Ocean Engineering 137 (2017) 367–381

Contents lists available at ScienceDirect

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

A study on prediction of ship maneuvering in regular waves MARK

a,b b,c,⁎ b
Wei Zhang , Zao-Jian Zou , De-Heng Deng
a
School of Petroleum Engineering, China University of Petroleum (East China), Qingdao 266580, China
b
School of Naval Architecture, Ocean and Civil Engineering, Shanghai Jiao Tong University, Shanghai 200240, China
c
State Key Laboratory of Ocean Engineering, Shanghai Jiao Tong University, Shanghai 200240, China

A R T I C L E I N F O A BS T RAC T

Keywords: A simulation method is developed for predicting ship maneuvering in regular waves. Based on a two-time scale
Ship maneuvering in waves model, the total ship motion is divided into the low frequency maneuvering motion and the high frequency
Wave drift force wave-induced motion. The maneuvering analysis is based on a MMG model which takes the mean second-order
Two-time scale model wave loads into account. In order to evaluate the second-order wave loads, a velocity potential is introduced and
Rankine panel method
decomposed into a basic part and a perturbation part, which are related to the maneuvering motion and the
Time domain analysis
wave-induced motion, respectively. The basic part is evaluated based on the double-body model with a trailing
vortex sheet, while the perturbation part is solved via a time domain Rankine panel method. The effects of
maneuvering motion on the wave forces are considered through the m-terms as well as the leading-order terms
kept in the boundary conditions on the free surface. By using the proposed method, turning and zig-zag
maneuvers of the S-175 container ship in regular waves are simulated. The predicted turning trajectories and
10°/10° and 20°/20° zig-zag maneuvers are compared with the experimental data, which show fairly good
agreements. The drift forces and moment on the ship turning in waves are also discussed.

1. Introduction turning and zig-zag maneuvers. Yasukawa (2006, 2008) conducted

Prediction of ship maneuverability is typically carried out under
calm water condition. This provides valuable information at the ship
design stage. However, an actual seagoing ship usually maneuvers in
the presence of waves. From the viewpoint of ship navigation safety, it
is meaningful to understand the maneuvering behavior of a ship in
waves. According to the report of ITTC Maneuvering Committee (2011)
and the literature review given by Tello Ruiz et al. (2012), the existing
methods for investigating ship maneuvering in waves can be generally
classified as experimental methods, simulation methods based on two-
time scale models, simulation methods based on hybrid approach, and
simulation methods using CFD.
Experiment methods are supposed to be the most reliable method
to investigate ship maneuvering in waves. Several experimental
investigations have been published. For example, Ueno et al. (2003)
carried out turning, zig-zag and stopping tests in regular waves using a
VLCC model, and discussed the effects of wave length and encounter
angle to waves as well as the effect of loading condition on maneuvering
motion based on the experimental data. Lee et al. (2009) investigated
the effects of the wave amplitude and wave length on the maneuver-
ability of a ship, with tests on a KVLCC model. The results showed that
second-order wave force has a dominant influence on the trajectory for
turning tests as well as zig-zag and stopping maneuvers with the SR108
ship model, and compared the model test results with their numerical
results.
CFD methods in principle provide an adequate description of all
physics. However, this approach is still considered as a subject of the
state-of-the-art research rather than application in engineering practice.
On one hand, these methods require large computational resources and
long CPU time. On the other hand, according to Skejic (2013), a lot of
technical difficulties concerning the analysis of ship maneuvering in
waves are still unsolved, such as the adequacy of the selected turbulence
models, the adequacy of the propeller and rudder models under large
angle of attack, the appearance of the crossflow shed vortices and so
forth.
Compared with CFD methods, the hybrid methods and the two-
time scale methods are much more widely applied in numerical
simulation of ship maneuvering in waves. Both of these two methods
are based on the potential flow theory but they are different in specific
treatments:
The hybrid approach integrates maneuvering motion and wave-
induced motion into a generic set of rigid body motion equations to
describe the ship maneuvering in waves. Several works falling into this
category can be found in the literature, for instance, Bailey et al.

⁎
Corresponding author at: School of Naval Architecture, Ocean and Civil Engineering, Shanghai Jiao Tong University, Shanghai 200240, China.
E-mail address: zjzou@sjtu.edu.cn (Z.-J. Zou).

http://dx.doi.org/10.1016/j.oceaneng.2017.03.046
Received 28 November 2015; Received in revised form 1 January 2017; Accepted 24 March 2017
Available online 20 April 2017
0029-8018/ © 2017 Elsevier Ltd. All rights reserved.
W. Zhang et al. Ocean Engineering 137 (2017) 367–381

Nomenclature v′ Non-dimensional transverse speed

wP0 Wake fraction in straight moving
A Amplitude of the incident wave wP Wake fraction
aH Rudder force increase factor ⎯⇀
W Velocity vector due to the low frequency motion
Cf Frictional resistance coefficient xG x-coordinate of the center of gravity
Cr Residuary resistance coefficient x′P Non-dimensional longitudinal coordinate of propeller
CRr , CRrrr , CRrrv Experimental constants for expressing the inflow xR Longitudinal coordinate of the rudder
velocity to rudder xH Longitudinal coordinate of acting point of the additional
[Cij ] Matrix of hydrostatic restoring coefficients lateral force induced by steering
Dp Propeller diameter XH , YH , NH Low frequency hydrodynamic surge force, lateral force,
Fi High frequency hydrodynamic components yaw moment acting on ship hull except added mass
⇀(2) components
F Generalized second-order wave force vector
FN Normal rudder force XP Surge force due to propeller
G Rankine source XR , YR, NR Surge force, lateral force, yaw moment by rudder
Izz Moment of inertia about z axis XW , YW , NW Wave drift forces and moment
JP Propeller advance ratio αR Relative inflow angle to rudder
Jzz Added moment of inertia about z axis βP Geometrical inflow angle to the propeller
kf Form factor χ Incident angle of wave defined in the global frame
k Wave number of the incident wave δ Rudder angle
KT Thrust coefficient ε, κ , τ , γ Coefficients representing the interaction among the hull,
L Ship length propeller and rudder
m Ship mass ϕ(⇀ x , t ) Perturbation potential
mx , m y Added masses in x-axis direction and y-axis direction, Φ(⇀ x , t ) Basic potential
respectively η Wave elevation
mi m-terms λ Wave length
[Mij ] Inertial matrix for the hull φI (⇀x , t ) Incident wave potential
n Propeller revolution φ(⇀ x , t ) Perturbation potential except incident wave potential
⇀n (0), ⇀
n (1), ⇀n (2) Zero-, first- and second-order components of the φI (⇀
x , t)
normal vector on the hull surface, respectively Λ Rudder aspect ratio
ni Component of the generalized normal vector ρ Water density
Nt Ratio of τL to τH τL Time step for the maneuvering computation
o − xyz Reference frame with the origin at midship τH Time step for the seakeeping computation
O − XYZ Global frame fixed in space υ Damping strength of the numerical damping beach
p(1) First-order hydrodynamic pressure ω Frequency of the incident wave
r Yaw rate of the low frequency motion ωe ‘Local’ encounter frequency during maneuvering process
r′ Non-dimensional yaw rate of the low frequency motion ⇀
ξT Translational displacements due to high frequency motion
R (u ) Ship resistance ⇀
ξR Rotational displacements due to high frequency motion
S Wetted hull surface ψ Ship heading angle
SB Mean wetted hull surface Ψ (⇀x , t ) Total velocity potential
t Time ζ (x, y, t ) Wave elevation except incident wave elevation ζI (x, y, t )
tp Thrust deduction fraction ζI (x, y, t ) Incident wave elevation
tR Steering thrust deduction factor ξz Vertical displacement of point (x, y) due to high frequency
u Forward speed of the low frequency motion motion
UR Relative inflow velocity to rudder
v Transverse speed of the low frequency motion

(1997), Fang et al. (2005), Sutulo and Guedes Soares (2006), and Lin
et al. (2006). A defect common to all these works is that none of them
incorporated the effects of second-order wave forces. The omission of
the steady components of the second-order wave forces will heavily
decrease the accuracy of the maneuvering predictions. Recently,
Subramanian and Beck (2015) developed a body-exact strip theory
based model to simulate maneuvering of a ship in a seaway. In their
methodology, the second-order wave effects are considered approxi-
mately via the square of the velocity potential's gradient.
Different from the hybrid approach, the two-time scale methods
separate the total ship motion into two parts: the one for high
frequency wave-induced motion and the other for the low frequency
maneuvering motion, based on the fact that the wave-induced motions
are generally much faster than the maneuvering motion. The motion
separation brings in a favorable merit that the second-order wave
forces can be evaluated in a more accurate manner. Within the
framework of the two-time scale model, Skejic and Faltinsen (2008)
carried out a unified seakeeping and maneuvering analysis of ships in
regular waves and compared the wave drift forces obtained by using
four different approaches based on strip theory. Yasukawa and
Nakayama (2009) conducted the 6-DOF motion simulations of a
turning ship in regular waves. The high-frequency wave forces in their
study were estimated based on the strip theory, while the wave drift
forces were evaluated by Maruo's method. It should be noted that the
seakeeping analysis in these two works was carried out based on the
quasi-steady assumption, which means the memory effects due to the
ship motion were not considered. Seo and Kim (2011) conducted the
numerical analysis of the coupled maneuvering and seakeeping pro-
blem. In their study, two different time scales were used in solving the
maneuvering and seakeeping problems, respectively; while the wave
forces were solved directly in time domain by a 3D Rankine panel
method.
In this paper, a numerical model for investigating ship maneuver-
ing in waves is developed. The simulation method is based on the
procedure suggested by Seo and Kim (2011), i.e. the maneuvering
motion is predicted using MMG model, whereas the wave effects are

368
W. Zhang et al. Ocean Engineering 137 (2017) 367–381

determined by solving a linearized boundary value problem (BVP) in
time domain. However, the BVP for the wave potential in the present
study is different: The linearization is carried out based on the double-
body model with the trailing vortex sheet, while Seo and Kim (2011)
used the Neumann-Kelvin linearization. Such modification aims at
taking into account the effects of the trailing vortices on the wave drift
forces, and subsequently increasing the accuracy of the numerical
predictions.
In the previous study (Zhang and Zou, 2016), the developed model
has been used in simulation of wave-induced motion of a maneuvering
ship. Based on those results, the turning and zig-zag maneuvers of the
ship in waves are investigated in the present study. Numerical
simulations are carried out with a S-175 container ship. The predicted
turning trajectories are compared with the model tests, which were
carried out at the Ocean Engineering Model Basin of Shanghai Jiao
Tong University. The predicted time histories of heading angle and
rudder angle for zig-zag maneuvers are compared with the existing
experimental data (Yasukawa, 2008).
2. Mathematical formulations
2.1. Coordinate systems
Two right-handed coordinate systems are adopted, as shown in
Fig. 1. The first one ⇀
x = (x, y, z ) is a reference frame which moves with
the speed of ship maneuvering motion, with the positive x towards the
bow and the positive z pointing upward. The xy plane is always
coincident with the calm water level and the origin of the frame is at
[Mij ]ξj̈ (t ) + [Cij ]ξj (t ) = Fi (i, j = 1, 2, …, 6) (3)
⇀ ⇀
where ξT = (ξ1, ξ2, ξ3) and ξR = (ξ4, ξ5, ξ6 ) represent the ship's transla-
tional and rotational displacements, respectively; [Mij ] is the inertial
matrix for the hull, [Cij ] is the matrix of hydrostatic restoring
coefficients and Fi denotes the high frequency hydrodynamic forces
including F-K force, radiation force and diffraction force.
In order to carry out numerical simulations of ship maneuvering in
waves, the hydrodynamic forces on the right hand side of Eq. (2) must
be determined. In this study, it is assumed that the waves have no effect
on the low frequency hydrodynamic forces acting on the hull (H),
propeller (P) and rudder (R). These forces are written according to the
typical maneuvering theories and evaluated based on the model test
data. Meanwhile, a velocity potential is introduced to present the flow
field around the hull, and the wave drift forces and moment are
determined by solving the boundary value problems for the velocity
potential.
2.3. Low frequency hydrodynamic forces acting on the hull
(XH , YH , NH )
The low frequency hydrodynamic forces on the hull are given in the
following forms:
⎧ X = (1/2)ρL2V 2X′ (v′, r′) − R(u )
⎪ H H
⎨ YH = (1/2)ρL2V 2YH′ (v′, r′)
⎪
⎪ N = (1/2)ρL3V 2N ′ (v′, r′)
⎩ H H (4)
where ρ is the water density, L represents the ship length, V = u + v 2 2

⎯⇀ is the ship speed, v′ = v / V and r′ = r (L / V ) denote the non-dimensional

the midship; the second one X = (X , Y , Z ) is a global frame fixed in
space. The two coordinate systems coincide at the initial instant and lateral speed and yaw rate, respectively. R(u ) is the ship resistance,
have the following relations: which is estimated based on the following equation:

⎧ X = x cos ψ (t ) − y sin ψ (t ) + X0(t ) R(u ) = (k f Cf + Cr )⋅(1/2)ρu 2S (5)

⎪
⎨Y = x sin ψ (t ) + y cos ψ (t ) + Y0(t )
⎪ where Cf is the frictional resistance coefficient determined by the 8th
⎩Z = z (1) ITTC formula, Cr is the residuary resistance coefficient, k f is the form
⎯⇀ factor, S is the wetted surface area.
where (X0 , Y0 ) is the midship position in the X frame, ψ is the heading
X′H , Y′H and N′H are expressed as polynomial functions of v′ and r′:
angle, and t denotes time.
The boundary value problem is formulated with respect to the ⎧ X′ = X ′ v′r′ + X ′ v′ 2 + X ′ r′ 2
frame ⇀
⎯⇀
x , while the frame X is employed to determine the global ⎪
⎪
H vr vv rr
′ v′ 3 + Yrrr
⎨YH′ = Yv′v′ + Yr′r′ + Yvvv ′ r′ 3 + Yvvr
′ v′ 2r′ + Yvrr
′ v′r′ 2
position of the ship as well as the direction of the incident waves. ⎪
⎪ N ′ = N ′v′ + N ′r′ + N ′ v′ 3 + N ′ r′ 3 + N ′ v′ 2r′ + N ′ v′r′ 2
⎩ H v r vvv rrr vvr vrr (6)
2.2. Basic motion equations
X
ψ
Assume that the ship is rigid and undergoing six degrees of freedom χ
(6-DOF) oscillations while translating with forward speed u, transverse
speed v and rotating with yaw rate r in regular waves. Since the
frequencies of maneuvering motion and wave-induced motion are very
x
different, the total ship motion is separated into two parts: the low
frequency maneuvering motion and the high frequency wave-induced u
motion.
The maneuvering motion is regarded as the basic motion of the
-v
ship, which is governed by the following equation: o r
⎧ (m + m )u ̇ − (m + m )vr − mx r 2 = X + X + X + X
⎪
⎪
x y G H P R W
⎨ (m + m y )v ̇ + (m + mx )ur + mxGr ̇ = YH + YR + YW
⎪
⎪ (I + J + mx 2 )r ̇ + mx (v ̇ + ur ) = N + N + N Incident
⎩ zz zz G G H R W (2)
y Wave
where m is ship's mass, Izz is moment of inertia about z axis; xG
represents the x-coordinate of the center of gravity; mx, my and Jzz are
the corresponding added masses and added moment of inertia; the
subscripts H, P, and R denote the low frequency hydrodynamic forces
Y
on the hull, propeller and rudder, respectively; the subscript W denotes O
the wave drift forces.
The equations of 6-DOF wave-induced motion are expressed as: Fig. 1. Coordinate systems.

369
W. Zhang et al. Ocean Engineering 137 (2017) 367–381

where X ′vr , Y ′v , Y ′r etc. are the non-dimensional hydrodynamic

derivatives.
z
2.4. Hydrodynamic force due to propeller ( XP )
The hydrodynamic force XP due to propeller is expressed as

XP = (1 − tp )ρn 2DP4KT (JP ) r

(7) x u
o
where tp, n , Dp , KT are the thrust deduction fraction, propeller
revolution, propeller diameter, and thrust coefficient, respectively. KT v
is approximately expressed as 2nd polynomial function of propeller trailing vortex sheets
y
advance ratio JP :
Fig. 2. The vortex model.
KT (JP ) = j0 + j1 JP + j2 JP2 (8)

u(1 − wP ) order of O(1) and is related to the double-body flow induced by

JP = maneuvering motion. It satisfies the following boundary conditions:
nDp (9)
⎧ ∂Φ ⎯⇀ ⇀
The wP is the wake fraction changing with maneuvering motions in ⎪
⎪ ∂n = W ⋅ n , on SB
general. Several formulae for wP have been summarized by Yasukawa ⎨
and Yoshimura (2015). In this study the following formula is used ⎪ ∂Φ = 0, on z= 0
⎪
⎩ ∂z (20)
wP = wP 0⋅exp( − 8.0βP2 ) (10) ⎯⇀ ⇀ ⇀ ⇀
where W = (u − ry ) i + (v + rx ) j + 0 k , S B is the mean wetted hull
where wP0 denotes the wake fraction at propeller position in straight surface.
moving, βP is the geometrical inflow angle to the propeller in In order to take into account the lifting effect associated with the
maneuvering motion and defined as maneuvering motion, trailing vortex sheets are introduced which are
βP = β − x′P r′ = tan−1( − v / u ) − x′P r′ assumed shed from the hull at both the keel line and the stern end, as
(11)
shown in Fig. 2.
where β denotes the drift angle, x′P is the non-dimensional longitudinal The perturbation potential ϕ(⇀ x , t ) and the wave elevation η are
coordinate of propeller. assumed to be order of O(ε ). They are related to the free surface flow
and further decomposed as:
2.5. Rudder forces and moment ( XR ,YR ,NR )
ϕ(⇀
x , t ) = φ(⇀
x , t ) + φI (⇀
x , t) (21)
Effective rudder forces XR , YR and NR are expressed as
⎧ XR = − (1 − tR )FN sin δ η(x, y, t ) = ζ (x, y, t ) + ζI (x, y, t ) (22)
⎪
⎨ YR = − (1 + aH )FN cos δ where φI (⇀
x , t ) represents the incident wave potential, ζI (x, y, t ) is the
⎪ N = − (x + a x )F cos δ
⎩ R R H H N (12) incident wave elevation. φ(⇀ x , t ) and ζ (x, y, t ) denote the remaining
parts of the perturbation potential and wave elevation, respectively.
where FN denotes the rudder normal force; δ is the rudder angle, xR is
In the frame ⇀ x , the incident wave potential in deep water is given
the longitudinal position of the rudder, tR, aH , xH represent the
as:
interaction between the hull and rudder.
The rudder normal force FN is evaluated by g
φI (⇀
x , t ) = A e kz sin {k[x cos(χ − ψ ) + y sin(χ − ψ ) + X0 cos χ + Y0 sin χ ]
ω
1 6.13Λ
FN = ρ Ad UR2⋅ sin αR − ωt} (23)
2 Λ + 2.25 (13)
where Λ is the rudder aspect ratio. UR, αR represent the relative inflow where A, ω and k denote the amplitude, frequency and wave number of
velocity to rudder and the inflow angle, which are estimated as follows: the incident wave, respectively; χ is the incident angle of wave defined
⎯⇀
in the frame X .
αR = δ − tan−1(vR / uR ) (14) By carrying out the linearization about the double-body flow,
linearized BVP for φ(⇀x , t ) reads as follows:
uR = uPε 1 + 8κKT /(πJP2 ) (15)
⎧ ∂ ⎯⇀ ∂ 2Φ ∂φ
⎪[ ∂t − ( W − ∇Φ)⋅∇]ζ = ∂z2 (ζ + ζI ) + ∂z − ∇Φ⋅∇ζI , on z= 0
uP = u[(1 − wP ) + τβP2] (16) ⎪ ∂
⎪[ − ( ⎯⇀ W − ∇Φ)⋅∇]φ = − gζ − ∇Φ⋅∇φI , on z= 0
⎨ ∂t
vR = vR′ V = (γv′ + CRrr′ + CRrrrr′ 3 + CRrrvr′ 2v′) V (17) ⎪ 6
⎪ ∂φ = ∑ ( ∂ξj n + ξ m ) − ∂φI , on S
UR2 = uR2 + vR2 ⎪ ∂n ∂t j j j ∂n B
(18) ⎩ j =1 (24)
The formulae shown in Eqs. (15)–(17) are taken from Son and where (n1, n 2 , n3) = ⇀
n , (n4 , n5, n6 ) = ⇀
x ×⇀
n , mj is the so-called m-terms
Nomoto (1981). The coefficients ε ，κ ，τ ，γ represent the interaction (Newman, 1978), which can be evaluated using the basic flow:
among the hull, propeller and rudder, while CRr ,CRrrr and CRrrv are
⎯⇀
experimental constants for expressing vR accurately. (m1, m 2, m 3) = (⇀
n ⋅∇)( W − ∇Φ )
⎯⇀
(m4, m5, m6 ) = (⇀n ⋅∇)[⇀x × ( W − ∇Φ )] (25)
2.6. Wave drift forces and moment (XW ,YW ,NW )
The wave drift forces are determined by a seakeeping analysis. To Theoretically, an interaction exists between the trailing vortices and
this end, a velocity potential Ψ (⇀x , t ) is introduced and decomposed the perturbation potential ϕ(⇀ x , t ). In the present study, however, the
x , t ) and a perturbation potential ϕ(⇀
into a basic potential Φ(⇀ x , t ): influence of the wave potential on the trailing vortices is neglected,
because of the difficulty to determine the strength of the vortex in the
Ψ (⇀
x , t ) = Φ(⇀
x , t ) + ϕ(⇀
x , t) (19)
presence of free surface and incident waves. The error introduced by
The basic potential is assumed to be the main component with its this treatment is supposed to be small, since ϕ(⇀ x , t ) is smaller than

370
W. Zhang et al. Ocean Engineering 137 (2017) 367–381

Φ(⇀x , t ) by one order of magnitude. On the other hand, the effects of than the wave-induced motions, two different time scales, denoted by
the trailing vortices on the wave potential φ(⇀ x , t ) are considered τL and τH , are used in maneuvering and seakeeping computation,
through the m-terms as well as the leading-order terms kept in the respectively. The ratio of τL to τH is denoted by Nt, which is much
linearized free surface boundary conditions. larger than 1. Referring to Seo and Kim (2011), a parallel time
Using Bernoulli's equation, the first-order hydrodynamic pressure marching scheme is used. The algorithm is generally summarized as
from the perturbation flow is obtained: follows:
∂ ⎯⇀
p(1) = − ρ( − ( W − ∇Φ )⋅∇)ϕ (a) The maneuvering equation is solved for one time step to obtain
∂t (26) ⎯⇀
ship velocity W and global ship position (X0 , Y0 ).
The generalized hydrodynamic force for high frequency wave- (b) The basic flow is recalculated to update the linearized BVP for the
induced motions can be determined by: perturbation potential, Eq. (24).
(c) The updated linearized BVP, Eq. (24), is solved for Nt time steps to
Fj = ∫ ∫S B
p(1) njds, j = 1, 2, …, 6
(27) obtain the wave-induced ship motions as well as the wave drift
forces and moment.
By substituting Fj into Eq. (3), the wave-induced motions can (d) The averaged drift forces and moment are substituted back to
⇀(2) perform the maneuvering simulation at next time step.
be determined. Then the second-order wave force F =
(2) (2) (2) (2) (2) (2)
(Fx , Fy , Fz , Mx , My , Mz ) is evaluated based on the pressure
integration proposed by Joncquez (2009): The above four steps are repeated until the end of time-marching
1 ⇀(2) ⎡ ⎯⇀ 1 ⎤ (2) 1
procedure. Fig. 3 shows a sketch map of the algorithm, in which the
⎯
∬S ⎢ − W ⋅∇Φ + ∇Φ⋅∇Φ + gz ⎥⇀ ∬ n (0)ds
(∇ϕ⋅∇ϕ)⇀
ρ
F =−
B⎣ 2 ⎦
n ds −
2 SB step (c) is referred to as “Seakeeping’.
⎧ ∂ϕ ⎯⇀
⎯ ⎡ ⎯⇀ ⎯ 1 ⎤ ⇀⎫ (1)
− ∬ ⎨ − ( W − ∇Φ )⋅∇ϕ + gξz + ∇⎢ − W ⋅∇Φ + ∇Φ⋅∇Φ ⎥⋅ ξ ⎬⇀n ds
SB ⎩ ∂t ⎣ 2 ⎦ ⎭
⎧ ⎡ ∂ϕ ⎯⇀
⎯ ⎤ ⇀⎫ (0) 1
− ∬S ⎨∇⎢ − ( W − ∇Φ )⋅∇ϕ⎥⋅ ξ ⎬⇀
n ds + g(η − ξz )2⇀
∫ n (0)dl 3.2. Determination of the basic potential and m-terms
B ⎩ ⎣ ∂t ⎦ ⎭ 2 WL
⎧ ⎡ 1 ⎤ ⇀⎫
− ∫WL ⎨⎩∇⎢⎣ − ⎯⇀
⎯
W ⋅∇Φ(0) +
2 ⎦ ⎭
n (0)dl
∇Φ⋅∇Φ(0)⎥⋅ ξ ⎬(η − ξz ) ⇀ The basic potential is determined by a constant panel method.
⎡ ⎯⇀⎯ 1 ⎤ During each time step of maneuvering simulation, the maneuvering
− ∫ n (1)dl
⎢ − W ⋅∇Φ(0) + ∇Φ⋅∇Φ(0)⎥(η − ξz )⋅⇀
WL ⎣ 2 ⎦ (28) motion is assumed to be steady and a trailing vortex sheet with given
shape is adopted. The Kutta condition is modeled by using the Morino's
where ξz = ξ3 + ξ4y − ξ5x is the vertical displacement of point (x, y) due
method (Morino and Kuo, 1974). Details of the solving method of the
to the wave-induced motions, ⇀ n (0), ⇀
n (1) and ⇀
n (2) denote the zero-, first-
basic potential as well as the related validations can be found in Zhang
and second-order components of the normal vector on the hull surface,
and Zou (2016).
defined as:
Although the basic flow can be used to obtain the low frequency
⎧ ⇀ ⎫ ⎧ ⇀ ⇀ ⎫ ⎧ ⎫ hydrodynamic forces acting on the hull, the present study elects to
⎪ ξR × n ⎪ H⇀
n
n (0) = ⎨⇀ n ⇀⎬ , ⇀ n (1) = ⎨⇀ n (2) = ⎨ ⇀ ⇀
⎪ ⎪
⇀ ⎬, ⇀ ⇀ ⇀ ⎬
⎩x × n ⎭ ⎪ ⇀ ⎩ H ( x × n ) + ξT × ( ξR × ⇀
n )⎭ determine the hydrodynamic forces by using the model test data, for
⎩ ξT × ⇀ n + ξR × (⇀ n )⎪
x ×⇀
⎪ ⎪

⎭
⎡ 2 2 ⎤ the reasons that the predicted values from the basic flow may be lower
⎢−(ξ5 + ξ6 ) 0 0 ⎥
H = ⎢ 2ξ4ξ5
1
−( ξ 2
+ ξ 2
) 0 ⎥ than the experimental ones at larger drift angle due to the neglect of the
2⎢ 4 6 ⎥
⎢⎣ 2ξ ξ 2ξ ξ 2 2 ⎥
−(ξ + ξ )⎦
viscous effect (see Zhang and Zou, 2016). On the other hand, the basic
4 6 5 6 4 5
flow is used in the linearized BVP for φ(⇀ x , t ) and determining the m-
(29) terms, so that the effects of maneuvering motion on the perturbation
⇀(2) flow could be taken into account in a more reasonable manner. In
The mean values of the first, second and sixth components of F
are the wave drift forces and moment XW , YW , and NW , respectively. practice, the m-terms are calculated by solving the boundary value
problems of Dirichlet type, based on the scheme proposed by Wu
3. Numerical implementation (1991).

3.1. Fundamental time marching scheme

3.3. Determination of the drift forces and moment
In the present study, the low frequency maneuvering motion
Eq. (2) is solved by using a fourth-order Runge-Kutta scheme, The perturbation potential φ(⇀ x , t ) is solved by using a time domain
while the seakeeping analysis is conducted based on a time domain Rankine panel method, following the approach of Kring (1994). The
Rankine panel method. Since the maneuvering motion is much slower integral equation for φ(⇀
x , t ) reads as

Fig. 3. Time marching scheme.

371
W. Zhang et al. Ocean Engineering 137 (2017) 367–381

Table 1 on the free surface and the potential on the hull surface) on each panel
Main particulars of the S-175 container ship. are assumed slowly varying and the variation is approximated by using
the quadratic B-Spline basis function B as
Specifications Full scale Model
∞
ζ (⇀
x , t ) ≃ ∑ j =−∞ ζj (t )Bj (⇀
x)
Hull Length between perpendiculars, 175.0 3.0337
⇀ ∞
L (m) φ( x , t) ≃ ∑
z
φ (t )B (⇀
j =−∞ z j
x) j
Beam, B (m) 25.40 0.4403 ∞
Draft, T (m) 9.50 0.1647 φ(⇀
x , t ) ≃ ∑ j =−∞ φj (t )Bj (⇀
x) (32)
Displacement, ▽(m3) 24742 0.127
Block coefficient, CB 0.57 0.57 The coefficients ζj (t ), φzj (t ) and φj (t ) on each panel are the spatially
Radius of gyration in roll, kxx 0.33B 0.33B
discrete unknowns.
Radius of gyration in pitch, kyy 0.25L 0.25L
Radius of gyration in yaw, kzz 0.269L 0.269L The major advantage of using this B-Spline basis function is the
Position of the center of gravity (−0.042 m, 0, effective treatment of the differential operators. Referring to Nakos
−0.01 m) (1990), the basis function can be differentiated analytically up to twice.
Hence, when treating the free surface conditions, the spatial gradients
Propeller Diameter, Dp (m) 6.5064 0.1128
Pitch ratio 0.915 0.915
of ζ and φ can be simply written as
Number of blades 5 5 ∞
∇ζ (⇀
x , t ) ≃ ∑ j =−∞ ζj (t )∇Bj (⇀
x)
Rudder Area, Ad (m2) 32.46 0.0098 ⇀ ∞ ⇀
∇φ ( x , t ) ≃ ∑ φ (t )∇B ( x )
j =−∞ j j (33)
Height, H (m) 7.7 0.1335
Aspect ratio 1.8268 1.8268
In this way the introduction of finite difference scheme is avoided.
For the time integration of free surface conditions, the kinematic
and dynamic free surface conditions are satisfied through the explicit
and implicit schemes, respectively. Adopting the Einstein's summation
convention, the discrete equations are given as follows:

ζ j(n +1) − ζ j(n) ⎯⇀

Bij − ζ j(n)( W − ∇Φ )⋅∇Bij = (ζ j(n)Bij + ζI )Φzz + φzj(n)Bij − ∇Φ⋅∇ζI
Δt
(34)

φj(n +1) − φj(n) ⎯⇀

Bij − φj(n +1)( W − ∇Φ )⋅∇Bij = − ζ j(n +1)gBij − ∇Φ⋅∇φI
Δt (35)

At the instant t = tn+1, the kinematic free surface boundary condi-

tion (34) uses the solution for vertical flux at t = tn to update the wave
elevation. The dynamic free surface boundary condition (35) uses the
wave elevation just obtained to update the potential. The integral Eq.
(30) is then solved to determine the vertical flux on free surface and the
potential on hull surface.
In order to satisfy the radiation condition, the numerical damping
beach is applied. Over the outer part of free surface, two numerical
Fig. 4. Bodylines of the S-175 container ship.
damping items are added into the kinematic free surface boundary
condition as follows:

∂ ⎯⇀ υ2
[ − ( W − ∇Φ )⋅∇]ζ = Φzzζ + φz − 2υζ + φ
∂t g (36)

where υ is the so-called damping strength. Details about the numerical

damping beach can be found in Huang (1997).
During the interval of each maneuvering time step τL , the linearized
BVP, Eq. (24), is solved for Nt time steps, and the employment of Eq.
⇀(2)
(28) yields Nt discrete data of F . Since the second-order force
oscillates two times faster than the linear wave force, the Nt discrete
⇀(2)
data of the surge component of F is fitted with the following
function:
f (t ) = a cos(2ωet ) + b sin(2ωet ) + c (37)
Fig. 5. The panel arrangement for S-175 container ship model.
where ωe is the ‘local’ encounter frequency, which is estimated as
∂G(⇀
x ′, ⇀
x) ⇀ ∂φ ⇀ ⇀ ⇀ ωe = ω − k (X0̇ cos χ + Y0̇ sin χ ) (38)
2πφ + ∬ F ∪ SB
φ
∂n′
dx′ − ∬F ∪S B ∂n′
G ( x ′, x )d x ′ = 0
(30)
It should be noted that in Eq. (38) the time derivative with respect
where G represents the Rankine source: to heading angle ψ is assumed negligible during the maneuvering time
step τL . The coefficients a and b in Eq. (37) are related to the second-
G(⇀
x ′, ⇀
x ) = 1/ (x − x′)2 + (y − y′)2 + (z − z′)2 (31) order wave-induced motions, which are not considered in the present
The hull surface (SB ) and the undisturbed free surface (F ) are study. The coefficient c equals the mean surge drift force XW . Similar
⇀(2)
discretized by plane quadrilateral panels. The unknown quantities fitting is carried out for sway and yaw components of F to obtain YW
(including the free surface elevation, the normal flux and potential and NW , respectively.

372
W. Zhang et al. Ocean Engineering 137 (2017) 367–381

Fig. 6. Comparisons of mean drift forces and moments on S-175 container ship in forward motion.

Table 2
Coefficients used in maneuvering simulation.

Coefficient Value Coefficient Value Coefficient Value

Hull m′x 0.000238 Y′v −0.0116 N′v −0.00385

m′y 0.007049 Y′r 0.00242 N′r −0.00222
J′zz 0.000419 Y′vvv −0.109 N′vvv 0.001492
X′vv −0.00386 Y′vvr 0.0214 N′vvr −0.0424
X′vr −0.00311 Y′vrr −0.0405 N′vrr 0.00156
X′rr 0.0002 Y′rrr 0.00177 N′rrr −0.00229
Cr 0.001 kf 1.04

Propeller tP 0.175 1-wP0 0.816 x′P −0.47

j0 0.5179 j1 −0.1179 j2 −0.3618

Rudder tR 0.29 x′H −0.48 x′R −0.5

aH 0.237 ε 0.921 γ {0.088 (v′ < 0)
0.193 (v′ ≥ 0)
CRr −0.156 CRrrr −0.275 CRrrv 1.96
k 0.631

Fig. 7. Comparison of turning trajectories of S-175 model in calm water (δ = ± 35°).

4. Free running model tests condition and rudder angle.

The free running model tests were conducted at the Ocean (1) wave length: λ / L = 0.5, 0.75, 1.0 and 1.25
Engineering Model Basin of Shanghai Jiao Tong University, the (2) wave amplitude: A / L =0.005
dimension of the basin is 50 m×30 m. The water depth for the model (3) wave direction: χ = 90°, 180°
test is 5 m. A S-175 container ship model is used as the subject ship. (4) rudder angle: δ = ± 35°.
The principal particulars of the hull, propeller and rudder of the model
are listed in Table 1. The bodylines of the S-175 container ship are All these tests were performed for constant propeller revolution, i.e.
shown in Fig. 4. 10.0 rps. This revolution can produce a speed of 0.818 m/s (Fn=0.15)
The turning tests were firstly carried out in calm water condition, when the ship model is running forward in calm water. The steering
and then in regular waves under the following combinations of wave rate was 13.0°/s. The trajectory of the ship was obtained by a ship

373
W. Zhang et al.
Fig. 8. Time histories of the second order surge force.
Fig. 9. Comparison of the turning trajectories.
position detecting system, according to the LED lights fixed on the
model. Due to the limitation of the basin size, the turning tests stopped
when the ship model completed about 630° turning.
5. Numerical results and discussions
5.1. Validations of the solving method for drift forces and moment
Since the mean drift forces and moment play an important role in
the present study, the S-175 container ship advancing straightly in
374
Ocean Engineering 137 (2017) 367–381
regular waves is tested firstly to validate the related solving methods.
Within theses computations, the maneuvering model is temporarily not
considered. Alternatively, a constant forward speed (u = u 0 , v = r = 0 )
is directly given to the ship. The computational domain on the free
surface extends ±1.0L in y-direction, while 0.5L upstream and 1.0L
downstream in x-direction. The panel number reaches 75 × 20 on half
of the free surface. The half hull is represented by 30 × 10 panels. Fig. 5
illustrates the grid used in these computations, in which the bold line
represents the boundary of the numerical damping beach.
The mean drift forces and moments for the S-175 container ship at
Fn = 0.15 are shown in Fig. 6. The incident wave length is varying in
the range of 0.5 < λ /L < 1.5. At χ = 180°, i.e. in head sea, the surge
force (i.e. added resistance in waves) is calculated and compared
with the experimental data by Fujii and Takahashi (1975); at
χ = 90°, i.e. in beam sea, the surge force, together with the sway force
and yaw moment are computed and compared with the experimental
data by Yasukawa and Adnan (2006). As shown in the comparisons,
the calculated added resistance and yaw moment show fairly good
agreement with the experimental data, but the discrepancy in the
sway force is remarkable. Such discrepancy seems to be due to the
limitation of the pressure integration method because it involves
the second gradients of the perturbation potential. At the stern of the
S-175 hull, these gradients are difficult to evaluate due to the presence
of the flare. For a better prediction of the drift forces, more study is
needed.
5.2. Turning tests in calm water
In this subsection, the δ = ± 35° turning tests of the S-175 contain-
er ship in calm water are investigated. The related hydrodynamic
derivatives and coefficients used in the numerical simulation are listed
in Table 2. These values are taken from the model test data of Son and
W. Zhang et al. Ocean Engineering 137 (2017) 367–381

Fig. 10. Comparison of the mean drift forces and moment.

Fig. 11. Comparison of the port-side 35° turning trajectories in regular waves at χ = 180°.

Fig. 12. Comparison of the starboard-side 35° turning trajectories in regular waves at χ = 180°.

Nomoto (1981), except Cr , k f and j0 , j1, j2 which are estimated by
empirical formulae and propeller open water test, respectively.
Fig. 7 shows the comparisons of the turning trajectory in clam
water. It can be found that the simulated trajectories generally agree
with the test results, although the advances of the computational
results are a little larger than the experimental data.
5.3. The drift forces and moment acting on the ship turning in regular
waves
In this subsection, the mean drift forces and moment acting on the
turning hull is investigated. The port side 35° turning of the S-175
model is selected as the computational example, while the incident

375
W. Zhang et al. Ocean Engineering 137 (2017) 367–381

Fig. 13. Comparison of the port-side 35° turning trajectories in regular waves at χ = 90°.

Fig. 14. Comparison of the starboard-side 35° turning trajectories in regular waves at χ = 90°.

wave is chosen to be λ / L = 1.0 , A / L =0.005 and χ = 180°(head sea before As it can be seen in Fig. 10, the computed wave drift forces and
entering the maneuver). The initial ship speed corresponds to moment generally demonstrate periodicities on the whole. The period
Fn = 0.15. corresponds to the time that the heading angle changes 360°. Due to
Based on Eq. (38), it could be estimated that the encounter wave the presence of the drift angle and yaw rate, the magnitudes of the
period generally varies from 1(s) to 2(s) during the turning process. lateral wave drift force are different when the ship is under the
Within the computations, a common time step τH = 0.02(s ) is adopted. starboard (i.e. ψ ≈ 90°, 450°) and the port (i.e. ψ ≈ 270°, 630°) beam
Numerical computations show that this time step is small enough to sea conditions. The largest surge drift force generally appears under the
provide sufficient accuracy for the seakeeping analysis. In order to head sea condition (i.e. ψ ≈ 0°, 360°), while the largest sway drift force
determine the time step τL for the maneuvering simulation, two generally appears in quartering sea (i.e. ψ ≈ 250°, 610°) rather than in
different time scale ratio Nt = 15 and Nt = 25 are tested. Fig. 8(a) beam sea.
illustrates the computed time histories of the second-order surge force
and its mean value (i.e., surge drift force XW ) over the time region of
0s ∼ 120s . As shown in the figure, there is no remarkable difference
between the case of Nt = 15 and Nt = 25. For improving the calculation
5.4. Simulations of the turning tests in regular waves
efficiency, Nt = 25 is employed in the subsequent computations.
Fig. 8(b) illustrates the detailed results of Nt = 25, over the time region
After examining the drift forces and moment acting on the turning
of 60s ∼ 65s . It can be found that there are 25 discrete data of Fx (2) in
hull, the turning tests of S-175 ship in regular waves with different
each interval of maneuvering time step, based on which the computed
wave length and wave direction are simulated. The simulated turning
mean force seems reasonable.
trajectories are compared with the model test results. Figs. 11 and 12
In order to illustrate the effects of considering the double body flow
show the comparisons of the turning trajectories of δ = + 35°(port-
and the trailing vortex, results of the present linearization are
side) and δ = − 35°(starboard-side) turning tests, respectively, with the
compared with those of the Neumann-Kelvin (N-K) linearization. It
wave direction χ = 180° and the wave length λ / L = 0.5, 0.75, 1.0 and
should be noted that the boundary value problem resulted from N-K
1.25. Figs. 13 and 14 show the similar comparisons for wave direction
linearization is similar as Eq. (24), except that the basic potential is
χ = 90°. As shown in the figures, the measured turning trajectories drift
simply set to Φ = 0 . Correspondingly, the m-terms are simplified as
⎯⇀ both along and normal to the wave direction. These tendencies are in
(m1, m 2, m 3) = 0 , (m4, m5, m6 ) = ⇀
n × W.
accordance with the test results obtained by Yasukawa (2006), but
Fig. 9 shows the turning trajectories predicted by using the present
different from the measured results by Lee et al. (2009), which shows
method and the N-K linearization. As it can be seen, the present result
that the trajectories drift only along the wave direction. It can be seen
shows a better agreement than the N-K result, when compared with the
that the longest drift distance occurs at λ / L = 0.5. This is easily
experimental measurement.
understood since the drift sway force and drift yaw moment increase
Fig. 10 shows the computed mean drift forces and moment, which
quickly with the decrease of the wave length. However, the measured
are plotted against the heading angle ψ of the ship. As it can be seen,
drift distances at λ / L = 0.75 and λ / L = 1.0 do not appear to be much
there are some discrepancies between the predicted YW and NW by the
different. At λ / L = 1.25, the measured drift distance is insignificant. The
pres2ent method and those by the N-K linearization. Meanwhile, a
numerical results can generally give the drift tendency of the turning
remarkable discrepancy between the two results occurs in the predic-
trajectories, but the predicted drift distances are generally more
tion of the surge wave drift force XW , when the ship is under head sea
remarkable than the experimental measurements. The reason for this
condition (i.e. ψ ≈ 0°, 360°). The reason of these discrepancies is
difference is due to the inaccurate evaluation of the mean drift forces.
thought to be due to the introduction of the trailing vortex, which
Despite such difference, the overall agreement of the predicted results
changed the flow field around the ship stern.
with the experimental measurements is thought to be acceptable.

376
W. Zhang et al.
Fig. 15. Comparison of the time histories of heading angle and rudder angle for 10°/10° (starboard-side) zig-zag tests at χ = 180°.
5.5. Simulations of the zig-zag tests in regular waves
Numerical simulations of the zig-zag tests of the S-175 ship in calm
water and in regular waves are also conducted. In these simulations,
the steering rate is 12.0°/s; the ship speed before entering the
maneuver corresponds to Fn=0.15; and the incident wave amplitude
is A / L =0.01. The numerical results are compared with the experimental
data obtained by Yasukawa (2008). In the comparisons, the present
rudder angle and heading angle are multiplied by −1, in order to be in
accordance with the definition of the experiment. Figs. 15 and 16 show
the time histories of the rudder angle and the heading angle for 10°/
10° and 20°/20° (starboard-side) zig-zag maneuvers in calm water and
in regular waves with the wave direction χ = 180° and the wave length
377
Ocean Engineering 137 (2017) 367–381
λ / L = 0.5, 0.7, 1.0 and 1.2. Figs. 17 and 18 show the similar
comparisons for wave direction χ = 90°.
It can be seen from Figs. 15–18, the simulated time histories of the
rudder angle and the heading angle show fairly good agreements with
the experimental measurements for the calm water condition. The
discrepancies mainly occur in the predictions at λ / L = 0.7, 1.0 and 1.2,
especially for the tests at χ = 180°. As it can be seen from Figs. 15(c)–
(e) and 16(c)–(e), the steering timings in the simulations are different
from those in the experiments; and the overshoot angles (OSA),
especially the second OSA, are underestimated. These discrepancies
stem from two aspects. On the one hand, because the drift angle and
the rudder angle change frequently during the zig-zag tests, the wave
drift forces and moments acting on the hull are very difficult to be
W. Zhang et al.
Fig. 16. Comparison of the time histories of heading angle and rudder angle for 20°/20° (starboard-side) zig-zag tests at χ = 180°.
evaluated. Although the numerical models of the wave drift forces and
moments have been improved in the present model, it is still difficult to
give a very precise evaluation. On the other hand, the rudder forces and
propeller-rudder interactions used in the present computations are
determined from the experimental data given by Son and Nomoto
(1981). The wave effects on the rudder forces and propeller-rudder
interactions are not considered. The neglect of the wave effects on the
propeller and rudder forces may bring in a considerable error of the
rudder force prediction, especially when the rudder angle is small.
Since the zig-zag maneuver is known to be sensitive to the rudder, the
simulated zig-zag results show some remarkable discrepancies when
compared with the experimental ones.
For zig-zag tests in beam sea, it is found from Figs. 17(b) and 18(b)
378
Ocean Engineering 137 (2017) 367–381
that the time histories of the rudder angle and the heading angle at
λ / L = 0.5 are different from those in calm water, since the wave drift
sway force and yaw moment are significant. The present method can
generally capture the changes of the steering timings. As also shown in
Figs. 17(c)-(e) and 18(c)-(e), the simulated results at λ / L = 0.7, 1.0 and
1.2 in beam sea show generally better agreement than those in head
sea, when compared with the corresponding experimental measure-
ments.
6. Concluding remarks
In the present study, a simulation method is presented for predicting
ship maneuvering in regular waves. Based on the two-time scale model,
W. Zhang et al.
Fig. 17. Comparison of the time histories of heading angle and rudder angle for 10°/10° (starboard-side) zig-zag tests at χ = 90°.
the total ship motion is divided into the low frequency maneuvering part
and the high frequency seakeeping part. The maneuvering motion is
calculated by using a MMG model, while the seakeeping problem is
solved based on a time domain Rankine panel method. The mean
second-order wave loads, evaluated in the seakeeping computation, are
included in the MMG model as the wave drift forces and moment.
In order to validate the present method, the turning tests of S-175
model in regular waves are simulated. The numerical simulations are
compared with the results of model tests, which were carried out at the
Ocean Engineering Model Basin of Shanghai Jiao Tong University. The
comparisons show that the incident waves can have a significant
influence on the maneuvering performance of a ship, especially when
the wave length is small. Although there is some room for improve-
ment, the predicted trajectories show overall acceptable agreement
379
Ocean Engineering 137 (2017) 367–381
with the measured ones. It is also found from the numerical computa-
tion that the magnitudes of the wave drift sway force are different when
the ship is under starboard and port beam sea conditions during the
turning process; and the largest drift sway force generally appears in
quartering sea rather than in beam sea. These phenomena are to be
validated experimentally in the future study.
Besides the turning tests, numerical simulations of the zig-zag tests
of the S-175 ship in regular waves are also conducted; the numerical
results are compared with the experimental data by Yasukawa (2008).
It is found that the simulations give generally better agreement under
beam sea condition than under head sea condition, when compared
with the corresponding experimental results. For a better prediction, a
more complex numerical model seems to be necessary, especially
including the wave effects on the rudder forces.
W. Zhang et al.
Fig. 18. Comparison of the time histories of heading angle and rudder angle for 20°/20° (starboard-side) zig-zag tests at χ = 90°.
Acknowledgements
This study is supported by the National Natural Science Foundation
of China (Grant no. 51279106) and the Lloyd's Register Foundation
(LRF) to which the authors are most grateful. LRF helps to protect life
and property by supporting engineering-related education, public
engagement and the application of research.
References
Bailey, P.A., Price, W.G., Temarel, P., 1997. A unified mathematical model describing the
manoeuvring of a ship travelling in a seaway. Trans. R. Inst. Nav. Archit. 140,
131–149.
Fang, M.C., Luo, J.H., Lee, M.L., 2005. A nonlinear mathematical model for ship turning
circle simulation in waves. J. Ship Res. 49 (2), 69–79.
380
Ocean Engineering 137 (2017) 367–381
Fujii, H., Takahashi, T., 1975. Experimental study on the resistance increase of a ship in
regular oblique waves. In:Proceedings of the 14th International Towing Tank
Conference. Ottawa, Canada, 351–360.
Huang, Y.F., 1997. Nonlinear Ship Motions by a Rankine Panel Method (Ph.D Thesis).
Massachusetts Institute of Technology, Cambridge, MA, USA.
Joncquez, S.A.G., 2009. Second-order Forces and Moments Acting on Ships in Waves
(Ph.D. Thesis). Technical University of Denmark, Copenhagen, Denmark.
Kring, D.C., 1994. Time Domain Ship Motions by a Three-dimensional Rankine Panel
Method (Ph.D Thesis). Massachusetts Institute of Technology, Cambridge, MA, USA.
Lee, S.K., Hwang, S.H., Yun, S.W., Rhee, K.P., Seong, W.J., 2009. An experimental study
of a ship manoeuvrability in regular waves, In: Proceedings of the International
Conference on Marine Simulation and Ship Manoeuvrability (MARSIM 2009),
Panama City, Panama.
Lin, W.M., Zhang, S., Weems, K., Liut, D., 2006. Numerical simulations of ship
maneuvering in waves. In: Proceedings of the 26th Symposium on Naval
Hydrodynamics, 17–22.
Morino, L., Kuo, C.C., 1974. Subsonic potential aerodynamics for complex
configurations: a general theory. AIAA J. 12 (2), 191–197.
W. Zhang et al.
Nakos, D.E., 1990. Ship Wave Patterns and Motions by a Three Dimensional Rankine
Panel Method (Ph.D Thesis). Massachusetts Institute of Technology, Cambridge,
MA, USA.
Newman, J.N., 1978. The theory of ship motions. Adv. Appl. Mech. 18, 221–283.
Seo, M.G., Kim, Y., 2011. Numerical analysis on ship maneuvering coupled with ship
motion in waves. Ocean Eng. 38 (17–18), 1934–1945.
Skejic, R., Faltinsen, O.M., 2008. A unified seakeeping and maneuvering analysis of ships
in regular waves. J. Mar. Sci. Technol. 13 (4), 371–394.
Skejic, R., 2013. Ships maneuvering simulations in a seaway - How close are we to
reality? In: Proceedings of the International Workshop on Next Generation Nautical
Traffic Models 2013, Delft, The Netherlands.
Son, K.H., Nomoto, K., 1981. On the coupled motion of steering and rolling of a high
speed container ship. J. Soc. Nav. Archit. Jpn. 20, 73–83.
Subramanian, R., Beck, R.F., 2015. A time-domain strip theory approach to maneuvering
in a seaway. Ocean Eng. 104, 107–118.
Sutulo, S., Guedes Soares, C., 2006. A unified nonlinear mathematical model for
simulating ship manoeuvring and seakeeping in regular waves. In: Proceedings of
the International Conference on Marine Simulation and Ship Manoeuvrability
(MARSIM 2006), Terschelling, The Netherlands.
Tello Ruiz, M., Candries, M., Vantorre, M., Delefortrie, G., Peeters, D., Mostaert, F.,
2012. Ship manoeuvring in waves: a literature review. Version 2_0. WL Rapporten,
00_096. Flanders Hydraulics Research & Ghent University, Antwerp, Belgium.
381
Ocean Engineering 137 (2017) 367–381
The Manoeuvring Committee, 2011. Final Report and Recommendations to the 26th
ITTC. Proceedings of 26th ITTC, Vol. I, 123-181.
Ueno, M., Nimura,T., Miyazaki, H., 2003. Experimental study on manoeuvring motion of
a ship in waves. In: Proceedings of the International Conference on Marine
Simulation and Ship Manoeuvrability (MARSIM 2003), Kanazawa, Japan.
Wu, G.X., 1991. A numerical scheme for calculating the mj terms in wave-current-body
interaction problem. Appl. Ocean Res. 13 (6), 317–319.
Yasukawa, H., 2006. Simulation of ship maneuvering in waves (1st report: turning
motion). J. Jpn. Soc. Nav. Archit. Ocean Eng. 4, 127–136.
Yasukawa, H., Adnan, F.A., 2006. Experimental study on wave-induced motions and
steady drift forces of an obliquely moving ship. J. Jpn. Soc. Nav. Archit. Ocean Eng.
3, 133–138.
Yasukawa, H., 2008. Simulation of ship maneuvering in waves (2nd report: zig-zag and
stopping maneuvers). J. Jpn. Soc. Nav. Archit. Ocean Eng. 7, 163–170.
Yasukawa, H., Nakayama, Y., 2009. 6-DOF motion simulations of a turning ship in
regular waves. In: Proceedings of the International Conference on Marine Simulation
and Ship Manoeuvrability (MARSIM 2009), Panama City, Panama.
Yasukawa, H., Yoshimura, Y., 2015. Introduction of MMG standard method for ship
maneuvering predictions. J. Mar. Sci. Technol. 20 (1), 37–52.
Zhang, W., Zou, Z.J., 2016. Time domain simulations of the wave-induced motions of
ships in maneuvering condition. J. Mar. Sci. Technol. 21 (1), 154–166.