Proceedings of the 26th International Conference on Offshore Mechanics Proceedings of OMAE2007

and Arctic Engineering

26th International Conference on Offshore Mechanics and Arctic OMAE2007
Engineering
June10-15,
June 10-15,2007,
2007,San
SanDiego,
Diego,California,
California,USA
USA

OMAE2007-29756
OMAE2007-29756

TIME DOMAIN SIMULATIONS OF SIDE-BY-SIDE MOORED VESSELS

LESSONS LEARNT FROM A BENCHMARK TEST

Mamoun Naciri
Single Buoy Moorings Inc.
24 avenue de Fontvieille PO Box 199
MC98007 Monaco Cedex
Email: mamoun.naciri@singlebuoy.com

Olaf Waals Jaap de Wilde

MARIN Wageningen MARIN Wageningen
The Netherlands The Netherlands
o.waals@marin.nl j.dewilde@marin.nl
Key words :
Side-by-side mooring, LNG, Offloading, Time domain simulations, Downtime assessment, Multiple-body
diffraction.

ABSTRACT and used at SBM, LIFSIM and aNySIM both

Time domain simulations are used extensively to
assess the availability of turret moored terminals for
offloading operations. The availability analysis
provides a key input when assessing the economical
viability of a project. In this context, it was deemed
important to perform a benchmark study. Three
programs have been used: AQWA-DRIFT an
ANSYS product maintained by Century Dynamics
developed in MARIN. Five environments including
wind, waves and current have been selected one of
which with all components in-line and the remaining
four with oblique wind and current. Simulations
have been performed with identical input for the
three programs. Results are compared statistically,
but also in the time domain. Lessons learned will be
discussed and areas of improvement identified.

Figure 1.1- Overview of new offshore LNG offloading systems.

1 Copyright 2007 by ASME

 Figure 1.2(a-b)- Overview of new offshore LNG offloading systems. (a) SQM system (b) FSRU.

1. INTRODUCTION 2. SIDE-BY-SIDE LNG OFFLOADING

FROM FSRU
Liquefied Natural Gas (LNG) is becoming an
increasingly important energy source. Due to the
cryogenic nature of LNG, the fluid transfer between 2.1. FSRU floating terminal
offloading LNG carrier and the FSRU (Floating
Storage and Regasification Unit) cannot yet be
carried out with floating hoses. For LNG carriers on A Floating Storage and Regasification Unit (FSRU)
jetties, special LNG loading arms are in use, but they is a new promising concept for offshore LNG import.
have so far not been applied on floating structures Locations are envisaged in the Gulf of Mexico (see
and their allowable motion envelope is limited. This Ye et al (2005)), US East Coast, California or the US
makes the subject of offshore LNG offloading one of West Coast. The FSRU will receive regular
the most challenging in the offshore LNG utilisation. shipments of LNG from LNG carriers, transfer it to
For floating LNG offloading, there are basically two the onboard storage tanks and send it to shore in
concepts under investigation: side-by-side mooring gaseous phase through a sub sea pipeline. The FSRU
and tandem (stern-to-bow) mooring. In the side-by- will be permanently moored on station, free to
side situation the use of existing LNG loading arms is weathervane about a single point mooring (SPM),
possible, although they should be evaluated for use allowing continuous operation throughout the year
on a moving structure. Relative motions and mooring and survival during extreme storm conditions.
loads are the main aspects that need to be considered Marinized loading arms will be used for loading the
in the evaluation of the concepts. Other novel LNG from the alongside moored LNG carrier to the
concepts, such as soft yoke systems have been FSRU. A side-by-side mooring configuration is
developed in recent years for harsh environments; see assumed, consisting of 16 to 20 mooring lines and 4
Naciri et al (2003). Gravity Based Structures (GBS) to 6 fenders suspended from the FSRU and located
have also been considered as LNG receiving near the waterline. An artist impression of an FSRU
terminals. An overview of several concepts is given is shown in Figure 1.2b.
in Figure 1.1 and 1.2.
A barge type FSRU is considered, with main
More recently, the Soft Quay Mooring system (SQM) particulars as presented in Table 2.1. The calculated
has been developed to enhance FSRU terminal natural roll period at 11.0 m draft is 15.7 seconds.
availability. The principle as shown in Figure 1.2a Designation FSRU LNGC
above is to ease the delicate fender kissing phase by Length between perpendiculars
323.48 274.00
providing a lateral yoke-mounted fender system. [m]
Breadth [m] 59.4 44.20
The purpose of this paper is to focus on conventional
Depth [m] 32.4 25.00
side-by-side offloading of an LNG carrier to a turret
moored FSRU. This is described in some detail in 2. Draft (even keel) [m] 11.0 11.00
The numerical modelling of side-by-side offloading Displacement weight [ton] 211,799 99,390
is covered in 3. Results obtained with the three COG above baseline [m] 21.8 16.30
software tools for one environment are compared in
Transverse metacentric height
4, before drawing conclusions in 5. 11.3 4.8
[m]
Roll radius of gyration [m] 21.0 15.15
Roll period [s] 15.7 16.1
Table 2.1- Main particulars of FSRU and LNGC.

2 Copyright 2007 by ASME

2.2. LNG Carrier
40

FSRU 20
A 135,000 m3 storage capacity LNG tanker is
0
considered whose main particulars are listed above. -200 -150 -100 -50 0 50 100 150 200
-20
The natural roll period of the loaded tanker at 11.0 m
-40
draft is 16.1 seconds. The two vessels have almost 4-6 7,8 9,10
11-13
the same roll period. 1-3 LNGC -60
14-16
-80

-100

2.3. FSRU turret mooring

Figure 2.1 - Side-by-side mooring arrangement.

The FSRU is turret moored on 4 bundles of 4 lines The four pneumatic fenders of 4.5 m diameter have a
each in 60m water depth. The internal turret is rated load of 4,532 kN at 55% compression. The safe
located 54.8 m from the fore perpendicular. The working load of the fenders is assumed to be
horizontal pretension is 638.5 kN with a pretension 2,492kN.
angle of 37.9 w.r.t. the horizontal. The anchor
fairlead distance is 837.5m. The restoring at 10 m
horizontal excursion is about 20,000 kN. The FSRU 2.6. Modal analysis
will be equipped with a thruster at its stern for
heading control. The thruster is modelled by a
constant transverse force in the FSRU local axis The side-by-side moored vessels have in principle six
system. low frequency natural modes in the horizontal plane,
involving in-phase and out-of-phase combinations of
surge, sway and yaw of each vessel. A modal
2.4. LNG loading arms
analysis was performed with the dynamic module of
AQWA-LIBRIUM. The results obtained in calm
Special marinized LNG loading arms are being water are presented in Table 2.2 below:
developed for use on floating terminals. Typically
Designation Per. Freq. Damping
four 16-inch loading arms are required, of which
[s] [rad/s] [%]
three will be used for loading the cryogenic LNG and In-phase yaw & sway 175 0.036 12.0
one for vapor return. For jetty mooring, special LNG In-phase surge 140 0.045 4.8
loading arms are in use, but they have so far not been In-phase yaw & out-of-phase 92 0.068 8.9
applied on floating structures and their allowable surge
Out-of-phase surge 62 0.102 2.9
motion envelope is limited. Demonstrated values for
Out-of-phase sway, surge & yaw 41 0.152 2.4
the motion envelope of the loading arm have been Out-of-phase yaw & sway 35 0.181 2.8
presented by van der Valk (2004).
- Low frequency horizontal rel. motions: +/-4.0 m Table 2.2 - Natural modes and damping values.
- Wave frequency horizontal rel. motions: +/-2.0 m
- Vertical relative motions: +/-2.0 m It is noted that the in-phase horizontal modes tend to
be at low frequencies and are relatively well damped.
The out-of-phase horizontal modes are, in contrast,
However, manufactures claim that larger operating
weakly damped (less than 3% of critical) and
envelopes are technically feasible.
correspond to rather short periods down to 35s.
Mooring lines connecting the LNGC to the FSRU are
2.5. Side-by-side mooring arrangement more sensitive to the out-of-phase modes. It is
noteworthy that horizontal resonant modes can have
The side-by-side mooring system consists of 16 such a low natural period (35s) compared to the
identical lines and 4 identical pneumantic fenders. natural frequency of moored single hulls. The above
Each steel wire mooring line is fitted with a 25m table shows that difference frequency wave loads
Nylon tail. The mooring lines are loaded at 15 tons in should be computed with a bandwidth of at least
the neutral position in calm water. The safe working 0.2rad/s if sources of low frequency wave excitation
load (SWL) of the 44 mm steel wires is assumed to are not to be omitted.
be 681 kN (55% of MBL). The 16 line mooring
arrangement in Figure 2.1 consists of 3 stern lines, 3
aft breast lines, 2 aft spring lines, 2 fore spring lines, 2.7. Downtime assessment
3 fore breast lines and 3 bow lines. The lines are
numbered from aft to fore.
Time domain simulations are used extensively to
assess the availability of turret moored terminals for

3 Copyright 2007 by ASME

offloading operations. The availability analysis
provides key input when assessing the economical
viability of a project. The mooring line loads, fender
loads and relative motions between the two vessels
are calculated for a large number of wind, wave and
current conditions. A systematic series of wave
heights and peak periods in the scatter diagram can
be run to establish the weather limits for the side-by-
side mooring operation. OCIMF (1988 and 1992)
safe working loads (SWL) are used as the criterion
for the mooring line loads and the fender loads. For
the cryogenic offloading arms, an operating envelope
of typically 2 m can be assumed, based on presently
available technology. More details regarding
downtime analyses can be found in Poldervaart
(2006) and Grin (2005).
The uptime of a floating LNG terminal needs
typically to be at least 98%. The time domain
simulation shows that offloading is possible for
significant wave heights above 3m. The threshold for
long period swells (Tp > 14 s) is usually lower and
also the thresholds for waves under an angle are
smaller than for head waves. Conventional side-by-
side mooring is normally considered possible for
waves up to 2.5 m significant and wind speeds up to
30 knots. The higher limits predicted with our
calculations can be explained by the use of the longer
(25 m) nylon tails and additional safety margins used
in real operations. Furthermore, it must be borne in
mind that the actual offloading is only one of many
stages of the whole berthing and offloading
operation, between LNGC arrival within a mile from
the FSRU and its sail away, see Poldervaart (2006).
3. NUMERICAL APPROACHES
3.1. Diffraction analyses and surface lid
method.
The first order and second order low frequency
hydrodynamics were calculated in the frequency
domain, using a three-dimensional source distribution
technique. The mean wetted part of the hull was
modelled by a large number of panels. The
distribution of source singularities on these panels
forms the velocity potential describing the fluid flow
around the hull. The pressure distribution on the hull
is calculated from the velocity potential. The shallow
water effects are incorporated using the so-called
Pinkster approximation; see Pinkster (1981). The
applied panel distributions for the side-by-side
configuration are depicted in Figure 2.2.
Figure 2.2- Fish-eye view of element distribution for
FSRU and LNGC.
Detailed investigation of the multiple-body
diffraction analysis led to the observation that
unrealistic wave oscillations can occur in the gap
between the two vessels, due to the absence of
viscous flow effects in the potential flow diffraction
analysis. To suppress these unrealistic wave
phenomena in the diffraction analysis, a free surface
lid was placed on the gap between the FSRU and the
alongside moored LNG carrier. A detailed discussion
of this technique has been presented by Huijsmans,
Pinkster and de Wilde (2001). As shown by Buchner,
van Dijk and de Wilde (2001), the low frequency
wave excitation and related low frequency motions
can be reduced to values comparable with model tests
with this approach. An alternative approach is to deal
with the free surface resonance in the gap as
discussed in Fournier et al (2006).
3.2. Time domain simulations
The linear response of a floating structure in waves
can be presented in the frequency domain. Such an
approach is however not suited for studying side-by-
side moored vessels, due to non-linearities as for
instance coming from the non-linear force-
displacement characteristics of the mooring lines and
fenders. Time domain simulations are therefore
needed. The impulse response theory, developed by
Cummins (1962) and Ogilvie (1964) and, can be used
to describe the fluid reactive forces in the time
domain (Van Oortmerssen (1973 and 1981)):
6 t
( M kj + mkj )x&& j + Rkj ( t )x& j ( ) d + Ckj x j = Fk ( t )
j =1
where:
xj = motion in j-direction
Fk(t) = external force in the k-mode of motion:
Mkj = inertia matrix
mkj = added inertia matrix at asymptotic
frequency
Ckj = matrix of hydrostatic restoring forces
Rkj = matrix of retardation functions
k,j = mode of motion.
4 Copyright 2007 by ASME
The duration of the retardation functions in LIFSIM
and aNySIM was 50s. In AQWA this duration is
120s.
Once the system of coupled differential equations is
obtained, arbitrary in time varying loading such as
wave excited forces, current forces, non-potential
fluid reactive forces and non-linear mooring forces
may be used as external forces on the right hand side
of the equation.
In the case of close proximity the equations of
motions needs to be solved in a coupled manner. For
a 2-body system, there are 12 degrees of freedom.
The 2 bodies can be subject to wave-induced forces,
hydrodynamic reaction forces and mechanical
coupling effects (either linear or non-linear). The
inertia and added inertia matrices and also the
matrices of the retardation functions are derived from
multi-body diffraction analysis in the frequency
domain. For multiple bodies not in close proximity,
as for instance in tandem offloading, only the
diagonal terms (Mi,j)i,j=1,6 and (Ri,j) i,j=1,6 are needed.
This implies that wave shielding of one body behind
another body is taken into account, but that the
hydrodynamic couplings are neglected. For side- by-
side offloading, Buchner, van Dijk and De Wilde
(2001) found that it was not possible to neglect the
off-diagonal blocks of the mass and retardation
function matrices.
The equations in the time domain were solved with a
multi-stage method using a time step of 0.5 s. The
duration of the simulations was 3.5 hours, with the
first 0.5 hour ignored due to transient effects.
3.3. Bandwidth of slow drift load P & Q
matrices
The second order wave drift forces were derived for
wave directions from 0 to 360 degrees with 15
increments and for frequencies from 0.2 rad/s to 1.5
rad/s with a 0.05 rad/s increments, using the second
order transfer functions:
N N {
Pij cos ( i j ) t + ( i j ) }
F(
2)
( t ) = i j
i =1 j =1
{
+Qij sin ( i j ) t + ( i
in which Pij is the in-phase part of the second order
transfer function and Qij the quadrature part of the
second order transfer function.
The P- and Q-matrices were truncated at their 8th off-
diagonal band, meaning that the slow drift wave
forces were calculated up to 0.4 rad/s difference
frequencies. This is a sufficiently large bandwidth in
view of the highest low frequency natural frequency
of the 2-body system of 0.18 rad/s, as shown in the
modal analysis in section 2.6.
3.4. Close proximity damping
One of the most important issues in the study of the
motions of moored vessels is the viscous damping of
the low frequency motions due to the low frequency
viscous reaction forces. For a single moored vessel,
this problem has been studied in detail by Wichers
(1987). This resulted in a complex model for the
viscous damping in both still water and in current,
making use of oscillation tests of tanker models in the
horizontal plane.
For side-by-side mooring the situation is even more
complex, considering the relative sway and yaw
motions between the two vessels. The water in the
small gap between the two vessels is oscillating in
and out, resulting in large water velocities around the
bilges. The resulting flow separation provides an
important contribution to the viscous damping.
Linear equivalent relative damping coefficient (Bry,
Br) have been derived from dedicated model tests,
by subtracting the damping of the shuttle alone (By,
B) from the damping of the shuttle next to the fixed
FPSO. This results in the following sway and yaw
damping forces in the coupled equation of motions
for the FPSO and shuttle tanker:
FFSRU = BFSRU .y& FSRU Bry .( y& FSRU y& LNGC )
y y
FFSRU = BFSRU .& FSRU Br .(& FSRU & LNGC )
FLNGC y = BLNGC y .y& LNGC + Bry .( y& FSRU y& LNGC )
FLNGC = BLNGC .& LNGC + Br .(& FSRU & LNGC )
3.5. Fender model
In LIFSIM, the point of application of fender loads is
kept fixed during the simulation. The fender friction
force is computed as f*Fnormal and is applied in the
horizontal direction regardless to the orientation of
the relative velocity vector. The direction of the
friction force is opposite the relative surge velocity.
j ) }
The fender compression force was based on the
specified non-linear load-compression characteristic
of the pneumatic fender. A fender friction coefficient
f of 12.5% was assumed, based on manufacturers
specification.
Ffriction _ FSRU = f Fnormal sign(x& FSRU x& LNGC )
Ffriction _ LNGC = f Fnormal sign(x& FSRU x& LNGC )
In which x& FSRU and x& LNGC are the surge velocity of
respectively the FSRU and the LNG carrier.
5 Copyright 2007 by ASME
It should be noted that in LIFSIM the fender friction
force is set to zero when the relative surge velocity
between the two vessels is zero.
In AQWA and aNySIM, the point of application of
the fender friction load is computed at each time step
and its direction is opposite to the relative velocity
vector. Consequently, this force can have a vertical
component. The magnitude of the friction force is
also based on a 12.5% friction.
In aNySIM the fender friction is a function of the
pressure force normal to the hull. The friction force
goes to zero for small values of the normal force.
3.6. Modelling differences LIFSIM,
AQWA and aNySIM
The three time domain simulation tools are in essence
very similar. However the following differences can
be noted:
The relative heading of wind, waves or current is
based on:
o AQWA: slow drift heading
o LIFSIM: instantaneous heading
o aNySIM: instantaneous heading
The 1st and 2nd order (low frequency) wave loads
are computed as follows:
o AQWA: evaluated at each time step
o LIFSIM: pre-calculated for several
wave heading and interpolated in the
time domain
o aNySIM: pre-calculated for 7 min
segments and recomputed when a 1 deg
heading criterion is exceeded.
Vertical drift loads:
o AQWA: Included,
o LIFSIM: Not included, but possible
o aNySIM: Not included, but possible
Direction of fender friction force is opposite
(see section 3.5):
o AQWA: Relative velocity,
o LIFSIM: Relative surge velocity,
o aNySIM: Relative velocity.
Solver type:
o AQWA: 2 step predictor/corrector,
o LIFSIM: MARIN Dverk solver,
o aNySIM: MARIN Dverk solver.
4. COMPARISON LIFSIM, AQWA &
ANYSIM
The benchmark study was started in 2005 with
LIFSIM and AQWA only. Five environments (A, B,
C, D and E) had been selected thus covering typical
wave heights, periods and directionalities. Case A
was an in-line case. All other cases had current at
90 and wind at 30 to the waves. Since LIFSIM
and AQWA differed in the encounter frequency
effects, cases with a current at 90 were more
suitable for a meaningful comparison. All LIFSIM
input quantities were provided and implemented in
AQWA-DRIFT. In particular the LIFSIM wave
elevation time series was imported to insure a
meaningful comparison. For the sake of brevity, the
paper focuses on Case B.
4.1. Case B
The following environment is considered (see Figure
4.1):
- 4.0 m significant waves from 180 degrees
- 16 s peak period waves.
- 16.7 m/s wind from 210 degrees
- 1.0 m/s current from 270 degrees
Hs/Tp = 4.0/16
180 deg
+x
+y
Uw 16.7 m/s
210 deg Uc 1.0 m/s
270 deg
Figure 4.1- Environment B
In this simulation, the FSRU is subjected to a 500kN
thruster force in order to reduce the relative wave
heading.
Time domain simulations of 3.5 hour duration are
compared. The time step selected in these simulations
is 0.5s. The starting positions of the FSRU and
LNGC are with a zero heading, that is with the FSRU
bow pointing in the positive X-direction. Comparison
of the three programs is done in terms of statistics
and of time series of motions and loads.
6 Copyright 2007 by ASME
4.2. Statistics for Case B AQWA and aNySIM agree very well for relative
surge and sway (less than 5%). Relative yaw is
improved compared to LIFSIM and is within 10%.
4.2.1. Mean FSRU heading 4.2.4. Tensions
The mean FSRU headings derived from the three
The maximum tension in the LNGC mooring lines
time domain simulations are shown in Table 4.1
must not exceed the safe working load for the
below.
terminal to be available for offloading for a given
LIFSIM AQWA aNySIM seastate. Before comparing the maximum values
22.3 25.2 24.9 obtained by the three programs, the RMS tensions are
compared for lines 1, 4, 7, 10, 13 and 16 in Figure
Table 4.1- Mean FSRU heading ().
4.2.
AQWA and aNySIM agree very well and predict a 80

RMS Tension [kN]

somewhat larger heading than LIFSIM. A good
agreement in mean heading is a pre-requisite to 68
comparable behaviour in the time domain.
56

44
4.2.2. FSRU and LNGC vertical motions
32
Vertical motion RMS are shown in Table 4.2 below. 20
LIFSIM AQWA aNySIM 1 4 7 10 13 16
Heave [m] 0.120 0.21 0.20
FSRU Roll [] 0.15 0.16 0.23 Line Number
Pitch [] 0.27 0.28 0.26
Heave [m] 0.41 0.42 0.39 LIFSIM
LNGC Roll [] 0.30 0.49 0.49 AQWA
Pitch [] 0.58 0.63 0.55 aNySIM
Table 4.2- FSRU and LNGC RMS motions.
Figure 4.2- RMS mooring line tension.
Heave responses are recovered within less than
The general shape of the RMS curves is consistent.
10%. Slightly larger differences are seen in the pitch
All three programs predict the largest RMS response
response especially for the LNGC (12.6%). Although
small, the roll responses predicted by LIFSIM and for line 13 (breast line near the bow as shown in
Figure 2.1). AQWA and aNySIM predict consistent
aNySIM differ by more than 50% for both vessels.
values for line 13 (4% difference). Other RMS values
differ by up to 30% (line 1 at the stern). Maximum
calculated tensions for a 3-hour storm are shown in
4.2.3. FSRU/LNGC relative horizontal Figure 4.3 below.
motions
1000
Maximum Tension [kN]

Rather than comparing X and Y motions of the FSRU

and LNGC (as defined previously in Figure 4.1), 800
surge and sway motions are derived for both vessels
and then relative surge, sway and yaw motions. These
motions are more easily correlated with loads in 600
mooring lines. Furthermore, these relative motions
are needed to assess whether the manifold motions on
400
the LNGC are in within the loading arm operational
envelope. Table 4.3 below summarizes the RMS
relative motions: 200
LIFSIM AQWA aNySIM 1 4 7 10 13 16
Surge [m] 0.41 0.46 0.46
Sway [m] 0.13 0.14 0.14 Line Number
Yaw [] 0.10 0.11 0.13
Table 4.3- RMS FSRU/LNGC relative motions. Figure 4.3- Maximum mooring line tensions.

7 Copyright 2007 by ASME

 Excellent agreement is found for the heave and pitch
AQWA and aNySIM agree better than AQWA and motions. Roll predictions differ by at most 1 over
LIFSIM. In particular the largest tension predicted by the selected time window.
AQWA and aNySIM differ by only 50kN. It is
noteworthy that breast lines (1 to 6 and 11 to 16) are 4.3.3. Relative motions
the most loaded suggesting predominance of relative
sway and yaw motions. The most loaded line is
Figure 4.6 shows comparison of relative surge, sway
predicted consistently.
and yaw motions. Relative motions are computed as
LNGC motions minus FSRU motions.
4.3. Time series of Case B Relative Surge [m]
1

4.3.1. Wave elevations -1

-2
Max Tension Line 13
-3
Wave elevations generated by LIFSIM are imported 1.223 1.2235 1.224 1.2245 1.225 1.2255 1.226 1.2265 1.227 1.2275
4
x 10
in AQWA. Figure 4.4 below shows the comparison 1
Relative Sway [m]

of the two time series for a 100s-long window 0.5

centred around the time of maximum tension in 0

mooring line 13. -0.5

Max Tension Line 13
-1
1.223 1.2235 1.224 1.2245 1.225 1.2255 1.226 1.2265 1.227 1.2275
Wave elevations [m] 4
x 10
5 Relative Yaw [deg]
Max13
0.6
AQWA
0.4
LIFSIM
0.2 ANYSIM

0
0
Max Tension Line 13
-0.2
1.223 1.2235 1.224 1.2245 1.225 1.2255 1.226 1.2265 1.227 1.2275
time[s] 4
x 10
LIFSIM
AQWA Figure 4.6 - Comparison of relative surge, sway and
5
4 4 4 4 4 4
yaw.
1.22 . 10 1.222 . 10 1.224 . 10 1.226 . 10 1.228 . 10 1.23 . 10

The above figure shows that of all programs AQWA

Figure 4.4- Wave elevations. predicts the largest relative surge (LNGC is surging
Wave elevation RMS computed by LIFSIM and almost 2m aft relative to the FSRU) at the time of
imported in AQWA differ by only 0.3%. maximum load in breast line 13. The difference in
relative yaw of about 0.2deg amounts to a relative
sway motion of about 0.35m at 100m from COG. For
4.3.2. Vertical motions reference, the Nylon elongation required for 15 tons
is 1.4m and that for reaching the mooring line safe
Figure 4.5 shows comparisons of the LNGC heave, working load (69 tons) is 3.6m.
roll and pitch motions.

2
LNGC Heave [m] 4.3.4. Tensions
Max Tension Line 13
1

0
Figure 4.7 below shows the time series of tension in
-1
the most loaded line (Line 13) obtained with the three
-2
1.223 1.2235 1.224 1.2245 1.225 1.2255 1.226 1.2265 1.227 1.2275
4
programs.
x 10 Tension Line 13 [kN]
LNGC Roll [deg] 1000
2 Max Tension AQWA
Max Tension Line 13 800 LIFSIM
1
ANYSIM
600
T13 [kN]

-1 400

-2 200
1.223 1.2235 1.224 1.2245 1.225 1.2255 1.226 1.2265 1.227 1.2275
4
x 10 0
1.223 1.2235 1.224 1.2245 1.225 1.2255 1.226 1.2265 1.227 1.2275
LNGC Pitch [deg]
4 time [s] 4
x 10
Max Tension Line 13
2 Figure 4.7- Time series of tensions in line 13.
0
AQWA
-2 LIFSIM
ANYSIM The figure above shows that the time of maximum
-4
1.223 1.2235 1.224 1.2245 1.225
time[s]
1.2255 1.226 1.2265 1.227 1.2275
4
tension is consistently predicted by AQWA and
x 10
aNySIM. However, the peak heights differ. In order
Figure 4.5 - Snapshots of LNGC vertical motions.
to identify the source of these differences, the relative
surge, sway, heave, roll, pitch and yaw are plotted at
the time of the maximum tension in Figure 4.8

8 Copyright 2007 by ASME

(presented at the end of paper). The extreme values
obtained with AQWA for each of the above relative The previous figure shows that LIFSIM predicts a
motions are indicated by horizontal red dash lines. At stick-slip situation between 25 and 75s whether the
the time of maximum tension in line 13, only the other two programs always predict relative surge
relative surge and pitch reach an extreme value. The motion.
remaining relative motions assume average values.
5. CONCLUSIONS
The maximum tension in the simulation therefore
corresponds to a combination of extreme relative LF
A benchmark study has been performed with three
surge and HF pitch motions. Owing to the difference
different programs: AQWA, LIFSIM and aNySIM,
is vessel depths (FSRU deck is 7.4m higher than
for side-by-side offloading from an LNG Carrier to a
LNGC deck), when the LNG carrier pitches more
turret-moored FSRU in intermediate water depth.
than the FSRU (positive pitch is bow down), the
Identical environmental input has been used. The
relative elevation of the mooring line fairleads
three programs give very similar results in statistics
increases and so does the tension.
(mean and RMS) and time traces.
It is interesting to note that neither the absolute nor
However, important differences in the extreme line
the relative roll motions are correlated with the
loads can be observed. These may be explained by
maximum tension event.
small differences in the modelling methods that were
used, such as the interpolation of the wave forces, the
4.4. Decay test in surge application of vertical drift forces, the numerical
solver and the fender friction model. It should be
In order to shed some light on the origins of the noted that, due to the strong non-linearities of the
observed differences a decay test was performed. In mooring lines and the fenders, the mooring line
this numerical experiment, the FSRU is in its nominal forces are very sensitive to small differences in the
position in calm water and the LNGC is displaced (relative) motions.
backwards by one meter in surge. The lateral distance
between the two vessels is kept at 4m. Figure 4.9 It is however expected that the observed sensitivity in
below illustrates the FSRU and LNGC surge motions. the calculated line loads will be acceptable for a first
0.4
Surge FSRU [m] downtime analysis of the side-by-side configuration.
0.2
This is because it is plausible that the variation in the
extremes will average out over a large number of
Surge [m]

AQWA
simulations for a specific site.
-0.2
LIFSIM
ANYSIM
-0.4
0 100 200 300 400 500 600 700 800 900 1000 The present study has resulted in an improved
1
Surge LNGC [m] understanding of the complex dynamic behaviour of
two side-by-side vessels. For instance, the maximum
0.5
line load in the breast lines for case B occurs after a
Surge [m]

0
combination of a low frequency relative surge
-0.5
extreme and a maximum pitch response of the LNG
-1
0 100 200 300 400 500 600 700 800 900 1000 carrier.
time [s]

The following practical conclusions can be drawn:

Figure 4.9- Time series of surge motions during High availability can only be achieved by
decay simulation. Top: FSRU; Bottom: LNGC. increasing the standard 11m-long Nylon
tails significantly. In this study 25m long
Striking differences are seen between the three tails are selected giving a range of
programs. These differences are attributed to the elongation of 3.6m before the SWL is
implementation of fender friction. Figure 4.10 below attained.
shows the relative surge predicted by the three It is preferable to have the mooring
programs. attachment points on the FSRU and LNGC
Relative Surge LNGC [m] Relative Decay
0.5 at comparable elevation with respect to
MSL. This reduces relative pitch induced
Relative Surge [m]

0
tensions.
-0.5
AQWA
It is found that large relative roll events are
LIFSIM generally not correlated to peak values of
ANYSIM
-1
0 50 100 150 200 250 300 350 400 mooring line tensions for the range of
time [s]
relative wave headings considered.
Figure 4.10- Time series of relative surge motion
during decay simulation.

9 Copyright 2007 by ASME

 It is important to recall that offloading is
only one phase in the LNGC chain. The
LNGC has first to make its approach and
connect its mooring lines and loading arms
before offloading can commence. These
phases are heavily dependant on the
operability of tugs. An industry-wide project
is on-going to assess these aspects.
A sufficient number of off-diagonal QTF
terms should be included in the computation
of the 2nd order wave drift loads. This is
related to the relatively high natural periods
(30-200s) in the side-by-side mooring
configuration.
Future investigation should focus on the following
important issues:
Fender modelling (direction of the friction
force and its point of application) and
characterisation of reliable friction values
for design.
Wind and current coefficients for side-by-
side vessels (wind and current shielding
effects).
Modelling of the wave resonance in the gap
between two side-by-side vessels. How to
best deal with this resonance for first and
second order results of diffraction analyses (
see Fournier et al (2006)).
6. REFERENCES
[1] Naciri M. & J.P. Quau The Soft Yoke
Mooring and Offloading System for LNG
Offloading Application. Proceedings of the
22nd OMAE Conference, Paper OMAE-03-
37135 Cancun Mexico June 8th -13th 2003.
[2] Ye W., Luo, Y. and Pollack, J.: LNG Floating
Regasification Unit (FRU) Side-by-side
Mooring Analysis, Proceedings of the 24th
OMAE Conference, Paper OMAE2005-67117,
Halkidiki Greece June 2005.
[3] Valk, C.A.C. van der et. al.; Chiksan LNG
Marine Loading Arm Enhanced for
Application in Exposed Areas, LNG-14
Conference, Doha, Qatar, 2004.
[4] Oil Companies International Marine Forum
(OCIMF), Mooring Equipment Guidelines,
First Edition, 1992.
[5] Oil Companies International Marine Forum
(OCIMF), Ship to Ship Transfer Guide
(Petroleum), Second Edition, 1988.
[6] Poldervaart L., H. Oomen & J. Ellis. Offshore
LNGC transfer: A worldwide Review of
Offloading Availability. Proceedings of the
Offshore Technology Conference Paper OTC
18026 PP. Houston Texas USA May 2006.
[7] Grin R.,J. de Wilde & J. van Doorn.
''Assessment of LNG Transport Chains Using
Weather Based Voyage Simulations '' ,
SNAME 2005-ETT-003
[8] Pinkster, JA (1981). "Low-Frequency Second
Order Wave Exciting Forces on Floating
Structures," PhD Thesis, Delft University of
Technology.
[9] Huijsmans R.H.M., Pinkster J.A. and de Wilde
J.J. (2001), Diffraction and radiation of waves
around side-by-side moored vessels, ISOPE
2001, Stavanger.
[10] Buchner B., van Dijk A.W.V. and de Wilde
J.J. (2001), Numerical multiple-body
simulations of side-by-side mooring to an
FPSO, ISOPE 2001, Stavanger.
[11] Fournier J.-R., M. Naciri & X.-B. Chen
Hydrodynamics of Two Side-by-Side
Vessels. Experiments & Numerical
Simulations. Proceedings of the ISOPE
Conference. Paper ISOPE-06-pt-02. San
Francisco, California, USA. June 2006.
[12] Cummins, WE (1962). "The Impulse
Response Function and Ship Motions",
DTMB Report 1661, Washington D.C.
[13] Ogilvie, TF (1964). "Recent Progress toward
the Understanding and Prediction of Ship
Motions", Fifth Symposium on Naval
Hydrodynamics, Bergen.
[14] Oortmerssen, G van (1973): "The Motions of
a Moored Ship in Waves", NSMB Publication
No. 510.
[15] Oortmerssen, G van (1981): "Some
Hydrodynamical Aspects of Multi-Body
Systems", Int. Symposium on Hydrodynamics
in Ocean Engineering.
[16] Wichers, JEW (1987). A Simulation Model
for a Single Point Moored Tanker, PhD thesis
Delft University of Technology.
10 Copyright 2007 by ASME
 Maximum tension in Line 13 (aNySIM), Aqwa: 860 kN, Lifsim: 667 kN aNySIM: 810 kN
Relative-Surge [m] Relative-Sway [m] Relative-Heave [m]
2 Max Tension
1 Max Tension
2 Max Tension
max aqwa max aqwa max aqwa
1
0.5 1
0
0 0
-1
AQWA
-0.5 min aqwa -1
-2 LIFSIM min aqwa
min aqwa
ANYSIM
-3 -1 -2
1.223 1.224 1.225 1.226 1.227 1.223 1.224 1.225 1.226 1.227 1.223 1.224 1.225 1.226 1.227
4 4 4
x 10 x 10 x 10
Relative-Roll [deg] Relative-Pitch [deg] Relative-Yaw [deg]
2 2 max aqwa 1 Max Tension
max aqwa 1 max aqwa
1
0.5
0
0
-1
0
-1
min aqwa
-2 min aqwa
min aqwa
Max Tension Max Tension
-2 -3 -0.5
1.223 1.224 1.225 1.226 1.227 1.223 1.224 1.225 1.226 1.227 1.223 1.224 1.225 1.226 1.227
4 4 4
x 10 x 10 x 10
Roll FSRU [deg] Roll LNGC [deg] Fline13 [kN]
1 Max Tension
2 1000 Max Tension
max aqwa
max aqwa max aqwa
800
1
0.5
600
0
400
0
-1
200
min aqwa
Max Tension min aqwa
-0.5 -2 0 min aqwa
1.223 1.224 1.225 1.226 1.227 1.223 1.224 1.225 1.226 1.227 1.223 1.224 1.225 1.226 1.227
time[s] 4 time[s] 4 time[s] 4
x 10 x 10 x 10

Figure 4.8 Snapshot time traces of motions, relative motions and line loads, calculated with LIFSIM, AQWA
and aNySIM, for maximum line load event in case B.

11 Copyright 2007 by ASME