Ocean Engineering 195 (2020) 106731

Contents lists available at ScienceDirect

Ocean Engineering
journal homepage: www.elsevier.com/locate/oceaneng

Prediction of dynamic responses of FSRU-LNGC side-by-side

mooring system
Jingxia Yue a, *, Weili Kang a, Wengang Mao b, Pengfei Chen c, Xi Wang d
a
School of Transportation, Wuhan University of Technology, Wuhan, China
b
Department of Mechanics and Maritime Sciences, Chalmers University of Technology, Gothenburg, Sweden
c
Shanghai Investigation, Design & Research Institute Co., Ltd, Shanghai, China
d
CCS Wuhan Rules & Research Institute, Wuhan, China

A R T I C L E I N F O A B S T R A C T

Keywords: Floating Storage and Regasification Unit (FSRU) becomes one of the most popular equipment in the industry for
FSRU providing clean energy because of its technical, economic and environmental features. The interaction between
Side-by-side mooring the FSRU and Liquified Natural Gas Carrier (LNGC) under the combined loads from wind, wave and current is
Multi-floating mooring system
quite complex to model. In this paper, a configuration for the offloading operation of the FSRU-LNGC side-by-
Dynamic response
The damping lid method
side mooring system is proposed to predict the motion responses, forces on the cables and fenders of the multi-
floating mooring system. The damping lid method is adopted to improve the overestimated hydrodynamic co­
efficients calculated from conventional potential flow theory in the frequency domain. The dynamic response of
the side-by-side mooring system including six degrees of freedom motion, relative motions, cable tensions and
fender forces are provided and analyzed. The numerical results are validated using the experimental data. The
proposed coupled analysis model and the numerical analysis can properly predict the dynamic response of the
multi-floating mooring. The sensitivity analysis of pretension of the connecting cables on the dynamic responses
of the two vessels are provided. Moreover, the non-dimensional damping parameters can be acted as a good
reference to the dynamic response analysis of similar multi-floating mooring systems.

1. Introduction ● The simulation of the viscous effect of free surface between FSRU and
LNGC;
With the wide application of Liquified Natural Gas (LNG), the ● The detailed analysis of the dynamic responses of the multi-floating
Floating Storage and Regasification Unit (FSRU) has attracted wide mooring system;
attention. During the LNG offloading operation, the multi-floating
mooring system consisting of a FSRU and an LNG Carrier (LNGC) can For the multi-floating mooring system in side-by-side offloading
be in side-by-side or tandem configuration. The side-by-side offloading operation, experimental methods are still recognized as the most reliable
operation for the multi-floating mooring system is of more interest techniques to obtain of the dynamic responses (Zhao et al., 2017). For
because more accurate analysis of hydrodynamic interactions between example, Hong (Hong et al., 2005) presented a higher-order boundary
FSRU and LNGC moored in close proximity is needed (Hong et al., element method (HOBEM) combined with generalized mode approach
2005). To get accurate dynamic responses of the mooring system, rela­ and verified it by the experiment. Numerical results using HOBEM show
tive motion of the two vessels, and the load distribution on the cables good agreements with experiments for global and local motion response
and fenders due to wind, wave and current is essential to the side-by-side and wave drift force of side-by-side moored vessels in both regular and
offloading operation safety. During the prediction of the system dynamic irregular waves. However, exception is observed for a wave drift force in
response, the following items should be considered carefully: very narrow frequency region where strong interaction occurs due to
Helmholtz resonance. Zhao (Zhao et al., 2017) provided an experi­
● The arrangement of the side-by-side configuration; mental investigation on dynamic responses of the connection system in
the FLNG system during side-by-side offloading operations. In this

* Corresponding author.
E-mail address: [email protected] (J. Yue).

https://doi.org/10.1016/j.oceaneng.2019.106731
Received 17 May 2019; Received in revised form 25 October 2019; Accepted 11 November 2019
Available online 25 November 2019
0029-8018/© 2019 Elsevier Ltd. All rights reserved.
J. Yue et al. Ocean Engineering 195 (2020) 106731

research, relationships between relative vessel motion and the load born validate the numerical results. In order to achieve the above objective,
by the connection system are obtained and features of dynamic Section 2 gives the description of the configuration, Section 3 provides
connection system responses are summarized, but the detailed discus­ the numerical calculation method and calculation process, Section 4
sions on mooring system are not presented. Inoue (Inoue and Islam, introduces the model test set up, Section 5 verifies the numerical results
1999) compared numerical and experimental results of parallel con­ according to the experiment and carries out a detailed analysis of the
nected FPSO and LNG carriers in waves. A numerical multiple body dynamic response of the system, Section 6 analyzes the sensitivity of the
simulation model for the reliable prediction of relative motions and dynamic response of the system to the pretension of the connecting
mooring loads during side-by-side offloading operations is also devel­ cable.
oped. Huang (Huang et al., 2018) provided the numerical approach and
model for gangway response between nonparallel side-by-side flotel and 2. Description of the FSRU-LNGC side-by-side configuration
FPSO to predict the physics of various gangway responses. Numerical
simulation and measurement reach pretty good agreement, which shows 2.1. Features of the FSRU-LNGC side-by-side mooring system
that the established physical model can properly represent physics of
gangway response. But this numerical approach and model are not The conceptual FSRU-LNGC Ship to Ship Mooring Arrangement
applicable to the side-by-side mooring system. Pessoa (Pessoa et al., developed by Dalian Shipbuilding Industry Engineering and Research
2016) validated a frequency domain numerical method for calculating Institute of China is selected as a reference. The side-by-side configu­
second order responses to irregular wave excitations through compari­ ration of the multi-floating mooring system is proposed based on the
sons with experimental data obtained in an offshore wave basin. It is Mooring Equipment Guidelines (OCIMF, 2008) and taking into account
shown that the second order low frequency loads cannot be neglected the configuration of the wharf. The FSRU-LNGC side-by-side mooring
when calculating the tension on the side by side mooring system. Zhao system consists of FSRU-wharf mooring system and FSRU-LNGC con­
(Zhao et al., 2013) studied two different ways of connecting FLNG vessel necting system, the general arrangement of which is illustrated in Fig. 1.
and LNG carrier. The numerical results show that there is significant The FSRU is designed to be moored at the wharf, and the LNGC is
difference at the hydrodynamic performance in the two ways of tandem connected alongside the FSRU during the side-by-side offloading oper­
offloading operations, which means that the connection between FLNG ation. Full loaded FSRU and ballast LNGC are considered in this study
vessel and LNG carrier plays an important role. Vieira (Vieira et al., because test results showed more severe motions and load responses in
2018) investigated the influence of the liquid inside the tanks in the this loading condition (OCIMF, 2008). The principal dimensions of the
wave behavior of FLNG vessels in side-by-side offloading operations FSRU and LNGC are introduced in Table 1. The mooring system is
experimentally. The study showed that the analysis of coupled systems designed to be a “4-3-2” symmetrical arrangement and consists of 18
considering all the effects is very important for the correct definition of cables and 4 fenders. Similarly, the connecting system consists of 24
the dynamics of the vessels, which include coupled hydrodynamic be­ cables and 8 fenders and is designed to be a “2-2-3-3-2” symmetrical
tween vessels, the impact of LNG tanks free surface on systems motions, arrangement. The detailed particulars of the two systems are shown in
the resonance effect on the free surface of the gap between hulls and the Table 2 and Table 3. The side-by-side distance between FSRU and LNGC
influence of mooring lines and fenders. However, the influence of each is set to be 4.5 m, which is the same as the length of the fenders. It should
factor on the dynamic responses of the system is not quantified. be noted that the safety requirements can be satisfied when the loads of
Theoretical and numerical researches on the multi-floating mooring the cables or the forces of the fenders are less than their safe working
system were studied in order to analyze the side-by-side configuration load (SWL). In order to verify the safety and rationality of the above
realistically. Because of the complex multi-floating mooring system, the configuration, numerical calculation and test methods are used to
viscous effect of free surface between two ships is very obvious. Some analyze the dynamic response of the multi-floating mooring system. The
assumptions and simplifications should be adopted to obtain more cor­ analysis methods are shown in Section 3 and Section 4, respectively.
rect numerical results (Huang et al., 2018). Newman (Newman, 2001;
Newman and Nicholas, 1977) reviewed the extensive analytical results 2.2. Environmental condition
and accomplishments from numerical simulation. Buchner et al. (2001)
introduced a rigid lid on the free surface between the vessels within the In the numerical and experimental analysis, wave, wind and current
multi-body diffraction analysis to suppress the unrealistic resonant wave are considered in the side-by-side offloading of the FSRU-LNGC multi-
oscillations. Similar studies were also shown by Naciri (Naciri et al., floating mooring system. According to the long-term distribution data of
2007), which verified the simulation results through experiment. Chen coastal waves in China (Yangshan et al., 2007) and the Design Specifi­
(2004) proposed the damping lid method. They added a dissipative term cation for LNG Terminals (Ministry of Communications of the People’s
in the free surface boundary condition inside the gap. Chen and Mal­ Republic of China, 2009), the relevant parameters of the waves are
enica (2005) discussed the damping method applied in the multibody determined. The Mooring Equipment Guidelines (OCIMF, 2008) gives
problem. The method basically reduced the source strength by adjusting detailed wind and current environment requirements for the system
the terms associated with the damping effects. moored at the wharf. According to the guidelines (OCIMF, 2008), the
Previous research activities described above focused on the detailed
study of partial problems, such as the hydrodynamics interaction, the
resonant wave oscillations of the free surface between the side-by-side
vessels, or the implementation of newly developed methods. However,
researches on the prediction of dynamic responses of the side-by-side
configuration are far from enough. More experimental researches
should be carried out to illustrate response characteristics of multi-
floating mooring system. The objective of this paper is to propose a
safe and reliable configuration of FSRU-LNGC multi-floating mooring
system and to predict the dynamic response of the system. In the nu­
merical calculation, the damping lid method is presented to correct the
distortion of hydrodynamic coefficients caused by fluid resonance be­
tween two floating bodies, so as to obtain a more realistic dynamic
response to verify whether the configuration meets the specification
requirements. Corresponding model tests were also carried out to Fig. 1. The side-by-side configuration of the FSRU-LNGC mooring system.

2
J. Yue et al. Ocean Engineering 195 (2020) 106731

Table 1
Principal dimensions of the FSRU and LNGC.
Item Symbol Unit FSRU (full) LNGC
(ballast)

Length over all L m 340.5 290

Length between perpendiculars Lpp m 330 278
Breath B m 60 45.6
Depth D m 25.72 29
Draft T m 11.98 9.61
Displacement Δ MT 202246 95887
Center of the gravity above base KG m 17.14 12.53
Center of the gravity from AP LCG m 168.12 143.37
Radius of roll gyration Kxx m 20.42 18.78
Radius of pitch gyration Kyy m 83.62 72.77
Radius of yaw gyration Kzz m 86.95 74.18

Table 2
Main parameters and attributes of the cables.
Item Unit Mooring system Connecting system

Diameter mm 52 40
Weight in air kg/m 1.46 0.881
SWL kN 1734.6 1107.4

Table 3
Main parameters and attributes of the fenders.
Item Unit Mooring system Connecting system

Diameter � length m�m 4.5 � 9 4.5 � 9

Maximum deformation m 2.7 2.7
SWL kN 19600 19600

relevant parameters of wind and current are determined. The irregular

wave is described by a three-parameter JONSWAP spectrum with a
significant wave height of 2 m, a peak period of 8 s and a peak
enhancement factor of 3.3. The mean velocity of the steady wind is
20 m/s. The current velocity near the surface is 1.1 m/s. In the model
tests and the numerical simulations, wave and current approach the
multi-floating mooring system with the heading Angle of 45� , wind
Fig. 2. Procedures to the numerical analysis to estimate the motion and forces
approaches the system with the heading Angle of 90� .
of the LNGC and FSRU interaction under actual sea conditions.

3. Numerical modeling

During the operation, liquefied natural gas will be transferred from are obtained through time-domain calculation. The calculation process
LNGC to FSRU. In the numerical simulation, the transfer is regarded as a is shown in Fig. 2.
quasi-static process, and only the coupled motion of the multi-floating
mooring system in the initial state is studied. Numerical analysis is 3.1. Frequency-domain analysis
carried out in ANSYS-AQWA based on the potential flow theory (Hong
et al., 2009). The AQWA-Line, AQWA-Librium and AQWA-Drift modules Based on the potential flow theory, the velocity potential satisfies the
are used for frequency-domain calculation, static equilibrium calcula­ Laplace equation (Hong et al., 2009).
tion and time-domain calculation respectively. The hydrodynamic co­
∂2 Φðx; y; z; tÞ ∂2 Φðx; y; z; tÞ ∂2 Φðx; y; z; tÞ
efficients such as added mass, first-order and second-order wave drift þ þ ¼0 (1)
∂x2 ∂y2 ∂z2
force and potential damping of the two floating bodies are calculated in
frequency-domain. The damping lid method is used to improve the
where Φðx; y; z; tÞ is the velocity potential function of the coordinates x, y
conventional potential flow theory for the above-mentioned hydrody­
and z and the time t.
namic coefficients. The static equilibrium position of FSRU and LNGC
The velocity potential can be expressed as
and the pretension of cables in the multi-floating mooring system are � �
obtained by static equilibrium calculation under the given environment Φðx; y; z; tÞ ¼ Re φðx; y; zÞe iwt (2)
condition. The dynamic responses of the multi-floating mooring system

3
J. Yue et al. Ocean Engineering 195 (2020) 106731

where w is the oscillation frequency. the velocity of the fluid and opposite to its direction. The function of the
The Φðx; y; z; tÞ can further be divided into the radiation potential inner force is the same as the viscous effect and energy dissipation of the
φR ðx; y; zÞ and scattering potential φs ðx; y; zÞ. Scattering potential fluid between two floating bodies, but the vortex is not introduced so
φs ðx; y; zÞ can further be divided into incident-wave potential φI ðx; y; zÞ that the existence of velocity potential is guaranteed. The non-
and diffraction potential φD ðx;y;zÞ. Thus, the φ (x, y, z) can be expressed dimensional damping parameter of the lid is reasonably evaluated by
as referring to the model tests. Fig. 3 presents the numerical model of the
FSRU-LNGC side-by-side mooring system with a lid added to the free
φðx; y; zÞ ¼ φR ðx; y; zÞ þ φI ðx; y; zÞ þ φD ðx; y; zÞ (3)
surface between two ships to simulate the viscosity and the energy
The velocity potential can be transformed into the integral equation dissipation. The mesh area of the gap region is empirically defined as
on the surface by using the Green formula. Each part of the potential in explained by Chen (Chen and Malenica, 2005), which generally covers
Eq. (3) can be solved under their corresponding boundary conditions. the major part of the gap along the LNGC. The length and width of the
Wave forces acting on FSRU and LNGC can be obtained by adopting the damping lid are 290 m and 4.5 m respectively, which are the same as the
incident-wave potential and diffraction potential through the Bernoulli length of LNGC and the width of the gap between the two ships.
equation. The hydrodynamic coefficients induced by the oscillation of Based on this fairly perfect fluid, the classical Bernoulli’s equation is
the floating bodies can be solved by the radiation potential, like the modified as:
damping forces and the added masses.
P 1
2 3 þ gz þ Φt þ ðrΦÞ2 þ μΦ ¼ 0 (7)
ZZ ρ 2
∂φ
amn ðwÞ ¼ Re4ρw φm n ds5 ðm; n ¼ 1; 2; …; 6Þ (4)
∂n We define that the Cartesian coordinate system with xy-plane co­
S0
incides with the calm water and positive z-axis is pointing upwards.
2 3 Then the wave elevation at the free surface can be expressed as:
ZZ
∂φ
cmn ðwÞ ¼ Im4ρw φm n ds5 ðm; n ¼ 1; 2; …; 6Þ (5) z ¼ ζðx; y; tÞ (8)
∂n
At the free surface, the pressure in Bernoulli ‘s equation is identical to
S0

where φm represents the induced velocity potential when the floating the atmospheric pressure, that is:
body is oscillating in direction m with a unit speed. amn ðwÞ and cmn ðwÞ 1
gζ ¼ Φt ðrΦÞ2 μΦ (9)
means the added mass and the potential damping in direction m induced 2
by the body oscillation in direction n. ρ is the density of the fluid, and S0
Additionally, fluid particles on the free surface always remain on the
is the wet surface area of the floating body.
wave surface, which means that all the fluid particles on the free surface
can only make tangential movement along the surface and the normal
3.2. Improvement of hydrodynamic coefficients in frequency-domain
velocity of fluid particle is the same as that of the free surface. This
analysis
condition generates the kinetic condition of the free surface:

The theory of potential flow is not able to consider the viscosity and ζt þ Φx ζx þ Φy ζy Φz ¼ 0 (10)
energy dissipation of the hydrodynamic interaction between two ves­
sels. In the case of resonance, the hydrodynamic parameters of floating
3.3. Time-domain analysis
body calculated in frequency domain based on potential flow theory are
too large and have obvious distortion. Therefore, the numerical calcu­
In the multi-body modeling, the hydrodynamic interactions of the
lation method based on the potential flow theory needs to simulate the
two floating bodies are also taken into consideration, in the form of
damping due to the viscous separation of the fluid to model the real
coupled added masses and coupled retardation functions. The motion
circumstance. The damping lid method was introduced in detail by Chen
equations of the time-domain coupled analysis for FSRU or LNGC are as
(2004) and the sensitivity of the results for this method was presented by
follows (Hong et al., 2009):
Chen et al. (Chen and Malenica, 2005) and Chakrabarti (1978). In this
Z t
paper, the damping lid method based on the fairly perfect fluid is used to
½M þ að∞Þ�f€ξg þ D1 fξg
_ þ D2 f ðfξgÞ
_ þ Kfξg þ _ τ ¼ F wind
hðt τÞfξgd
add damping to the free surface between floating bodies. It results in a 0
linear damping term in free surface boundary condition. The charac­ þ Fwave þ Fcurrent þ Fext
teristics of the fairly perfect fluid are based on the assumption that the
(11)
fluid particle is subjected to an internal force proportional to the
magnitude of fluid velocity besides gravity. The internal force F can be where M is the generalized mass matrix for the vessel, að∞Þ is the added
expressed as: mass matrix at the infinite frequency, D1 and D2 are the linear and
F¼ μrΦ (6) quadratic damping matrices respectively, K is the hydrostatic restoring
stiffness matrix. Fwind , Fwave and Fcurrent represent the wind drag force, the
where Φ is the velocity potential and μ is the damping parameter and is wave drag force and the current drag force respectively. Fext represents
Rt
defined as a positive value. The internal force is directly proportional to any other forces. hðt τÞfξgd
0
_ τ means the retardation function matrix.
The hydrodynamic interaction between FSRU and LNGC needs to be
considered in the time-domain calculation of the multi-floating mooring
system. The effects of the hydrodynamic interaction on the frequency
dependent added mass and the damping forces are included in the
coupled added mass and the coupled retardation functions at the infinite
frequency. Thus, the two 6 � 6 matrices ½M það∞Þ� and ½hðt τÞ� in Eq.
(11) should be written as 12 � 12 matrices respectively.
� �
ðM þ að∞ÞÞi;i ðað∞ÞÞi;j
½M þ að∞Þ� ¼ (12)
ðað∞ÞÞj;i ðM þ að∞ÞÞj;j

Fig. 3. Numerical models based on the fairly perfect fluid.

4
J. Yue et al. Ocean Engineering 195 (2020) 106731

� �
hðt τÞi;i hðt τÞi;j would be used in the convergence analysis, and the non-dimensional
½hðt τÞ� ¼ (13)
hðt τÞj;i hðt τÞj;j damping parameters which would serve as a reference for the selec­
tion of the non-dimensional damping parameter used in damping lid
where the subscripts i and j represent the FSRU and LNGC respectively. It method. The damping parameters obtained in the test are used as a
should be indicated that the item with the same indices such as i,i or j,j is reference for non-dimensional damping parameter in numerical calcu­
equal to the item in the single vessel, and the item with the different lation, which provides a rough range for the value of non-dimensional
indices like i,j indicates the effect of LNGC on FSRU. Similarly, there is damping parameter in numerical calculation. In addition, the main
an item with a subscript as j,i. purpose of the combined wind, wave, and current test is to obtain an
Therefore, the time-domain coupled motion equation of multi- actual response of the FSRU-LNGC side-by-side offloading operation,
floating mooring system can be expressed as: which would be used to verify the validity of the numerical results.

� �� € � � �� _ � � � �� _ ��
ðM þ að∞ÞÞi;i ðað∞ÞÞi;jξi ðD1 Þi;i ðD1 Þi;j ξi ðD2 Þi;i ðD2 Þi;j ξi
þ þ f þ
ðað∞ÞÞj;i ðM þ að∞ÞÞj;j €
ξj ðD1 Þj;i ðD1 Þj;j ξ_ j ðD2 Þj;i ðD2 Þj;j ξ_ j
� �� � Z t � �� _ � � � (14)
ðKÞi;i ðKÞi;j ξi hðt τÞi;i hðt τÞi;j ξi Fi
þ dτ ¼
ðKÞj;i ðKÞj;j ξj 0 hðt τÞj;i hðt τÞj;j ξ_ j Fj

According to the experimental requirements, the basin can simulate

where Fi represents the forces on FSRU, which is the same as the right various marine environmental conditions which consists of wind, wave
side of Eq. (11). Similarly, the Fj means the forces on LNGC. The six and current and has adjustable water depth. The dimension of the basin
degrees of freedom motion response of LNGC-FSRU multi-floating is 50 m long, 40 m wide, 10 m in depth and the water depth was set at
mooring system can be obtained by solving the time-domain coupled 0.938 m, which is corresponding to actual water depth of 60 m.
motion equation. All coefficient matrices in Eq. (14) can be obtained in Advanced equipment and measuring instruments are carefully cali­
the frequency-domain calculation (Zhao et al., 2012). brated before the tests to ensure the reliability of the measured data. For
irregular waves, each test ran for more than 22.5 min corresponds to the
3.4. Cables and fenders 3 h. The motion responses of the FSRU and LNGC in 6 degrees of freedom
and the time series of the loads acting on the mooring system and con­
Based on the catenary theory, the cable tension can be calculated necting system were recorded with a sampling frequency of 40 Hz.
with given catenary parameters under the premise that the depth of During the combined wind, wave and current test, the mooring ca­
water and mooring cable wet weight per unit length are known. The bles with similar angle, length and axial stiffness between FSRU and
fender is defined by two contact surfaces and four nodes. In the mooring wharf can be simplified into one new cable. Therefore, a total of 10
system or connecting system, the fenders are located on the pier or simplified cables were used to replace 18 cables in the test. Similarly, the
FSRU. The two contact surfaces are located on the wharf-FSRU 24 connecting cables between FSRU and LNGC were replaced by 10
connection and FSRU-LNGC connection, respectively. The node is used simplified cables. The attributes of the 20 simplified cables are shown in
to define the normal direction of the contact surface. Under the premise Table 4. The length and stiffness of each cable are simulated according to
of known fender performance curve, the fender force can be solved by the principle of similarity to ensure that the mechanical properties of the
giving the fender parameters. model test. The nonlinear spring group and wire rope were used to
simulate the mooring and connecting cables. Taking into account the
nonlinear characteristics of the fenders, the nonlinear spring groups
3.5. Model test set up
were used to simulate them. Fig. 6 indicates mechanical property curve
of the cables and the nonlinear compression curve of the fenders. In
Model tests were conducted at a scale of 1:64 in the State Key Lab­
general, the loads on the fender at the bow and stern is larger than those
oratory of Ocean Engineering (SKLOE) basin at Shanghai Jiao Tong
in the middle of the vessels (Newman, 2001). In the experiment, the
University in China, which consists of static water attenuation test for
bearing capacity of the fender is mainly considered. Therefore, in the
the two vessels respectively (Fig. 4) and combined wind, wave and
connection system, pressure sensors are only installed at the four fenders
current test for the FSRU-LNGC side-by-side mooring system in side-by-
of the vessels at the bow and stern to measure the loads.
side configuration (Fig. 5). The main purpose of the static water atten­
uation test is to obtain the natural period of the FSRU and LNGC which

Fig. 4. Static water attenuation test: (a) shows the full loaded FSRU and (b) shows the ballast LNGC.

5
J. Yue et al. Ocean Engineering 195 (2020) 106731

Table 4
Attributes of the 20 simplified cables.
Simplified cable Original cable Weight in air (kN) Location
number number (kg/m)

Line A1 1/2 2.92 3469.2 FSRU-

Wharf
Line A2 3/4 2.92 3469.2 FSRU-
Wharf
Line A3 5 1.46 1734.6 FSRU-
Wharf
Line A4 6/7 2.92 3469.2 FSRU-
Wharf
Line A5 8/9 2.92 3469.2 FSRU-
Wharf
Fig. 5. Combined wind, wave and flow test of the FSRU-LNGC system in Line A6 10/11 2.92 3469.2 FSRU-
Wharf
the basin.
Line A7 12/13 2.92 3469.2 FSRU-
Wharf
4. Results and discussion Line A8 14 1.46 1734.6 FSRU-
Wharf
Line A9 15 1.46 1734.6 FSRU-
Numerical and experimental study on the dynamic responses of the Wharf
FSRU-LNGC multi-floating mooring system in side-by-side offloading Line A10 16/17/18 4.38 5203.8 FSRU-
operation have been carried out. The results include the six degrees of Wharf
freedom motion of the two vessels, hydrodynamic interactions, the Line B1 41/42 1.762 2214.8 FSRU-
LNGC
forces on the cables and fenders of the mooring and connecting systems.
Line B2 39/40 1.762 2214.8 FSRU-
The numerical results are compared with the experimental results and LNGC
the results obtained by Zhao (Zhao et al., 2013) to verify the correctness Line B3 36/37/38 2.643 3322.2 FSRU-
of the numerical calculation. LNGC
Line B4 33/34/35 2.643 3322.2 FSRU-
LNGC
4.1. Validation of the numerical results Line B5 31/32 1.762 2214.8 FSRU-
LNGC
Validation process of the numerical results is shown in Fig. 7. Based Line B6 29/30 1.762 2214.8 FSRU-
LNGC
on the static water attenuation tests, the convergence analysis of the
Line B7 26/27/28 2.643 3322.2 FSRU-
FSRU and LNGC simulation models is executed first. Then the non- LNGC
dimensional damping parameter of the lid used in the side-by-side off­ Line B8 23/24/25 2.643 3322.2 FSRU-
loading operation numerical analysis is reasonably selected. LNGC
To verify the mesh dependency and convergence, the influences of Line B9 21/22 1.762 2214.8 FSRU-
LNGC
different meshes on the calculation results are considered at first. From Line B10 19/20 1.762 2214.8 FSRU-
Mesh 1 to Mesh 4, 4537, 6383, 8530 and 10452 panels are adopted on LNGC
the FSRU-LNGC side-by-side mooring system respectively. The conven­
tional three-dimensional potential theory with zero μ was used in this
part of analyzes. The amplitude response operator (RAO) and added natural periods of the two vessels obtained from the model test and
mass in the sway mode for the FSRU of 4 different meshes are shown in simulations are in good agreement. This also illustrates that 3589 ele­
Fig. 8. The comparison results show that Mesh 1 has poor mesh quality ments for FSRU, 2743 elements for LNGC and 4119 elements for the
whose corresponding frequency of peak value is inconsistent with the wharf are enough to make the numerical simulation converge.
other meshes, while Mesh 3 and Mesh 4 have better convergence with As can be seen from Fig. 8, there are peak values in some frequency
nearly the same result except for small value discrepancy at the corre­ ranges that are not consistent with the actual situation. This may be
sponding frequency of peak value. In order to ensure accuracy, Mesh 4 is caused by irregular frequency and harmonic resonance. In the process of
used for further calculation and analyzes. The mesh used in this paper is using the panel method, with the increase of the frequency of hydro­
generated in the way shown in Fig. 3. The comparison of natural periods dynamic calculation, the irregular frequency will appear successively.
of FSRU and LNGC are listed in Table 5. The comparison shows that the The influence of irregular frequency on the calculated results can be

Fig. 6. Mechanical property curves: (a) shows the cables and (b) shows the fenders.

6
J. Yue et al. Ocean Engineering 195 (2020) 106731

Table 5
Comparison between experimental data and simulations of the natural periods.
Item Experimental natural periods (s) Numerical natural periods (s)

Roll Pitch Roll Pitch

FSRU (full) 13.2 9.784 13.34 9.84

LNGC (ballast) 11.9 9 11.84 9.03

Table 6
Non-dimensional damping parameters obtained from the experiment.
Item Experimental non-dimensional damping parameters
Fig. 7. Validation process of the numerical results.
Roll Pitch

reduced by increasing the number of model elements, but it cannot be FSRU (full) 0.01 0.422
LNGC (ballast) 0.016 0.239
eliminated fundamentally (Journ�ee and Massie, 2001). In order to
remove the irregular frequency, Teng (Teng and Li, 1996) proposed a
method of applying artificial lid on the inner water surface of the number of elements to minimize the impact of irregular frequency on the
floating body, but the results show that their method does not eliminate calculation results. On this basis, the damping lid method is used to
irregular frequency completely, but pushes irregular frequency towards modify the frequency-domain calculation results to obtain a more ac­
high frequency. Using the integral equation of simple Green’s function curate dynamic response of the multi-floating side-by-side mooring
(Yeung, 1975) or the mixed element method (Bai, 1972) will not lead to system. The dependence and convergence analysis of the mesh is shown
irregular frequency problems. However, their calculation amount will in Fig. 8 and Mesh 4 is selected for further calculation. On this basis, the
be greatly increased. Sun (Sun et al., 2008a, 2008b) proposed a new numerical calculation based on the potential flow theory is revised
method based on the extended integral region, which effectively mainly through the way of the damping lid method. Referring to the roll
removes the irregular frequency in the calculation results of the two and pitch non-dimensional damping parameters listed in Table 6 and
ships that are simplified to the fixed square box. However, its effec­ through the frequency-domain calculation, the non-dimensional damp­
tiveness in the multi-floating side-by-side mooring system without ing parameter of the lid used in the numerical calculation is chosen as
simplification still needs to be verified. 0.018. Fig. 9 shows the effects of the damping lid method on frequency-
When waves of certain frequency acts on two side-by-side floating domain calculation results of FSRU. It can be observed that adding the
bodies with small gaps, a large wave height will appear at the gap (Sun damping lid to the free surface between two ships can effectively reduce
et al., 2008a, 2008b). In the case of resonance, viscosity and energy the excessive peak value in the potential flow calculation at the reso­
dissipation of the fluid cannot be considered by using the nance frequency. The damping lid method simulates the viscous effect of
three-dimensional potential flow theory. The hydrodynamic parameters fluid resonance by adding damping on the free surface between the two
of the floating bodies and the wave surface rise value obtained by the vessels. It can make the numerical results closer to the real dynamic
calculation are too large, resulting in obvious distortion (Xin et al., responses.
2014). In order to reduce the resonance peak value and make the Both numerical and experimental analysis of the FSRU-LNGC system
calculation result closer to the real situation, this part mainly discusses are performed under the combined wind, wave and current. The oblique
the correction of the calculation result of potential flow. The damping lid sea waves will induce both transverse and longitudinal motions of the
method based on the assumption of quasi-ideal fluid is used to add vessels. However, it can be observed from the model tests that the surge
damping on the free water surface between two ships, and the correction and roll motions dominate FSRU-LNGC multi-floating mooring system
of frequency-domain calculation result can be achieved by selecting the responses compared to the motions in other degrees of freedom. Thus
appropriate non-dimensional damping parameter. The damping lid the results of surge and roll motion responses are showed to demonstrate
method was introduced in detail by Chen (2004) and the sensitivity of the dynamic response of the FSRU-LNGC system. Furthermore, the re­
the results for this method was presented by Chen (Chen and Malenica, sults of the cable 5 which is subjected to the largest forces in the mooring
2005) and Chakrabarti (1978). system are introduced in detail. In the experiment, the simplified cables
Therefore, this paper only adopts the method of increasing the of the connecting system all represent two or three original cables and

Fig. 8. Effects of different meshes on sway RAO and sway added mass of FSRU. (a) Sway RAO (b) Sway added mass.

7
J. Yue et al. Ocean Engineering 195 (2020) 106731

Fig. 9. Effects of damping lid method on frequency-domain calculation results of FSRU.

Fig. 10. Comparisons between numerical and experimental results of motion responses for the FSRU under combined wind, wave and current: (a) shows the surge
motions and (b) shows the roll motions.

Fig. 11. Comparisons between numerical and experimental results of motion responses for the LNGC under combined wind, wave and current: (a) shows the surge
motions and (b) shows the roll motions.

the forces of the original single connecting cable cannot be measured. response spectrums that the surge motions of the two vessels concentrate
Therefore, only the comparison of the most stressed cable group is given. on low frequencies, while the peak frequency of the roll motions is
Comparisons between the numerical and experimental results are pre­ located in the 0.75 rad/s, close to the natural frequencies of the two
sented by response spectrums. vessels. However, small discrepancies can be observed between the
The numerical and experimental comparisons for the FSRU and numerical and experimental results as listed in Table 7. In general, the
LNGC are presented in Fig. 10 and Fig. 11, respectively. It is shown that surge and roll motions for both the FSRU and LNGC estimated by the
the surge and roll motion responses obtained by numerical analysis are numerical analysis are a little bit larger than the test results. This may be
in good agreement with the test results. The variation trend and peak caused by the fact that in the numerical analysis, the damping lid
frequency of response spectrums of surge and roll motions are consistent method (Fig. 3) adds a linear damping term in free surface, but the
with the results of Zhao (Zhao et al., 2013). It can be seen from the actual natural damping should be nonlinear.

8
J. Yue et al. Ocean Engineering 195 (2020) 106731

Table 7 moored FSRU and LNGC under combined wind, wave and current, the
Comparison between numerical and experimental statistics. time series of the six degrees of freedom motion responses of the two
Item Max. Min. Mean Std. vessels from the proposed numerical analysis are shown in Fig. 13. Some
basic statistics of the time series responses are listed in Table 8.
FSRU Surge (m) Experiments 0.24 0.72 0.14 0.11
Simulations 0.29 0.67 0.07 0.12 As can be seen in Fig. 13, motion responses on the LNGC are in
FSRU Roll (deg) Experiments 0.37 0.51 0.05 0.08 general larger than on the FSRU, especially in surge, sway and roll
Simulations 0.38 0.50 0.03 0.08 motions. Moreover, the amplitude of the surge responses of the LNGC is
LNGC Surge (m) Experiments 0.18 1.52 0.39 0.22 much larger than that of FSRU as shown in Table 8. For the comparision
Simulations 0.22 1.57 0.42 0.24
LNGC Roll (deg) Experiments 1.45 1.36 0.03 0.30
of the roll motion, it is obvious that the amplitude of the motion
Simulations 1.50 1.32 0.09 0.30 response of the FSRU is so small that it can be negligible. This may be
Line A3/Cable 5 (kN) Experiments 481.91 112.60 187.89 39.08 related to the big differences of principal dimensions between the two
Simulations 445.22 117.60 192.29 37.46 floating vessels. The principal dimensions of FSRU are much larger than
Line B7/Cables 26/27/ Experiments 1588.19 18.90 582.22 199.64
those of LNGC. Furthermore, the FSRU is designed as a floating structure
28 (kN) Simulations 1600.79 0.13 492.74 209.59
with larger block coefficient and displacement (also inducing large
added mass) than LNGC. Thus, there is a smaller motion response on
Comparisons of the forces estimated from the numerical analysis and FSRU than LNGC under the same sea condition. According to Fig. 13, it
experimental tests are presented in Fig. 12 for the cable 5 in mooring can be observed that the response frequencies of six degrees of freedom
system (left plot) and the cable group consists of cables 26/27/28 in are quite different. The response frequency of surge motion is the lowest
connecting system (right plot). The forces estimated by the numerical compared with other motion modes such as sway, heave and roll. This
analysis are well consistent with measurements from the tests. It can be can be explained by the effect of the cables in both mooring system and
known from Fig. 12 that the dynamic tension of cable 5 in the mooring connecting system, which can increase the response period of surge
system mainly occurs in the low frequencies similar as the surge and roll motion and reduce its response frequency.
motions of the FSRU in Fig. 10. The dynamic tension of cables 26/27/28
mainly occurs around the wave frequencies also similar as the roll mo­ 4.2.2. Relative motion
tions of FSRU and LNGC. As can be seen in Table 7, there are small During the side-by-side offloading operation, the FSRU and LNGC are
discrepancies of the statistical motion and force responses between the close to each other and are prone to collision. Furthermore, the relative
numerical analysis and the experimental tests. Both the motion and motion of the two ships has a great influence on the force acting on the
force responses obtained from the numerical analysis are larger than cables and fenders of the connecting system. In order to ensure the
that of the experiments. The discrepancies might be caused by that the operational safety of the cables and fenders, the relative motion between
numerical analysis uses the original 42 cables from the original mooring the two ships should be seriously investigated. Additionally, heave, roll
design, while in the experiment test these 42 cables are simply replaced and pitch motions have a great impact on the safety performance of side-
by 20 cables as shown in Table 4. by-side offloading operation, and they should be carefully analyzed as
In general, the numerical analysis and experimental tests show well. Therefore, on the premise of analyzing the six degrees of freedom
satisfactory agreement. The numerical calculation method and model motion of FSRU and LNGC respectively, the relative heave, roll and pitch
tests used in this study are feasible, and the damping coefficient selected motions between the two vessels are investigated with FSRU as the
based on experimental data is reasonable. reference object.
Fig. 14 presents the relative heave, roll and pitch motions between
the FSRU and LNGC. According to Fig. 14, the amplitude of the relative
4.2. Dynamic responses of FSRU-LNGC side-by-side mooring system
heave and pitch motions is similar and that of the relative roll motion is
twice as large. This indicates that the relative roll motion between the
Based on the numerical model and the damping parameters that are
two vessels is much more important than the relative heave and pitch
validated above, the dynamic responses of the FSRU-LNGC side-by-side
motions. However, the relative motion response frequencies in three
mooring system under the specific sea condition are analyzed in this
directions are similar. The similar phenomenon could be found when the
section. The investigation of the response characteristics can be also
motion responses of FSRU and LNGC are analyzed separately. Through
used in the collision analysis between the two floating bodies.
the relative heave, roll and pitch motions analysis of the two vessels, it
can be concluded that relative roll motion dominates the dynamic re­
4.2.1. Ship motions in six degrees of freedom
sponses of the connecting system between two vessels under the given
In order to analyze the motion characteristics of the side-by-side

Fig. 12. Comparisons between numerical and experimental results of the forces on the cables under combined wind, wave and current: (a) shows the cable 5 and (b)
shows the cables 26/27/28.

9
J. Yue et al. Ocean Engineering 195 (2020) 106731

Fig. 13. Time series of the six degrees of freedom motions of the FSRU and LNGC.

sea condition.
Table 8
Summery of motion statistics of the FSRU and LNGC.
4.2.3. Cables and fenders
Item Max. Min. Mean Std. The FSRU-LNGC multi-floating mooring system contains multiple
Surge(m) LNGC 0.22 1.57 0.42 0.24 cables and fenders. The maximum force acting on the cables and fenders
FSRU 0.29 0.67 0.07 0.12 is of great significance to the safety of side-by-side offloading operation.
Sway(m) LNGC 0.81 0.40 0.32 0.12
The side-by-side configuration of the multi-floating mooring system is
FSRU 0.54 0.03 0.31 0.05
Heave(m) LNGC 0.22 0.89 0.28 0.12 described in detail in Fig. 1. The FSRU-wharf mooring system consists of
FSRU 0.09 0.49 0.27 0.04 18 cables and 4 fenders, of which from the bow to the stern the cable
Roll(deg) LNGC 1.50 1.32 0.09 0.30 number is #1~#18 and the fender number is #43~#46. Similarly, the
FSRU 0.38 0.50 0.03 0.08 FSRU-LNGC connecting system consists of 24 cables and 8 fenders, of
Pitch(deg) LNGC 0.46 0.44 0.00 0.10
which from bow to stern the cable number is #42~#19 and the fender
FSRU 0.16 0.17 0.00 0.04
Yaw(deg) LNGC 0.61 0.49 0.03 0.10 number is #54~#53. Fig. 15 and Fig. 16 show the statistics of maximum
FSRU 0.18 0.21 0.03 0.04 cable tension and maximum fender forces under the combined wind,
wave and current loads, respectively.
Fig. 15 shows that the cables arranged in different positions bear
different forces, and those placed in similar locations experience similar
loads. This can be explained by its similar angles, length and stiffness.

10
J. Yue et al. Ocean Engineering 195 (2020) 106731

Fig. 14. Time series of the relative motions between FSRU and LNGC.

Fig. 15. Maximum tension statistics of cables: (a) shows the cables in mooring system and (b) shows the cables in connecting system.

The maximum force of the connecting cable is 529.4 kN and that of the maximum force acting on the fenders of connecting system and mooring
mooring cable is 445.22 kN. The maximum force of the connecting cable system are 11400.2 kN and 1300 kN, respectively. They are smaller than
is greater than that of the mooring cable, which is consistent with the their SWL. The maximum force acting on the fenders of connecting
results shown by Zhao (Zhao et al., 2013). This may be due to the severe system is much larger than that of mooring system. This may be due to
relative motion between the two vessels. However, the largest forces the complex interaction of hydrodynamic force and relative motions
acting on both mooring cables and connecting cables are much smaller between FSRU and LNGC. For a more specific analysis of the fender
than their SWL under the given sea state. The safety margin of the cable forces in the connection system, Fig. 17 shows the time series of the
allows the FSRU-LNGC side-by-side mooring system to operate safely in fenders with large force at bow and stern in the connection system,
worse sea conditions. respectively. There are many intervals in the time series of the forces
According to Fig. 16, it is obvious that the difference of estimated acting on fenders, and the number of intervals occurring at the bow is
forces among the fenders of the mooring system is small, while that in larger than at the stern. This illustrates that there are many collisions
the connecting system is large. In the connecting system, the force acting between the bow and stern of two ships, and the fender at the bow ex­
on the fender 53 in the bow is much larger than the fender 48 in the periences more collisions than the fender at the stern.
stern, while other fender forces are negligible compared to them. The

11
J. Yue et al. Ocean Engineering 195 (2020) 106731

of freedom. It can be observed from Table 10 that the motion responses

can be significantly affected by the pretension of the connecting cables.
Both the maximum value and the amplitude of LNGC and FSRU motion
responses increase as the pretension of the connecting cables decrease.
One can also concluded from Table 10 that the influence of the con­
necting cables pretension on the motion amplitude of the vessels is
nonlinear. The motion response of LNGC is much more sensitive than
that of FSRU to the pretension of the connecting cables. When the length
of the connecting cables increases from þ2% to þ3%, the roll motion of
the two floating bodies significantly changes, which indicates that when
the connecting cable length is within this range, increasing the preten­
sion of the connecting cables can significantly improve the roll motion
characteristics of the two floating bodies. The large-amplitude roll mo­
tion of the vessels will affect the normal life of the crew and the safety
performance of the side-by-side offloading operation of the system. The
vessel will exhibit complex nonlinear dynamic behavior under large-
Fig. 16. Maximum force statistics of fenders in the FSRU-LNGC system. amplitude roll motion. Serious large-amplitude roll motion can cause
the vessel to capsize under the wave excitation or other excitation
disturbance. The severe roll motion of the two floating bodies in the
5. Sensitive analysis
multi-floating mooring system can be improved by increasing the pre­
tension of the connecting cables. Furthermore, the heave, roll and pitch
Due to the complex hydrodynamic interaction between the two
motions are more sensitive than that of the other degrees of freedom.
vessels, the pretension of the connecting cables has a great influence on
Therefore, based on the analysis of the influence of the pretension of the
the dynamic responses of the mooring system. The sensitivity analysis of
connecting cables on the six degrees of freedom motion response of the
the pretension of the connecting cables is of great significance to the
two ships, the sensitivity analysis of relative heave, roll and pitch be­
safety performance of the side-by-side offloading operation. Based on
tween the two ships are carried out.
the analysis of the dynamic responses of the FSRU-LNGC side-by-side
mooring system, the sensitivity study focuses on the investigation of the
connecting cable pretension on the six degrees of freedom motion, 5.2. Effects on the relative motions
relative motions, cable tension and fender forces of the two floating
bodies are presented. Fig. 18 shows the time series of the relative heave, roll and pitch
According to the guidelines (OCIMF, 2008), each cable of the system motions between the FSRU and LNGC. It can be observed from the figure
should maintain the same pretension, typically taking 10% of its that the relative heave, roll and pitch motions increase as the pretension
breaking force. This is easy to implement for a symmetrical arrangement of the connecting cables decrease. In order to make quantitative anal­
such as the single point mooring. However, the multi-floating mooring ysis, the statistics of the relative motions are illustrated in Table 11. As
system in this paper adopts wharf mooring, and the side-by-side shown in Table 11, the decrease of the pretension on the connecting
arrangement is not completely symmetrical, so it is difficult to achieve cables can lead to nonlinearly increase of both the maximum values and
the same pretension of 24 cables. Therefore, the pretension is changed amplitudes of the relative motions. The relative motions of the case 1 do
by changing the length of the connecting cables, and a certain propor­ not differ too much from that of the case 2, but obvious difference can be
tion is added to the original cable length. As the length of the connecting found from the case 2 to the case 3. Similar to the effect of the pretension
cables increases, the pretension decreases. The cases selected for the
sensitivity study are listed in Table 9. The other parameters not listed in
Table 9
Table 9 keep the same as in the base case.
Sensitivity cases configurations.
Case Connecting cable number Proportion
5.1. Effects on the ship motions in six degrees of freedom
Case 1 #19-#42 þ1%
Case 2 #19-#42 þ2%
For the motion response analysis of the vessels, besides the maximum Case 3 #19-#42 þ3%
value of motion, the amplitude of motion also needs to be investigated, Base case #19-#42 0
because it can reflect the range of motion of the ship on a certain degree

Fig. 17. Time series of the fenders with maximum force in connecting system: (a) shows fender 53 at stern and (b) shows fender 48 at bow.

12
J. Yue et al. Ocean Engineering 195 (2020) 106731

Table 10
Summary of the motion statistic in different pretension cases.
Item Case Max. Min. Mean Amplitude

LNGC FSRU LNGC FSRU LNGC FSRU LNGC FSRU

Surge(m) Case 1 0.24 0.31 1.50 0.67 0.40 0.08 1.74 0.98
Case 2 0.25 0.35 1.59 0.69 0.38 0.08 1.84 1.04
Case 3 0.42 0.36 1.62 0.72 0.41 0.09 2.04 1.08
Sway(m) Case 1 0.83 0.56 0.20 0.03 0.30 0.29 1.03 0.59
Case 2 0.85 0.62 0.30 0.01 0.32 0.36 1.15 0.61
Case 3 1.12 0.55 0.26 0.08 0.36 0.33 1.38 0.63
Heave(m) Case 1 0.22 0.09 0.87 0.5 0.25 0.28 1.09 0.41
Case 2 0.32 0.06 0.86 0.56 0.31 0.25 1.18 0.50
Case 3 1.92 0.89 0.99 0.64 0.27 0.22 2.91 1.53
Roll(deg) Case 1 1.56 0.41 1.23 0.54 0.10 0.05 2.79 0.95
Case 2 1.75 0.45 0.85 0.62 0.42 0.12 2.60 1.07
Case 3 7.43 1.03 6.84 1.8 0.06 1.31 14.27 2.83
Pitch(deg) Case 1 0.46 0.18 0.42 0.19 0.00 0.02 0.88 0.37
Case 2 0.48 0.18 0.45 0.23 0.02 0.01 0.93 0.41
Case 3 0.52 0.20 0.52 0.34 0.01 0.00 1.04 0.54
Yaw(deg) Case 1 0.62 0.21 0.48 0.21 0.02 0.01 1.10 0.42
Case 2 0.62 0.24 0.53 0.19 0.00 0.01 1.15 0.43
Case 3 0.69 0.28 0.55 0.15 0.01 0.02 1.24 0.43

Fig. 18. Time series of the relative motions between FSRU and LNGC in different pretension conditions.

of the connecting cables on ship motions in six degrees of freedom, when the connecting cables would induce severer relative motions between
the length of the connecting cable is increased from þ2% to þ3%, the the floating vessels. Furthermore, it can be seen from Table 11 that the
relative motions of the two floating bodies obviously increases, espe­ relative heave and roll are much more sensitive than that of the relative
cially the relative roll motion. This indicates that increasing the pre­ pitch.
tension of the connecting cables can significantly improve the relative
motion characteristics between the two floating bodies when the length
of the connecting cable is within this range. The severe relative roll 5.3. Effects on the forces acting on the cables and fenders
motion between the two floating bodies in the multi-floating system can
be improved by increasing the pretension of the connecting cables. The Table 12 shows the maximum forces acting on the representative
standard deviations of the relative heave, roll and pitch motions increase cables and fenders in both the mooring system and the connecting sys­
as the pretension decrease. It means that decrease of the pretension of tem under different pretension conditions. As shown in the table, the
most stressed cable in the mooring system has always been cable #5,

13
J. Yue et al. Ocean Engineering 195 (2020) 106731

Table 11 pretension of the connecting cables.

Summary of the relative motion statistic in different pretension cases.
Item Case Max. Min. Mean Std. Amplitude 6. Conclusions
Relative heave(m) Case 1 0.22 0.89 0.29 0.12 1.11
Case 2 0.27 1.55 0.78 0.25 1.82 In this paper, the damping lid method based on fairly perfect fluid
Case 3 1.20 1.57 0.17 0.33 2.77 hypothesis is used to simulate the hydrodynamic interaction between
Relative roll(deg) Case 1 1.45 1.36 0.03 0.29 2.81 the two floating bodies. The time-domain coupled model of FSRU-LNGC
Case 2 1.68 0.82 0.37 0.33 2.52 side-by-side mooring system is established. The results between nu­
Case 3 7.17 7.74 0.01 1.63 14.91
Relative pitch(deg) Case 1 0.40 0.41 0.01 0.08 0.81
merical results and model tests are compared to validate the correctness
Case 2 0.45 0.4 0.00 0.10 0.85 of the numerical calculation. The analysis of dynamic responses of the
Case 3 0.49 0.48 0.00 0.12 0.97 FSRU-LNGC side-by-side mooring system including the six degrees of
freedom motion, relative motions, loads on cables and fenders is pro­
vided. The sensitivity study which is focused on the effects of the con­
Table 12 necting cable’s pretension is developed. A safe and reliable
Summary of the statistics of the forces in different pretension cases. configuration of FSRU-LNGC multi-floating mooring system and the
Item Case cable Max. Min. Mean Std.
prediction of the dynamic responses of the system is obtained in this
number study. The design of side-by-side configuration and the numerical
calculation method can provide reference for the dynamic response
Mooring cable Case #5 452.32 110.34 183.42 38.25
(kN) 1 study of similar multi-floating mooring system.
Case #5 466.8 121.67 198.71 45.27 Based on the dynamic responses analysis and sensitivity study re­
2 sults, as well as some guidelines proposed for further studies of mooring
Case #5 472.21 118.4 120.56 60.32 system design, we can further conclude:
3
Mooring Case #46 1308.41 5.62 202.92 202.68
fenders(kN) 1 (1) Under the given sea condition, the FSRU-LNGC side-by-side
Case #46 1321.23 6.81 204.83 200.19 configuration designed in this paper meets the safety
2 requirements.
Case #46 1352.89 7.99 208.72 189.72
(2) Compared with FSRU, LNGC has more severe motion responses.
3
Connecting Case #26 570.1 8.94 200.49 150.1 Overall, the forces acting on the cables and fenders of the con­
cables(kN) 1 necting system are greater than that of the mooring system. In the
Case #24 624.43 14.61 208.51 149.62 connecting system, there are many collisions between the bows of
2 the two ships, as well as between the sterns. The fenders at the
Case #39 601.53 26.98 245.81 142.85
3
bow experiences more collisions than the fender at the stern.
Connecting Case #53 11513.4 25.71 1108.93 759.3 (3) The pretension of the connecting system can significantly affect
fenders(kN) 1 the dynamic response of the FSRU-LNGC mooring system in side-
Case #53 11526.72 30.1 1148.72 775.42 by-side offloading operation. The severe roll motion of the two
2
floating bodies and the severe relative roll motion between the
Case #53 11575.98 40.8 1231.01 800.51
3 two floating bodies in the multi-floating mooring system can be
improved by increasing the pretension of the connecting cables.
(4) The effects of the pretension of the connecting system on the
there is not much difference in the maximum forces on cable #5 in the dynamic responses of the mooring system in side-by-side off­
different pretension cases. However, the standard deviations of the cable loading operation are nonlinear. The severe collision between the
#5 forces increase as the pretension decrease, which is consistent with two floating bodies in the multi-floating mooring system can be
the trend of the relative motions as shown in Table 11. Similarly, the improved by increasing the pretension of the connecting cables.
fender with the largest force in the mooring system is fender #35 at the (5) In this paper, the influence of irregular frequency on the calcu­
bow. There is not much difference in the maximum forces on fender #35 lation results is reduced by increasing the number of elements,
when the pretension of the connecting cables get smaller, while the but the irregular frequency is not completely removed. In the
standard deviations decrease. This means that the decrease of the pre­ future research, corresponding methods will be considered to
tension reduces the probability of the collision events between FSRU and remove the irregular frequency in the calculation of multi-
the wharf. floating body side-by-side mooring system.
It can be observed from Table 12 that the most stressed cable in the
connecting system changes in different pretension conditions. The Declaration of competing interest
maximum forces acting on the connecting cables has no specific change
regulation. This is due to the complex hydrodynamic interaction be­ The authors declare that they have no known competing financial
tween FSRU and LNGC under the action of wind, wave and current, interests or personal relationships that could have appeared to influence
which will also lead to complex relative motions. Moreover, ship type the work reported in this paper.
difference and asymmetric arrangement of connecting cables would
change the whole force system of connecting cables. The maximum Acknowledgments
forces acting on the fender #53 increase when the pretension decrease,
so as the standard deviations. This indicates that the decrease of the This work was financially supported by the National Natural Science
pretension increases the maximum loads on the connecting fenders, as Foundation of China (Project no. 51679177), the Key International
well as the probability of the collision events between FSRU and LNGC. (Regional) Cooperative Research Projects of the National Natural Sci­
In addition, it can be seen from Table 12 that the standard deviation of ence Foundation of China (Project no. 51720105011). These sources of
fender #53 is much larger than that of fender #46, which indicates that support are gratefully acknowledged by the authors.
the collision between two ships is much more severe than that between
FSRU and wharf. The severe collision between the two floating bodies in
the multi-floating mooring system can be improved by increasing the

14
J. Yue et al. Ocean Engineering 195 (2020) 106731

References ASME 2007 26th International Conference on Offshore Mechanics and Arctic
Engineering. American Society of Mechanical Engineers, pp. 801–811.
Newman, J.N., 2001. Wave effects on multiple bodies. Hydrodyn. Ship Ocean Eng. 3,
Bai, K.J., 1972. A Variational Method in Potential Flows with a Free Surface. Ph.D.
3–26.
Dissertation, Department of Naval Architecture. University of California, Berkeley.
Newman, Nicholas, John, 1977. Marine Hydrodynamics. MIT press, Cambridge.
Buchner, Bas, Dijk, Van, Adri, Jaap De, Wilde, 2001. Numerical multiple-body
OCIMF, 2008. Mooring Equipment Guidelines. Oil Companies International Marine
simulations of side-by-side mooring to an FPSO. In: Proceedings of the Eleventh
Forum, London.
International Offshore and Polar Engineering Conference. International Society of
Pessoa, J., Fonseca, N., Soares, C.G., 2016. Side-by-side FLNG and shuttle tanker linear
Offshore and Polar Engineers.
and second order low frequency wave induced dynamics. Ocean Eng. 111, 234–253.
Chakrabarti, S.K., 1978. Comments on second-order wave effects on large-diameter
Sun, L., Teng, B., Liu, C.F., 2008. Removing irregular frequencies by a partial
vertical cylinder. J. Ship Res. 22.
discontinuous higher order boundary element method. Ocean Eng. 35, 920–930.
Chen, X.B., 2004. Hydrodynamics in Offshore and Naval Applications-Part I. Keynote
Sun, L., Taylor, P.H., Eatock Taylor, R., 2008. First and second order wave effects in
Lecture in the 6th International Conference on HydroDynamics, Perth (Australia).
narrow gaps between moored vessels. Marine operations specialty symposium,
Chen, X.B., Malenica, S., 2005. Interaction hydrodynamique d’un ensemble de flotteurs
pp. 113–124.
sur la surface libre. In: 10�emes Journ�
ees de l’Hydrodynamique, pp. 1–14 (Nantes,
Teng, B., Li, Y.C., 1996. A unique solvable higher order BEM for wave diffraction and
France).
radiation. China Ocean Eng. 10 (3), 333–342.
Hong, S.Y., Kim, J.H., Cho, S.K., et al., 2005. Numerical and experimental study on
Vieira, D.P., Mello, P.C.D., Dotta, R., et al., 2018. Experimental investigation on the
hydrodynamic interaction of side-by-side moored multiple vessels. Ocean Eng. 32
influence of the liquid inside the tanks in the wave behavior of FLNG vessels in side-
(7), 783–801.
by-side offloading operations. Appl. Ocean Res. 74, 28–39.
Hong, Yong-Pyo, Wada, Yojiro, Choi, Yong-Ho, et al., 2009. An experimental and
Xin, X., Yang, J.M., Xin, L., et al., 2014. Hydrodynamic performance study of two side-
numerical study on the motion characteristics of side-by-side moored LNG-FPSO and
by-side barges. Ships Offshore Struct. 9 (5), 475–488.
LNG carrier. In: Proceedings of the Ninth International Offshore and Polar
Yangshan, Dai, Jinwei, Shen, Jingzheng, Song, 2007. Ship Wave Load [M]. National
Engineering Conference. International Society of Offshore and Polar Engineers.
Defence Industry Press.
Huang, W., Li, B., Chen, X., et al., 2018. Numerical and experimental studies on dynamic
Yeung, R.W., 1975. A hybrid integral-equation method for the time-harmonic free-
gangway response between monohull flotel and FPSO in non-parallel side-by-side
surface flows. In: Proceeding: First International Conference on Numerical Ship
configuration[J]. Ocean Eng. 149, 341–357.
Hydrodynamics. Maryland, Gaithersburg, pp. 581–607.
Inoue, Yoshiyuki, Islam, M.R., 1999. Comparative study of numerical simulation and the
Zhao, W.H., Yang, J.M., Hu, Z.Q., 2012. Hydrodynamic interaction between FLNG vessel
experimental results for a parallely connected FPSO and LNG in waves. In:
and LNG carrier in side by side configuration. J. Hydrodyn. Ser. B 24 (5), 648–657.
Proceedings of the Ninth International Offshore and Polar Engineering Conference.
Zhao, W., Yang, J., Hu, Z., et al., 2013. Hydrodynamics of an FLNG system in tandem
International Society of Offshore and Polar Engineers.
offloading operation. Ocean Eng. 57 (Complete), 150–162.
Journ�ee, J.M.J., Massie, W.W., 2001. Offshore Hydromechanics, first ed. Delft University
Zhao, D., Hu, Z., Chen, G., 2017. Experimental investigation on dynamic responses of
of Technology, p. 6.
FLNG connection system during side-by-side offloading operation. Ocean Eng. 136,
Ministry of Communications of the People’s Republic of China, 2009. Design
283–293.
Specification for LNG Terminal [S]. China Communications Press, Beijing.
Naciri, Mamoun, Waals, Olaf, de Wilde, Jaap, 2007. Time-domain simulations of side by-
side moored vessels: lessons learnt from a benchmark test. In: Proceedings of the

15