 1475 1500 1525 1550
Time [s]

10 Copyright © 2004 by ASME


Figure 16 - Sample plot export riser loads (upper) Test No. 5 ACKNOWLEDGEMENTS
Test No. 5 - Hs = 1.75 m, Tp = 7.0 s We would like to thank Bluewater Energy Services B.V.
750
for their kind permission to use the results from the large scale
model tests with the moored CALM buoy, which were carried
Upper Export Riser FX [kN]

out on their behalf in July 2001, as validation material for our


500 simulations. Without these model test results the authors could
not have made the comparison between model tests and
simulation results as presented in this paper.
250 Model Test
Fully Dynamic Simulation
REFERENCES
Quasi-static Simulation
[1] Dercksen, A. and Wichers, J.E.W.: "A Discrete Element
0
Method on a Chain Turret Tanker Exposed to Survival
1525 1550 1575 1600
Time [s] Conditions", BOSS Conference 1992.

Figure 17 - Surge motion RAO from Test No. 5 [2] Bunnik, T.H.J., de Boer, G., Cozijn, J. L., van der
Cammen, J., van Haaften, E. and ter Brake E.: “Coupled
1.50
Model Test Mooring Analysis in Large Scale Model Tests on a
Dynamic
Deepwater CALM Buoy in Mild Wave Conditions”,
OMAE28056, 2002.
Surge Motion RAO [m/m]

Quasi-static
1.00
[3] Buchner, B.: “Numerical Simulation and Model Test
Requirements for Deep Water Developments”,
Symposium on Deep and Ultra Deep Water Offshore
0.50
Technology, New Castle 1999.

[4] Buchner, B., Wichers, J.E.W. and de Wilde, J.J.:


0.00 "Features of the State-of-the-art Deepwater Offshore
0.50 0.75 1.00 1.25 1.50 Basin", OTC10841, 1999.
Wave frequency [rad/s]

[5] Ogilvie, T.F.: "Recent Progress towards the


Figure 18 - Heave motion RAO from Test No. 5 understanding and prediction of ship motions".
2.00 Proceedings of the 5th Symposium on Naval
Model Test Hydrodynamics, 1964.
Heave Motion RAO [m/m]

1.50 Dynamic
[6] van den Boom, H.J.J., “Dynamic Behaviour of Mooring
Quasi-static
Lines”, BOSS Conference, 1985.
1.00

0.50

0.00
0.50 0.75 1.00 1.25 1.50
Wave frequency [rad/s]

Figure 19 - Pitch motion RAO from Test No. 5


5.00
Model Test

4.00 Dynamic
Pitch Motion RAO [deg/m]

Quasi-static
3.00

2.00

1.00

0.00
0.50 0.75 1.00 1.25 1.50
Wave frequency [rad/s]

11 Copyright © 2004 by ASME