International Journal of Naval Architecture and Ocean Engineering 11 (2019) 736e749

Contents lists available at ScienceDirect

International Journal of Naval Architecture and Ocean Engineering

journal homepage: [Link]
international-journal-of-naval-architecture-and-ocean-engineering/

Effects of diffraction in regular head waves on added resistance and

wake using CFD
Cheol-Min Lee a, Sung-Chul Park a, Jin-Won Yu b, Jung-Eun Choi b, Inwon Lee a, *
a
Department of Naval Architecture and Ocean Engineering, Pusan National University, Busan, South Korea
b
Global Core Research Center for Ships and Offshore Plants, Pusan National University, Busan, South Korea

a r t i c l e i n f o a b s t r a c t

Article history: This paper employs computational tools to investigate the diffraction effects in regular head waves on
Received 20 December 2017 the added resistance and wake on the propeller plane. The objective ships are a 66,000 DWT bulk carrier
Received in revised form and a 3,600 TEU container ship. Fixed and free to heave and pitch conditions at design speed have been
30 November 2018
taken into account. Two-phase unsteady Reynolds averaged Navier-Stokes equations have been solved
Accepted 19 February 2019
Available online 21 February 2019
using the finite volume method; and a realizable k-ε model has been applied for the turbulent closure.
The free surface is obtained by solving a VOF equation. The computations are carried out at the same
scale of the model tests. Grid and numerical wave damping zones are applied to remove unwanted wave
Keywords:
Wave diffraction
reflection at the boundaries. The computational results are analyzed using the Fourier series. The added
Added resistance resistances in waves at the free condition are higher than those at the fixed condition, which are nearly
Wake constant for all wavelengths. The wake velocity in waves is higher than that in calm water, and is
Bulk carrier accelerated where the wave crest locates on the propeller plane. When the vertical motion at the stern
KCS goes upward, the wake velocity also accelerated.
CFD © 2019 Production and hosting by Elsevier B.V. on behalf of Society of Naval Architects of Korea. This is an
open access article under the CC BY-NC-ND license ([Link]

1. Introduction becomes larger as the ship motions increase. This is significant in

the resonance frequency region of heave and pitch motions. The
Predicting added resistances of a ship in waves (RAW) is essential second is the diffraction force of the incident waves on the ship
to optimize the hull form, to predict speed losses, to analyze speed (Rdif
AW) which is primarily due to wave diffractions. This component
sea trial results, to estimate the sea margin, and to find the opti- is the most significant in short waves, where the ship motion is
mum seaway route for eco-operation. Many studies have been nearly negligible. The blunt bow shape generally provides larger
performed since the 1970s to analyze the characteristics of RAW Rdif
AW (Blok, 1983; Naito et al., 1996; Orihara and Miyata, 2003; Hirota
using numerical and experimental methods. The RAW has shown et al., 2005; M. Kim et al., 2017; Yu et al., 2017). The third is damping
dependency on ship motions, bow relative motions, ship speeds, forces associated with the forced heave and pitch motions in calm
wave length, height, and heading angles, and hull form and bow water, and insignificant compared to hydrodynamic damping of
shapes. Studies have shown that RAW demonstrates strong ship motions. These three components involve energy dissipation.
nonlinear behaviors (Gerritsma and Beukelman, 1972; Journee, A major part of this energy is transmitted to the wave radiating
1976; Lee et al., 2013; Shen and Wan, 2013; Yu et al., 2017). The from the hull and a very small part of the energy is lost due to
RAW composes three components, - which are vertical motions, viscous friction. Therefore, the RAW can be considered a non-viscous
diffraction and viscous effects (Arribas, 2007). The first is known as phenomenon. This is important in practice when scaling from
drift force which has a second order nature (Faltinsen, 1990), model experiments to full scale. The numerical methods based on
mainly caused by wave radiation via the ship motion (Rrad AW) and the potential-flow theory may be reliable except in short waves
where the RdifAW is dominant. Several semi-empirical formulations
are proposed to estimate the Rdif AW (Fujii and Takahashi, 1975;

* Corresponding author.
Faltinsen et al., 1980; Liu and Papanikolaou, 2016). For the accurate
E-mail addresses: prohami@[Link] (C.-M. Lee), fol009@[Link] estimation of the RAW, the potential flow theory combined with the
(S.-C. Park), [Link]@[Link] (J.-W. Yu), jechoi@[Link] (J.-E. Choi), inwon@ empirical formula (Seo et al., 2013; Yu et al., 2017) or the viscous-
[Link] (I. Lee). flow solver (CFD workshop Tokyo, 2015; [Link]
Peer review under responsibility of Society of Naval Architects of Korea.

[Link]
2092-6782/© 2019 Production and hosting by Elsevier B.V. on behalf of Society of Naval Architects of Korea. This is an open access article under the CC BY-NC-ND license
([Link]
 C.-M. Lee et al. / International Journal of Naval Architecture and Ocean Engineering 11 (2019) 736e749 737

[Link]) may be applied. El Moctar et al. (2017) pointed out from the Table 1
RANS results of a 14,000 TEU containership that the Rdif AW is domi-
Principal particulars of objective ships at full-load draft in model scale.

nant in short waves; however, the Rrad dif

AW are nearly equal to the RAW K-Supramax KCS
in long waves; the sum of diffraction and radiation forces exceeded Ship Model Ship Model
total resistance in long waves; and the interaction of these two
Scale ratio l 1.00 33.33 1.00 37.89
force components needed to be accounted for.
Length between perpendiculars [m] LPP 192.0 5.76 230.0 6.07
Understanding of the wake velocity in waves is necessary to Length on waterline [m] LWL 196.0 5.88 232.5 6.14
access propeller loads and added power. The wake velocity in Breadth [m] B 36.0 1.08 32.2 0.85
waves fluctuates and is affected by the longitudinal (or surge), the Draft [m] T 11.2 0.34 10.8 0.29
Wetted surface area [m2] S0 9,816 8.83 9,424 6.56
vertical (or heave and pitch) motions and the orbital motion of
Displacement [m3] V 65,005 1.76 52,030 0.96
water particles in waves. Tsukada et al. (1997) measured wake LCB from midship (þ: forward) [m] LCB 5.76 0.17 3.404 0.09
velocity by a five Pitot tube system. They presented that the wake Height of center of gravity [m] KG 7.02 0.21 7.28 0.378
velocity increased in longer waves because of larger ship motions. Moment of inertia [/LPP] kyy 0.25 0.25 0.25 0.252
Ueno et al. (2013) measured wake velocity by vane wheel current
meters for a container ship. They proposed the wake estimation
formula including the surge oscillation effect and the orbital mo- Neither rudder nor propeller was fitted to the model ship. The
tion of water particles in an attenuated wave at the stern assuming design speeds (VS) are 14.5 knots and 24.0 knots for K-Supramax
that the effects of pitch and heave motions are negligible. Wu et al. and KCS, respectively. The subscript M and S refer to the model and
(2014) measured unsteady wake velocity of KVLCC2 in ballast ship scale, respectively.
condition at fixed and free to heave and pitch in regular head The model tests for K-Supramax were performed in calm water
waves, l/LPP ¼ 0.6, 1.1 and 1.6 (where l is wave length and LPP is and 15 regular head waves at the design speed were performed at
length between perpendiculars), using 2D PIV system. They the towing tank of the Pusan National University (PNU) (Lee et al.,
showed that the wake velocity in waves was higher than that in 2017). FN (Froude number) ¼ 0.172 and RNM (Reynolds num-
calm water, and that at free to heave and pitch was greater than that ber) ¼ 1.21  106, which are non-dimensionalized by VM(¼1.29 m/
at fixed condition. s) and LPPM. For the regular waves, the wave length (l) ranges from
This paper analyzes the added resistance and wake velocity on 0.3LPP to 2.0LPP and the wave amplitude (za) is 0.005LPP. Four
the propeller plane in regular head waves of a 66,000 DWT bulk wave conditions are selected for the computational study as listed
carrier and a 3,600 TEU container ship. Fixed and free to heave and in Table 2(a). The model tests for KCS at the design speed were
pitch conditions at design speed are taken into account. The viscous performed at the FORCE Technology's towing tank in calm water
solver of STAR-CCM þ has been utilized. The fixed and free to heave and 5 regular head waves (CFD workshop Tokyo, 2015; [Link]
and pitch motions have been taken into account to appreciate the [Link]) as listed in Table 2(b), which are the same con-
diffraction effects. The results of the 3-D panel source distribution ditions for the computations. FN ¼ 0.260 and RNM ¼ 1.07  107,
method combined with the empirical formula and the model tests where VM ¼ 2.017 m/s.
in towing tanks have also been included. The computations are carried out at the design speed in model
scale. Two simulations are taken into account; motion-fixed
(hereinafter ‘fixed’) and free to heave and pitch motions (herein-
2. Objective ships and wave conditions after ‘free’) conditions.

The objective ships are a 66,000 DWT supramax bulk carrier

3. Theoretical background
(hereinafter ‘K-Supramax’) and a 3,600 TEU container ship (here-
inafter ‘KCS’). The K-Supramax was designed by Daewoo Ship-
The ship is advancing at a constant mean forward speed VS in
building & Marine Engineering Co., Ltd. (DSME), which features a
sinusoidal waves. The ship fixed right-handed global coordinate
bulbous and blunt bow. The KCS was designed by Korea Research
system(x,y,z) is defined as positive x in the flow direction, positive y
Institute for Ship and Ocean Engineering (KRISO) and features a
in the starboard, and positive z upward, as shown in Fig. 2. The
bulbous and slender bow. The body plans and the side view of the
origin is the intersection point of the midship, the centerplane and
K-Supramax and KCS are presented in Fig. 1.
the undisturbed free surface. The translatory displacements in the
The principal dimensions at the full-load draft are listed in
x, y and z directions are surge (x1), sway (x2) and heave (x3), and the
Table 1.
angular displacement of rotational motion about the x, y and z axes
are roll (x4), pitch (x5) and yaw (x6) respectively. Note that surge to
bow, sinkage to downward and trim by head are negative. The local

Table 2
Conditions of incident regular waves in model scale.

l[m] l/LPP za [m] za/l

(a) K-Supramax
2.90 0.50 0.028 0.0097
6.92 1.20 0.029 0.0042
8.02 1.39 0.028 0.0035
10.22 1.77 0.029 0.0028
(b) KCS
3.949 0.651 0.031 0.0079
5.164 0.851 0.039 0.0076
6.979 1.150 0.062 0.0089
8.321 1.371 0.075 0.0090
11.84 1.951 0.098 0.0083
Fig. 1. Body plans and side view of K-Supramax and KCS.
738 C.-M. Lee et al. / International Journal of Naval Architecture and Ocean Engineering 11 (2019) 736e749

Reynolds stresses can be written using Boussinesq's isotropic

eddy-viscosity hypothesis;
!
vUi vUj 2
rui u[ ¼ mt þ  rkdij (5)
vxj vxi 3

Where turbulence eddy viscosity is expressed as

k2
mt ¼ rCm (6)
ε


U*k
Fig. 2. Coordinate system. Cm ¼ 1 A0 þ AS (7)
ε
Details of Eqs. (1)e(7) are well describes in STAR-CCM þ user
coordinate system (xP,yP,zP) is also taken into account to evaluate
guide (CDAdapco, 2016).
the wake on the propeller, where the origin is at the center of
propeller.
The potential flow theory combined with empirical formula, and 3.1.3. Free surface
viscous flow solver have been used for the calculation of the RAW. The Volume of Fluid (VOF) model is used to model the air and
The in-house code is used for the potential-flow solver. The water free surface interactions. The VOF is an interface-capturing
potential-flow solver uses a panel method, and the RAW is obtained type scheme used to capture the free surface between two fluids.
from the near-field method with the direct pressure integration. The fraction of fluid in each cell represented by a function a, whose
Details and the formulations of the potential-flow solver are value is 1(one) in a cell full of fluid, while 0(zero) no fluid, and
extensively documented in Chun (1992). 0<a < 1 contains a free surface. The volume of fraction is governed
The Fujii and Takahashi's formula with correction (hereinafter by the following transport equation:
‘modified Fujii-Takahashi’) to predict the Rdif AW at the ship bow
(Kuroda et al., 2008; Tsujimoto et al., 2008; Yu et al., 2017) is also va va
þ Ui ¼0 (8)
applied. The results of empirical formula and potential-flow solver vt vxi
are smoothly connected using the cubic spline curve.

3.1. Viscous-flow solver

3.1.4. Wave damping zone
3.1.1. Governing equations Grid and numerical wave damping zones are applied to ensure
The governing equations for the incompressible turbulent flow that no unwanted wave reflection occurs at the boundaries of the
are the instantaneous conservation of mass (continuity equation) solution domain.
and momentum (RANS) equation. These equations are expressed in In grid damping zone, the grids are expanded. These course
tensor notation as follows: grids support wave damping by numerical dissipation, and the
waves can't propagate well.
vUi In numerical damping zone, the source term (Sdz ) is introduced
¼0 (1)
vxi (Choi and Yoon, 2009).

 ek  1
vUi vU vp v vU Sdz ¼ rðf1 þ f2 jWjÞ W (9)
r þ rU[ i ¼  þ m i  rui u[ þ rgi (2) e1  1
vt vx[ vxi vx[ vx[

nd
where Ui¼(U,V,W) is the velocity component in xi¼(x,y,z) direction, x  xsd
k¼ (10)
while p, r, m, rui u[ and gi are the static pressure, fluid density, fluid xed  xsd
viscosity, Reynolds stress and gravitational acceleration in i-direc-
Where xsd is starting point for wave damping (propagation in x-
tion, respectively.
direction) and xed is end point for wave damping (boundary); f1, f2
and nd are parameters of the damping models.
3.1.2. Turbulence modeling
A realizable k-ε eddy viscosity model with standard wall func-
tion was applied for the turbulence closure. Turbulent kinetic en- 3.2. Fourier series analysis
ergy (k) transport equation is expressed as

 The time history of the experimental and computational results
vk vk v k2 vk vk vU for encounter wave elevation, resistance, sinkage and trim are
r þ rU[ ¼ Ck þn  rui u[ i  rε (3)
vt vx[ vx[ ε vx[ vx[ vx[ analyzed by Fourier series (FS). Each unsteady history fðtÞ can be
expanded into the sum of FS:
Dissipation of turbulent energy (ε) equation is expressed as

 f0 X N
vε vε v k2 vε vε ε vU ε2 fðtÞ ¼ þ fn cosðue nt þ gn Þ; n ¼ 1; 2; 3; (11)
r þ rU[ ¼ Cε þn  Cε1 ui u[ i  Cε2 2
vt vx[ vx[ ε vx[ vx[ k vx[ k n¼1

(4)
where, fn is the nth harmonic amplitude, ue is encounter wave
The values of model constants are Ck ¼ 0.09, Cε ¼ 0.07, Cε1 ¼1.44 frequency and gn is the corresponding phase. Their values can be
and Cε2 ¼ 1.92. obtained by followings equations:
 C.-M. Lee et al. / International Journal of Naval Architecture and Ocean Engineering 11 (2019) 736e749 739

qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
fn ¼ a2n þ b2n (12)

8

>
> 1 bn
>
< tan  ; if bn  0
an
gn ¼
 (13)
>
> b
>
: tan1  n þ p; if bn < 0
an

in which,

ðT
2
an ¼ fðtÞcosðue ntÞdt (14)
T
0

ðT Fig. 3. Computational domain and boundary conditions: K-Supramax.

2
bn ¼ fðtÞsinðue ntÞdt (15)
T
0
scheme. The diffusion terms utilize the central difference scheme.
The 0th harmonic amplitude f0 is the averaged value of time The SIMPLE algorithm is applied to the velocityepressure coupling.
history of fðtÞ. The 1st harmonics f1 stands for the linear term of The temporal discretization is the 2nd order Adams-Bashforth
the unsteady histories. The 0th and 1st harmonics are the primary scheme. A High-Resolution Interface-Capturing (HRIC) scheme to
components in linear system. However, the amplitude of the higher compute the evolution of free surface by solving equation for vol-
order harmonics (n  2), which represent non-linear term of un- ume fraction of liquid was employed.
steady histories, play a crucial role in non-linear problems. The
larger the amplitudes are, the more non-linear the problem is. 4.1. Computational domain and boundary conditions

3.3. Added resistance in waves The computational domain and boundary conditions are pre-
sent in Fig. 3. The computational domain is a shape of rectangular
The RAW is the difference between the total resistance in waves with 1.5 < x/L < 4.5, 0 < y/L < 1.5, and 1.0 < z/L < 0.4
(RTW) and calm water resistance (RT) at the same ship speed. Ac- (whereas 1.0 < z/L < 1.0 for KCS). The fluid is water for the region
cording to previous works (Gerritsma and Beukelman, 1972; of 1.0 < z/L < 0.0 and air for 0.0 < z/L < 0.4 (whereas 0.0 < z/L < 1.0
Journee, 1976; Shen and Wan, 2013; Lee et al., 2013), the RAW has for KCS) at the initial condition. Damping zone is applied in the
quadratic dependence on wave amplitude. Hence, it seems feasible region of 2.5 < x/L < 4.5. Half domain is used where symmetry
to non-dimensionalize RAW using z2a , where the CAW is the added condition is applied at the centerplane.
resistance coefficient in waves; No-slip condition is applied for the hull surface. Standard wall
function is utilized. Velocity inlet and pressure outlet boundary
RAW conditions are applied for the inlet and outlet boundary plane,
CAW ¼ 2 2
(16)
rgza B LPP respectively. Symmetry condition is applied on symmetry and side
boundary plane. Slip wall boundary condition is applied on top and
Here, g is gravitational constant. bottom boundary plane.
Fujii and Takahashi (1975) proposed empirical formula to pre- The 1storder linear wave theory is applied for the inlet boundary
dict the RAW using an R function that combines the Rrad dif
AW and RAW. plane expressed as;

RAW ¼ RAW rad þ R2 ðke dÞRAW2 (17) zðx; tÞ ¼ za ,cosðk0 x  ue tÞ (19)

p2 ,I21 ðke ,dÞ Uðx; z; tÞ ¼ u0 ,za ,ek0 z cosðk0 x  ue tÞ (20)

R2 ðke dÞ ¼ (18)
p2 ,I21 ðke ,dÞ þ K 21 ðke ,dÞ
Wðx; z; tÞ ¼ u0 ,za ,ek0 z sinðk0 x  ue tÞ (21)
where ke ð¼ u2e =gÞ is encounter wave number and d is draft of ship.
RAW2 is Havelock's formula for the drift force that is based on wave
u20 ,za ,ek0 z u20 ,z2a ,e2k0 z
reflection and uses an empirical correction related to the mean pðx; z; tÞ ¼ cosðk0 x  ue tÞ  (22)
velocity of wave energy transfer. I1 and K1 are modified Bessel k0 2
function of the 1st and 2nd kind of order 1, respectively. The R Here, k(¼2p/l) is wave number and u is wave frequency.
function indicates that the contribution of Rdif AW increases with Subscript o and e are circular and encounter wave, respectively.
decreasing wavelength.
4.1.1. Grid generation
4. Computational methods and conditions The overset grid systems are applied, in which the overlapping
grids move with the ship motion, which is obtained using the 6-
To solve the governing equations, the flow domain is subdivided degrees of freedom (DOF) model. In this work, only the 2-DOF is
into a finite number of cells and the equations are changed into taken into account, i.e., heave and pitch motions.
algebraic form via the discretization process. The cell-centered The overset grid system is present in Fig. 4. The grids are
finite volume method is used for the space discretization. The generated using STAR-CCM þ by means of the trimmed mesh
convective terms are discretized using the second order upwind approach. Grid damping zone is applied in the region of 2.5 < x/
740 C.-M. Lee et al. / International Journal of Naval Architecture and Ocean Engineering 11 (2019) 736e749

L < 4.5 including numerical damping zone of 3.8 < x/L < 4.5. A
prismatic layer is generated near the wall to resolve the boundary
layer. Five layers are used. The space of the 1stgrid from the hull is
yþz120. Fine grids are set near free-surface. More than 50 cells per
wave length and 20 cells per wave height for the computation of
the shortest wave condition (l/LPP ¼ 0.5) are used. The same grid
system is used for the computations at the calm water and the other
wave conditions.

4.1.2. Wave generation test

After generating meshes, wave generation tests are performed
to know the mesh's quality and to find the computational param-
eters which give high resolution. This is done using background
grid, in which there is no hull. The computed results are compared
with the theoretical values as expressed in Eq. (19) as presented in
Table 3. The S%T denotes the percentage ratio of the computational
value to the theoretical.
The wave lengths are nearly same within differences of 1.2%. The
wave heights show a little difference within 8.9% for the K-Supra-
max and 1.6% for the KCS. Thus, it is reasonable to non-
dimensionalize the heave or pitch motions by using the
computed wave amplitude.
Fig. 5 presents the snapshots of the wave profiles and time
histories of wave elevations for the wave-generation tests in the
cases of l/L ¼ 0.5 and 1.8 for the K-Supramax. The time histories are
displayed when the wave crest locates at FP. In the snapshots, the
wave is well generated as shown in Table 3 and damped in the
damping zones. In the time histories, the encounter wave periods
(Te) are 0.845 s and 1.950 s for l/L ¼ 0.5 and 1.8, respectively,
whereas the theoretical values are 0.845 s and 1.952. That means
the periods of generated waves are well matched with the theo-
retical values.
The under-relaxation factors were set to 0.2 for momentum and
0.8 for pressure. The time step was set 0.0025 s, corresponding to 1/
500 of the short wave period (l/LPP ¼ 0.5, K-Supramax), with the
number of iterations per time step of 20. The maximum physical
time sets 60 s.

4.1.3. Side wall effect

No numerical damping zone is applied near the wall. Fig. 6
displays the snapshots of the wave patterns at the last time for
the K-Supramax. The generated waves do not reach at the side wall
in short waves of l/L ¼ 0.5, whereas they reach at the side wall in
long waves of l/L ¼ 1.2e1.8. However, these reflected waves on the
Fig. 4. Overset grid system: K-Supramax. side wall do not disturb the wave patterns near the ship. Thus, the
wave patterns near the ship are acceptable for the analyzation.
The calculation was done by a cluster consisting of 180 cores
CPU of Intel Xeon E5-2660v3/2.6 Ghz-25M and 576 memories of
RECC DDR4-2133. The computing time took 3e5 days.

Table 3
Computed incident wave length and height at model scale.

Theory CFD S%T

l [m] za[m] l/LPP za/l l [m] za [m] l [%] za [%]

(a) K-Supramax
2.880 0.029 0.500 0.010 2.889 0.029 100.3 99.1
6.912 0.029 1.200 0.004 6.911 0.031 100.0 107.9
8.064 0.029 1.400 0.004 8.071 0.032 100.1 108.9
10.368 0.029 1.800 0.003 10.367 0.031 100.0 105.8
(b) KCS
3.949 0.031 0.651 0.0079 3.945 0.031 99.9 98.4
5.164 0.039 0.851 0.0076 5.129 0.039 99.3 98.8
6.979 0.062 1.150 0.0089 6.927 0.061 99.3 98.5
8.321 0.075 1.371 0.0090 8.223 0.075 98.8 100.9
11.840 0.098 1.951 0.0083 11.826 0.097 99.9 98.6
 C.-M. Lee et al. / International Journal of Naval Architecture and Ocean Engineering 11 (2019) 736e749 741

Fig. 5. Snapshots of the wave profiles and time histories of wave elevations for the wave-generation tests in the cases of l/L ¼ 0.5 and 1.8 for the K-Supramax.

5. Results at design speed in model scale at the free and fixed conditions for
the K-Supramaxand KCS are compared in Table 4. The CTM is non-
The time histories of the computational results for the dimensionlized by r, S0 and VM.
encounter wave elevation, resistance, sinkage and trim are The error percentage of the computational prediction (E%D) of
analyzed by Fourier series. In the case of resistance, sinkage and the CTM is less than 2.6%. For K-Supramax, the computational pre-
trim, the 0th FS harmonics f0 denotes total resistance, averaged diction of CTM at the free condition is smaller by 2.5% than that at
sinkage and trim in waves, respectively. The 1st FS harmonics f1 the fixed condition. However, this trend is reverse in the case of
stands for the linear term of the unsteady histories. In the case of KCS. Both the K-Supramax and KCS are sunken and trimmed by
the sinkage and trim, f1 denotes the heave and pitch motions, head.
respectively. The E%D denotes the comparison error expressed as.

5.1. Resistance, sinkage and trim in calm water E%D ¼ ðD  SÞ=D  100 (23)

Total resistance coefficient (CTM), sinkage and trim in calm water Where D is the experimental value and S is the CFD value.

Fig. 6. Snapshots of the wave patterns at the last time step for the K-Supramax.
742 C.-M. Lee et al. / International Journal of Naval Architecture and Ocean Engineering 11 (2019) 736e749

Table 4
Total resistance coefficient, sinkage and trim in calm water of K-Supramax and KCS at design speed in model scale.

K-Supramax KCS

Free Fixed Free Fixed

Exp CFD E%D CFD Exp. CFD E%D CFD

CTM x103 4.509 4.442 1.5 4.552 3.835 3.734 2.6 3.634
Sinkage x103 [/LPP] 8.60 8.66 0.7 12.59 11.49 8.7
Trim [deg] 0.45 0.18 60.0 0.16 0.19 18.8

5.2. Resistance, heave and pitch motions in waves condition is higher than that at the fixed condition over all of the
wavelengths due to the radiation effect, except in the longest waves
Fig. 7 displays the averaged sinkage and trim, and added resis- for K-Supramax (l/LPP ¼ 1.8). For the fixed condition, the RAWs of
tance coefficients in regular head waves at the design speed. The the K-Supramax and KCS remain nearly constant for all wave-
averaged sinkage and trim in regular head waves show the same lengths. The RAW at the fixed condition is lower than that for the
tendencies as those of in calm water, which are indicated at l/ free condition in wavelengths near resonance; one half for the K-
LPP ¼ 0.0; model ships are sunken and trimmed by stern. The Supramax and one-fourth for the KCS. Those are similar to the
magnitudes are dependent on the wave lengths. The RAW at the free computational results of El Moctar et al. (2017); the Rrad
AW is nearly

Fig. 7. Sinkage, trim and added resistance coefficients in regular head waves at design speed.
 C.-M. Lee et al. / International Journal of Naval Architecture and Ocean Engineering 11 (2019) 736e749 743

zero in short waves, whereas the Rrad dif

AW are nearly equal to the RAW in strengths (or the lowest wake velocity regions) are fluctuating
long waves. To compute RradAW , they simulated the model ship in calm during Te. The shapes and wake regions at free condition are
water executing prescribed heave and pitch motions obtained from generally similar to those at fixed condition. Fig. 12 displays the
simulations in waves. They pointed out that summation of the Rrad AW effect of the orbital motion of water particles on the wake velocity.
and Rdif
AW did not yield the RAW, especially in longer waves. Obvi- When the wave crest/trough locates on the propeller plane, the
ously the physical interaction between the Rrad dif
AW and RAW caused a wake velocity increases/decreases.
reduction of the RAW. The Rdif
AW of the modified Fujii-Takahashi for- Snapshots of the axial velocity contours and velocity vectors on
mula becomes smaller as the wave length increases. The difference the propeller plane at t/Te ¼ 0.00, 0.25, 0.50 and 0.75 at the fixed
of Rdif
AW in long wave region between the modified Fujii-Takahashi and free conditions in the case of l/LPP ¼ 1.20 for K-Supramax are
formula and the present study can be explained through the displayed in Fig. 13. The wave crest and trough located on the
reflection coefficients. The modified Fujii-Takahashi formula ap- propeller plane are at t/Te ¼ 0.83 and 0.33, respectively. The char-
plies the correction factor for effect of draft and frequency (ad), acteristics at the fixed condition are similar to those in the case of l/
which is the square of the reflectance ratio for an upstanding wall LPP ¼ 0.50. At the free condition, the wake velocity is affected by
(Ursell, 1947). Ursell (1947) discussed the force acting on an up- not only the orbital motion but the ship motion. Details are
standing wall in waves consisted of the reflection and transmission described later in Fig. 16. The shapes of the ‘island’ are kept at the
components. The former is dominant in short waves, whereas the fixed condition. However, the shape of the ‘island’ become thinner
latter in long waves. The reflection effect except the transmission is and their strengths become weaker (or the lowest wake velocities
taken into account in the modified Fujii-Takahashi formula. That is become higher) at the free condition. The wake regions in waves
the reason why the CAW in the modified Fujii-Takahashi formula is also expand comparing to that in calm water. And the wake region
smaller than that in the present work in long wave region. at the free condition becomes thicker.
Fig. 8 displays the RAOs and corresponding phases of heave and Snapshots of the axial velocity contours and velocity vectors on
pitch motions in regular head waves at design speed. The phases the propeller plane at t/Te ¼ 0.00, 0.25, 0.50 and 0.75 at the fixed
are defined according to the time lag when the crest of the incident and free conditions in the case of l/LPP ¼ 0.65 for KCS are displayed
wave locates at the LCG. The CFD results of the RAOs and phases for in Fig. 14. The wave crest and trough located on the propeller plane
heave and pitch motions are similar to those of the experiments. are at t/Te ¼ 0.49 and 0.99, respectively. The characteristics also
The potential results well agree with the experiments since the shows the effect of the orbital motion of water particles, which is
double-body and steady flow approach is applied. If the Neumann- the same as that in the case of the l/L ¼ 0.5 for K-Supramax.
Kelvin approach without viscous damping model is applied, the Snapshots of the axial velocity contours and velocity vectors on
potential results will overestimate the experiments (Kim et al., the propeller plane at t/Te ¼ 0.00, 0.25, 0.50 and 0.75 at the fixed
2017). and free conditions in the case of l/LPP ¼ 1.15 for KCS are displayed
in Fig. 15. The wave crest and trough located on the propeller plane
5.3. Wake on the propeller plane in calm water and waves are at t/Te ¼ 0.87 and 0.37, respectively. The characteristics also
shows the effect of the orbital motion of water particles and ship
The axial velocity contour and velocity vectors on the propeller motions, which is the same as that of the l/LPP ¼ 1.20 for K-
plane in calm water are displayed in Fig. 9. The incoming flow on Supramax. The wake region at the free condition becomes thicker.
the propeller plane offers critical information in designing pro- The mean wake velocity at constant radius (¼URN(r,t)) is obtained
pellers. The low-speed region of u/U < 0.3 in the shape of an island by integrating the axial velocity distribution on the propeller plane
(hereinafter ‘island’) is well displayed in the results of the K- to the tangential direction (q), and the volume mean wake velocity
Supramax. This ‘island’ is the impact of the bilge vortices (counter- (¼UN(t)) is expressed as;
clockwise in starboard and clockwise in port side) which develop
around the stern. The bilge vortices are the causes of large sec- 2ðp
ondary flows as shown in the velocity vectors. In the case of KCS, 1
U RN ðr; tÞ ¼ Uðr; q; tÞ,dq (24)
there is no ‘island’, but as weak secondary flow is observed. In the 2p
0
case of the K-Supramax, another low-speed region of u/U < 0.2 with
counter-rotating vortices (clockwise in starboard and counter-
clockwise in portside) exists in the lower part of the propeller RðP 2ðp
1
axis which puts great load on the propeller. In the case of the KCS, UN ðtÞ ¼   Uðr; q; tÞ,rdq,dr (25)
this low-speed area does not exist. The results of the KCS show good p R2P  R2H RH 0
agreement with those of the experiments (Kim et al., 2001).
The snapshots of the axial velocity contours and velocity vectors Here the subscript P and H denote propeller tip and hub. The
on the propeller plane of at free condition of the KCS at l/L ¼ 1.15 time histories of the URN(r,t) and UN(t) during one encounter period
are compared with the experiments (Hossain et al., 2018) in Fig. 10. are present in Fig. 16. The solid line denotes value in calm water. The
The wave crest at t/Te ¼ 0.0 locates at the FP. The present CFD re- dashed-dot lines denote the UN(t). The dashed line denotes the
sults are similar to those of the experiments. wave elevation on the propeller plane (WEP) at the fixed condition,
Snapshots of the axial velocity contours and velocity vectors on and the vertical motion at the propeller plane (VMP) due to the
the propeller plane at t/Te ¼ 0.00, 0.25, 0.50 and 0.75 at the fixed heave and pitch motions at the free condition. The VMP is positive/
and free conditions in the case of l/LPP ¼ 0.5 for K-Supramax are negative when the ship model moves up/down. At the fixed con-
displayed in Fig. 11. The wave crest and trough located on the dition, the orbital motion of water particles affects the wake ve-
propeller plane are at t/Te ¼ 0.97 and 0.47, respectively. The wake locity as previous discussed in Figs. 11 and 13-15. The wake velocity
regions in waves at the free and the fixed conditions expand is accelerated/decelerated when the wave crest/trough locates. At
comparing to that in calm water. And the wake region is relatively the free condition except the inner region of the K-Supramax (r/
compressed at wave crest and expanded at wave trough not only at R < 0.4), the phase lags between the VMP and the WEP are negli-
the fixed but also at free condition, which reflects the effect of the gible. However, phase lags between the VMP and the wake velocity
orbital motion of water particles as shown in Fig. 12. The shapes of arise. The maximum wake velocity lags behind the maximum VMP.
the ‘island’ are kept as the wave changes over the Te. However, their This is deemed to delayed response of the fluid in the respect of the
744 C.-M. Lee et al. / International Journal of Naval Architecture and Ocean Engineering 11 (2019) 736e749

Fig. 8. RAOs and corresponding phases of heave and pitch motions in regular head waves at design speed.

ship motion. There is generally upper flow at the stern. When the condition of the K-Supramax, when the model ship moves down,
model ship moves down, this upper flow decreases, and the wake the bilge vortices become stronger, and the wake velocity is
velocity is decelerated. And the model ship moves up, the wake decelerated. When the model ship moves up, the wake velocity is
velocity is accelerated. At the inner region (r/R < 0.4) for the free accelerated. In the case of the KCS, the tangential gradient at the
 C.-M. Lee et al. / International Journal of Naval Architecture and Ocean Engineering 11 (2019) 736e749 745

Fig. 9. Axial velocity contour and velocity vectors on the propeller plane in calm water at free condition.

Fig. 10. Comparison of snapshots of the axial velocity contours and velocity vectors on the propeller plane with the experiments at l/L ¼ 1.15: KCS.

region of r/R > 0.5 evidently decreases. This is deemed to the effect The UN in waves are higher than that in the calm water by
of ship motion. 0.0e2.6% (free condition) and 7.1e7.5% (fixed condition) for the K-
As shown in these figures, the UN is fluctuating for an encounter Supramax, and 2.5e7.5% (free condition) and 3.9e6.0% (fixed
period and can be expressed as the time averaged value of volume condition) for the KCS. In the case of K-Supramax, the UN at the
average nominal velocity (UN ) and its amplitude(jUN j); free condition is lower than that at the fixed condition in the
whole wave lengths, whereas the jUN j shows the reverse. In the
case of KCS, the UN at the free condition is higher than that in fixed
UN ðtÞ ¼ UN þ jUN j,cosðue ,t þ bÞ (26)
condition for l/LPP>1.2, whereas nearly the same for l/LPP<0.85.
The jUN j becomes higher as the wave length is larger in both free
Here, b is phase angle.
and fixed conditions except for the free condition of the K-
Table 5 lists the UN and jUN j in regular head waves at the free
Supramax.
and the fixed conditions for K-Supramax and KCS.
746 C.-M. Lee et al. / International Journal of Naval Architecture and Ocean Engineering 11 (2019) 736e749

Fig. 11. Snapshot of the axial velocity contours and velocity vectors on the propeller plane at t/Te ¼ 0.00, 0.25, 0.50 and 0.75 at l/LPP ¼ 0.50; K-Supramax.

6. Conclusions

The diffraction and ship motion effects in regular head waves of

the 66,000 DWT bulk carrier and the 3,600 TEU container ship have
been investigated at the design speed using computational results.
The ships are sunken and trimmed by stern in calm water and in
waves. The added resistance in waves at the free condition is higher
than that at the fixed condition over all of the wavelengths due to
the radiation effect. The added resistance in waves at the fixed
Fig. 12. Effect of the orbital motion of water particles on the wake velocity.
condition remains nearly constant for all wavelengths, which
contradicts the result using the Fujii-Takahashi formula since the

Fig. 13. Snapshot of the axial velocity contours and velocity vectors on the propeller plane at t/Te ¼ 0.00, 0.25, 0.50 and 0.75 at l/LPP ¼ 1.20; K-Supramax.
 C.-M. Lee et al. / International Journal of Naval Architecture and Ocean Engineering 11 (2019) 736e749 747

Fig. 14. Snapshot of the axial velocity contours on the propeller plane at t/Te ¼ 0.00, 0.25, 0.50 and 0.75 at l/L ¼ 0.65; KCS.

Fig. 15. Snapshot of the axial velocity contours on the propeller plane at t/Te ¼ 0.00, 0.25, 0.50 and 0.75 at l/LPP ¼ 1.15; KCS.

correction factor of this formula applies the reflectance ratio except region of the K-Supramax (r/R < 0.4), the phase lags between the
the transmission of an upstanding wall in waves. The wake region vertical motion and the wave elevation are negligible. However,
in waves at the free is thicker than that at the fixed conditions, phase lags between the vertical motion and the wake velocity arise.
where both wake regions expand compared to that in calm water. At the inner region (r/R < 0.4) for the free condition of the K-
The orbital motion of water particles affects the wake region; the Supramax, when the model ship at the stern moves down/up, the
wake velocity is accelerated/decelerated at wave crest/trough. In bilge vortices become stronger/weaker, and the wake velocity is
the case of the K-Supramax, the shapes of the ‘island’ are kept as the decelerated/accelerated. The time averaged values of volume
wave changes over the encounter period. However, the lowest average nominal velocity in waves are higher than that in calm
wake velocity regions are fluctuating. At the fixed condition, the water by 0.0e2.6% (free condition) and 7.1e7.5% (fixed condition)
mean wake velocity at constant radius coincides with the orbital for the K-Supramax, and 2.5e7.5% (free condition) and 3.9e6.0%
motion of water particles. At the free condition except the inner (fixed condition) for the KCS.
748 C.-M. Lee et al. / International Journal of Naval Architecture and Ocean Engineering 11 (2019) 736e749

Fig. 16. Time histories of the mean wake velocity at constant r during one encounter period.

Table 5 References
Time averaged value of volume average nominal velocity and its amplitude.
Arribas, F.P., 2007. Some methods to obtain the added resistance of a ship advancing
l/LPP Free Fixed
in waves. J. Ocean Eng. 34, 946e955.
UN jUN j UN jUN j Blok, J.J., 1983. The Resistance Increase of a Ship in Waves. PhD thesis. Delft Uni-
versity of Technology.
(a) K-Supramsx CDAdapco, 2016. STAR-CCMþ User Guide, Version 11.04.
Calm 0.569 0.000 0.577 0.000 Choi, J., Yoon, S.B., 2009. Numerical simulations using momentum source wave-
0.50 0.578 0.017 0.620 0.026 maker applied to RANS equation model. Journal of Costal Engineering 56,
1.20 0.584 0.130 0.618 0.032 1043e1060.
1.39 0.581 0.110 0.620 0.046 Chun, H.H., 1992. On the added resistance of SWATH ships in waves. J. Soc. Nav.
1.77 0.569 0.071 0.618 0.058 Archit. korea 29 (4), 75e86.
(b) KCS El Moctar, O., Sigmund, S., Ley, J., Schellin, T.E., 2017. Numerical and experimental
analysis of added resistance of ships in waves. J. Offshore Mech. Arctic Eng. 139
Calm 1.399 0.000 1.396 0.000
(1), 011301e011309.
0.651 1.434 0.022 1.451 0.032
Faltinsen, O.M., 1990. Sea Loads on Ships and Offshore Structures. Cambridge Uni-
0.851 1.488 0.076 1.477 0.060
versity Press, Cambridge.
1.150 1.634 0.110 1.475 0.128 Faltinsen, O.M., Minsaas, K.J., Liapis, N., Skjørdal, S.O., 1980. Prediction of resistance
1.371 1.647 0.128 1.476 0.135 and propulsion of a ship in a seaway. In: Proceedings of the 13th Symposium on
1.951 1.576 0.198 1.480 0.190 Naval Hydrodynamics, pp. 505e529.
Fujii, H., Takahashi, T., 1975. Experimental study on the resistance increase of a ship
in regular oblique waves. In: Proceeding of 14th International Towing Tank
Conference, pp. 351e360.
Acknowledgement Gerritsma, J., Beukelman, W., 1972. Analysis of the resistance increase in waves of a
fast cargo ship. Int. Shipbuild. Prog. 19, 285e293.
Hirota, K., Matsumoto, K., Takagishi, K., Yamasaki, K., Orihara, H., Yoshida, H., 2005.
This work was supported by the National Research Foundation Development of bow shape to reduce the added resistance due to waves and
of Korea (NRF) grant funded by the Korean government (MSIT) verification of full scale measurement. In: Proceedings of the first International
through GCRC-SOP (No. 2011-0030013), the Ministry of Educa- Conference on Marine Research and Transportation (ICMRT05), Ischia, Italy,
September 19e21, pp. 63e70.
tion (No. 2015R1D1A1A01059973) and the Ministry of Trade, Hossain, Md A., Shibano, Y., Toda, Y., 2018. Forces, ship motions and velocity wake
Industry and Energy (No. 10073164), to which deep gratitude is field for KRISO container ship model in regular head waves. In: Proceedings of
expressed. the Twenty-Eighth (2018) International Ocean and Polar Engineering
 C.-M. Lee et al. / International Journal of Naval Architecture and Ocean Engineering 11 (2019) 736e749 749

Conference, Sapporo, Japan, pp. 226e233. above water bow shape with a small added resistance in waves. J. Kansai So.
Journee, J.M.J., 1976. Motion, Resistance and Propulsion of Ship in Regular Head Nav. Architect. 226, 91e98 ([in Japanese]).
Waves. Delf University of Technology. Report 0428. Orihara, H., Miyata, H., 2003. Evaluation of added resistance in regular incident
Kim, M., Hizir, O., Turana, O., Incecika, A., 2017. Numerical studies on added resis- waves by computational fluid dynamics motion simulation using an over-
tance and motions of KVLCC2 in head seas for various ship speeds. J. Ocean Eng. lapping grid system. J. Mar. Sci. Technol. 8, 47e60.
140, 466e476. Seo, M.G., Park, D.M., Yang, K.K., Kim, Y., 2013. Comparative study on computation of
Kim, W.J., Van, D.H., Kim, D.H., 2001. Measurement of flows around modern com- ship added resistance in waves. Ocean Eng. 73, 1e15.
mercial ship models. Exp. Fluid 31, 567e578. Shen, Z., Wan, D., 2013. RANS computations of added resistance and motions of a
Kim, Y.C., Kim, K.S., Kim, J., Kim, Y., Park, I.R., Jang, Y.H., 2017. Analysis of added ship in head waves. Int. J. Offshore Polar Eng. 23 (4), 263e271.
resistance and seakeeping responses in head sea conditions for low-speed full Tsujimoto, M., Shibata, K., Kuroda, M., Takagi, K., 2008. A practical correction
ships using URANS approach. Int. J. Nav. Architect. Ocean Eng. 9 (6), 641e654. method for added resistance in waves. J. Jpn. Soc. Nav. Archit. Ocean Eng. 8,
Kuroda, M., Tsujimoto, M., Fujiwara, T., Ohmatsu, S., Takagi, K., 2008. Investigation 177e184.
on components of added resistance in short waves. J. Jpn. Soc. Nav. Archit. Tsukada, Y., Hinatsu, M., Hasegawa, J., 1997. Measurement of unsteady ship wakes in
Ocean Eng. 8, 71e176. waves. J. Kansai Soc. Nav. Archit. Jpn. 228, 15e20.
Lee, C.M., Yu, J.W., Choi, J.E., Lee, I., 2017. Effect of bow hull forms on the resistance Ueno, M., Tsukada, Y., Tanizawa, K., 2013. Estimation and prediction of effective
performance in calm water and waves for 66k DWT bulk carrier. Int. J. Nav. inflow velocity to propeller in waves. J. Mar. Sci. Technol. 18 (3), 339e348.
Architect. Ocean Eng. [Link] ([on review- Ursell, F., 1947. The effect of a fixed vertical barrier on surface wave in deep water.
ing]). Article in press. Proc. Phil. Soc. 42, 374e382.
Lee, J.H., Seo, M.G., Park, D.M., Yang, K.K., Kim, K.H., Kim, Y., 2013. Study on the Wu, P.C., Okawa, H., Kim, H., Akamatsu, K., Sadat-Hosseini, H., Stern, F., Toda, Y.,
effects of hull form on added resistance. In: Proceedings of the 12th Interna- 2014. Added resistance and nominal wake in waves of KVLCC2 model ship in
tional Symposium on Practical Design of Ships and Other Floating Structures, ballast condition. In: Proceedings of the 31stSymposium on Naval Hydrody-
Changwon, South Korea, 329~337. namics, Hobart, Tasmania, Australia, 2-7 November 2014.
Liu, S.K., Papanikolaou, A., 2016. Fast approach to the estimation of the added Yu, J.W., Lee, C.M., Lee, I., Choi, J.E., 2017. Bow hull-form optimization in waves of a
resistance of ships in head waves. Ocean Eng. 112, 211e225. 66,000 DWT bulk carrier. Int. J. Nav. Architect. Ocean Eng. 9 (5), 499e508.
Naito, S., Kodan, N., Takagi, K., Matsumoto, K., 1996. An experiment study on the