Journal of

Marine Science
and Engineering

Article
CFD Simulation on Workability of a Seaweed
Harvesting Boat Due to Wake-Wash
Marco Polo Espinoza Haro 1 , Jong-Chun Park 1, * , Dong-Hyun Kim 2 and Sung-Bum Lee 3
1 Department of Naval Architecture & Ocean Engineering, Pusan National University, Busan 46241, Korea;
marco.polo.espinoza@live.com
2 Future Ship Research Department, Korea Shipbuilding & Offshore Engineering Co. Ltd., Gyeonggi-do 13591,
Korea; donghyun.kim@ksoe.co.kr
3 R&D Center, Haemin Heavy Industry Co. Ltd., Busan 46751, Korea; space489@naver.com
* Correspondence: jcpark@pnu.edu; Tel.: +82-51-510-2480




Received: 21 May 2020; Accepted: 20 July 2020; Published: 22 July 2020


Abstract: In the present study, a 2-ton class seaweed harvesting boat was optimized by employing
a W-shape hull form to reduce roll motion due to wake-wash from passing boats. A series of numerical
simulations were conducted using Star-CCM+, a commercial CFD (computational fluid dynamics)
software, to improve workability by optimizing the hull form from the conventional design (original
hull form). The 2-dimensional roll decay motion of various hull forms with W-shape midsection
were simulated and the hull form with the best performance in free roll decay test was selected.
To evaluate stability of each hull in wake-wash, the original or optimized hull was alternately located
at the middle of a computational domain as a target ship while an advancing ship (original hull)
moved forward generating Kelvin waves which impact the original or optimized boat. Two kinds
of working conditions, i.e., ballast and full loading conditions, of the target ship were considered
with and without initial roll angle. It was observed through the comparison of motion between
the original and optimized hulls a decrement of roll motion for the optimized ship demonstrating
the effectiveness of the W-shape hull. Decrement of roll motion was observed for both working
conditions. Additionally, the optimized W-shape hull showed an extraordinary performance under
the ballast condition without initial roll angle.

Keywords: seaweed harvesting boat; hull form optimization; roll motion; wake-wash; stability control

1. Introduction
Seaweed has been used across the world for centuries not only in coastal communities but in
many industries such as pharmaceuticals, industrial sectors (soil fertilizers), and food industry due to
its rich nutritional composition [1]. Seaweed can be harvested in farms or from the wild by hand from
open boats. Harvesting seaweed by hand is a dangerous activity because workers are biased toward
one side in the ship producing a change in the center of gravity generating a heel angle.
In addition, ship traffic of advancing ships near seaweed farms could induce a large energy
affecting the stability of the harvesting boats. The far-fields generated by advancing ships are classified
in transverse and divergent waves [2]; while transverse waves might not impact the harvesting boats’
stability, divergent waves might increase the energy at the seaweed farm. Therefore, divergent waves
produced by nearby ships, so-called ‘wake-wash’, can increase the heel angle leading to a risky situation
for the operational personnel [3].
Many researchers, institutions, and companies have worked to improve the stability of large ships
and cruisers by using passive stability systems such as bilge keels which have proved a reduction in the
roll motion [4]. On the other hand, design of seaweed boats has not changed largely in the recent leading

J. Mar. Sci. Eng. 2020, 8, 544; doi:10.3390/jmse8080544 www.mdpi.com/journal/jmse

J. Mar. Sci. Eng. 2020, 8, 544 2 of 11

to a lag of the improvement of the workers safety conditions. Divergent waves affect the stability of
the ships by increasing the roll motion which is known to be a nonlinear phenomenon related to effects
of viscous nature of water, such as eddy making, turbulent boundary layer development, and viscous
roll damping. Therefore, to improve the workability of the crew in harvesting boats viscous effects
must be considered. Analytical solutions based in the potential flow theory have frequently been
used to analyze ship motion successfully. However, methods based on the potential flow theory has
limitations for problems related with nonlinear phenomena such as roll motion. The CFD approaches
have provided good results, for roll motions [5]. Since the roll motion can be used to judge the stability
of a ship it is clear that CFD analysis can provide good motion prediction. The main purpose of
this work is to study an improvement in the design of seaweed harvesting boats which can lead to
a reduction of the roll motion increasing the safety of the workers and proposing a new design for the
mentioned boats. In this study, therefore, to reduce the possibility of accidents or casualties during
working conditions, an optimized ship hull was design based on W-shape to reduce the roll motion.
By numerical simulation, roll motion of 11 different hulls was carried out to find the most appropriate
W-shape in terms of roll decay performance. No experimental data of optimized hulls was available
to validate the results; therefore, a proper numerical setting was sought to produce reliable results.
For instance, roll decay motion of the rectangular barge based on experiments [6] was simulated.
Results of simulations were compared with experimental data until the proper configuration
was obtained. Then, a series of roll motion simulations were carried for 11 hulls using the confirmed
numerical settings. In addition, the increment of the hydraulic force was compared for the original and
W-shape hull under ballast and full loading condition and the effective horsepower was calculated and
compared with a sea trial. For the original and optimized hulls, the comparative studies for workability
were performed by simulating wake-wash under ballast and full loading conditions. In addition,
the influence of workers’ weight distribution was investigated by using an initial roll angle because
the position of workers is mostly biased towards one side during working hours.

2. Numerical Methods and Simulations

Numerical simulations were carried out using a commercial CFD software, Star-CCM+. Unsteady
Reynolds-averaged Navier–Stokes (URaNS) equations for continuity Equation (1) and momentum
Equation (2) were used as governing equations in the numerical simulation, and they are defined in
integral form as follows:

∂
Z ! I  Z
→ →  →
ρx dV + ρ v − vg · da = Su dV, (1)
∂t V A V

∂
Z ! I I I Z
→ → → →  → → → → →
ρx v dV + ρv ⊗ v − vg · da = − p I · da + T · da + f dV, (2)
∂t V A A A V
→ → →
where V is the cell volume, a the face area vector, ρ the density, v and v g the velocity and grid velocity,
→
respectively, Su the source term, ⊗ the tensor product, p the pressure, I the identity matrix, T the
→
viscous stress tensor, and f the external force. For more detailed explanation of governing equation
refer to Star-CCM+ Documentation User Guide [7].
The governing Equations (1) and (2) are discretized by a finite-volume manner. Viscosity effects
are modeled by using the SST (Shear Stress Transport) turbulence model [8]. It is well known that
the SST turbulence model has strength in estimating separation flow near the wall layers for flows
with moderate adverse gradient compared to the other turbulence model of RaNS [9]. In addition,
the turbulence viscosity is well resolved at the wake of the ship due to the viscosity limiter used in
the SST model. As the target ship has knuckles, chines, and transom stern which affect separation
flow, it can be determined that the SST model is suitable for this study. There are similar simulation
J. Mar. Sci. Eng. 2020, 8, 544 3 of 11

cases that used the SST model to estimate the flow field around small boat, which yielded a resistance
coefficient close to the experiment [10].
In this study the calculation of the damping force due to vorticity should be obtained accurately to
capture all the flow characteristics of the body motions, especially the roll motion. Therefore, 2nd order
of accuracy scheme was used for the diffusion term and 2nd order of accuracy TVD scheme was used
for the convection term. Second order of accuracy assures a high accuracy while the TVD scheme for
the convection schemes secures high accuracy avoiding unphysical oscillations. Implicit second order
discretization scheme was used for time advancing to maintain the stability of the solution during the
total time of the simulation.
PISO algorithm was used to decouple the pressure-velocity fields in the momentum equation.
For each time step the PISO algorithm was solved 10 times; in addition, a convergence criterion of 10−5
was set up. At each solution of the PISO algorithm the pressure and velocity fields were compared
with the previous solution and if the convergence criterion was achieved, before the PISO algorithm
was solved 10 times, the simulation advanced to the next timestep.
Prediction of hull motion was implemented through dynamic fluid body interaction (DFBI)
technique. The DFBI technique is a technique that predicts the motion of an object by solving the
six-degree-of-freedom motion equation. The external force of the motion equation is the fluid force
which is solved through the fluid governing equation. The pressure acting on the hull (dynamic
pressure, hydrostatic pressure) gravity, and buoyancy force of the hull are calculated at every time
step and substituted into the equation of motion, after the equation of motion is solved the body
position will be updated and the simulation will advance to the next time step. Sinkage, trim, and roll
motion were predicted by using the DFBI technique. The overset grid system was employed to handle
ship motion.

2.1. Validation of Numerical Modeling

To validate the CFD tool used, a roll decay test of a rectangular barge was simulated and the results
compared to the published experimental data by [6]. The simulation considered a 2-D barge 0.3 m wide,
0.1 m high, while the moment of inertia was I = 0.236 kgm2 and the density was ρ = 1049 kg/m3 .
The barge was allowed to rotate freely to its center of gravity (CG). It was inclined at an initial roll angle
J. Mar. Sci. Eng. 2020, 8, x FOR PEER REVIEW 4 of 11
of 15◦ with respect to the free surface. The draft was set at 0.05 m. Figure 1a shows the initial setup.

(a) (b)
1. Roll decay test of 2-D rectangular barge used for validation.
Figure 1. validation. (a) Dimensions
Dimensions of
of barge; (b)
system.
overset grid system.

Only
Figurerolling motion
2a shows was freely
the results represented
of roll decay test,using multicompared
that were blocks and overset
with grids while[6]other
the experiment and
the other CFD simulation [11]. The natural period obtained from the simulation is TN = 0.939s ,
motions were constrained. The inner grid was allowed to rotate while the outer grid remained static.
The
whichfineisgrid wasdifferent
0.945% used near the the
from freeexperiment
surface as shown in Figure
( TN = 0.93s 1b. Second
). Figure order
2b shows theschemes
curve of were used
extinction
+
for convection and temporal terms. The y value is defined as a non-dimensional distance (based on
for the roll decay motion. The slope in the graph was used to calculate the damping coefficient
local cell fluid velocity) from the wall to the first mesh node. Therefore, to guarantee the turbulence
following [12]. The damping coefficient calculated from the simulation results is 0.538 and shows an
model was applied correctly in the numerical simulation a y+ < 1 was used. The barge was discretized
error of 3.68% compared to 0.519 in the experiment [6], which gives a very good agreement. Therefore,
with a uniform grid near-body with spacing of 0.005 m and a Courant number (CFL) smaller than
it can be seen that the present numerical schemes and mesh setting can be considered to be valid for
the hull optimization to minimize roll motion and stability analysis related to the workability of 3-D
vessels due to wake-wash effects.
 J. Mar. Sci. Eng. 2020, 8, 544 4 of 11

(a) (b)
1 was Figure
used during the simulation.
1. Roll decay Calderer etbarge
test of 2-D rectangular al. (2014) [11]validation.
used for used four(a)grids with spacing
Dimensions of(b)
of barge; 0.008,
0.004, overset
0.002, and
grid 0.001
system.m and a time step of 0.0005 s was employed. Calderer et al. (2014) performed
LES simulations to study the vortex around the barge.
Figure2a2ashows
Figure showsthethe results
results of of
rollroll decay
decay test,
test, thatthat were
were compared
compared withwith the experiment
the experiment [6] the
[6] and and
the other
other CFD simulation
CFD simulation [11].
[11]. The The natural
natural period obtained
period obtained from thefrom the simulation
simulation is Ts,
is TN = 0.939 N = 0.939s
which is ,
which different
0.945% is 0.945%from
different from the experiment
the experiment (TN = 0.93 s). = 0.93s
( TNFigure 2b).shows
Figurethe2bcurve
showsofthe curve offor
extinction extinction
the roll
decay motion. The slope in the graph was used to calculate the damping coefficient
for the roll decay motion. The slope in the graph was used to calculate the damping coefficient following [12].
The damping
following coefficient
[12]. calculated
The damping from calculated
coefficient the simulation
fromresults is 0.538 and
the simulation shows
results an error
is 0.538 and of 3.68%
shows an
compared to 0.519
error of 3.68% in the to
compared experiment [6],
0.519 in the which gives
experiment [6], awhich
very gives
good aagreement. Therefore, it
very good agreement. can be
Therefore,
seen that
it can be the
seenpresent
that thenumerical schemes and
present numerical meshand
schemes setting can
mesh be considered
setting to be valid
can be considered tofor
be the hull
valid for
optimization to minimize
the hull optimization roll motion
to minimize rolland stability
motion and analysis
stability related
analysistorelated
the workability of 3-D vessels
to the workability of 3-D
due to wake-wash
vessels effects. effects.
due to wake-wash

(a) (b)
Figure2.2.Results
Figure Resultsofofverification
verificationcondition.
condition.(a)
(a)Roll
Rolldecay
decaymotion;
motion;(b)
(b)extinction
extinctioncurve.
curve.

2.2.
2.2.Roll
RollDecay
DecayMotion
Motionof Optimized W-Shape
of Optimized Hull Hull
W-Shape
AAseries
seriesofofroll
rolldecay
decaymotion
motionforfor1111hulls
hullswere
weresimulated
simulatedusing
usingthethenumerical
numericalschemes
schemesand andgrid
grid
system in the same manner of the prementioned validation case. Figure 3 illustrates
system in the same manner of the prementioned validation case. Figure 3 illustrates the 11 hulls used the 11 hulls used
ininthe
theoptimization
optimization stage whilewhile Case
Case11shows
showsa common
a common hull
hull geometry
geometry for for a seaweed
a seaweed harvesting
harvesting boat.
boat. Proposed
Proposed hull hull geometries
geometries (Cases
(Cases 2–11)2–11) included
included appendages
appendages attached
attached at the
at the hull
hull bottom
bottom based
based on
on different angles, length, and shapes to dissipate the rolling motion through
different angles, length, and shapes to dissipate the rolling motion through vortex. Based on the hull vortex. Based on
the hull geometry,
geometry, physicalphysical properties
properties of the hullofwere
the hull were modified,
modified, for example, formoment
example, of moment of inertia,
inertia, draft, added
draft,
mass,added mass,and
damping, damping, and force.
buoyancy buoyancy force. Therefore,
Therefore, different approaches
different approaches were considered
were considered to
to find the
find the proper balance which can produce roll motion reduction. Cases 2, 3,
proper balance which can produce roll motion reduction. Cases 2, 3, 5, 6, and considered big 5, 6, and 8 considered big
appendages,
appendages,
J. Mar.
on8,the
Sci. Eng. 2020,on
other
x the
hand,
FOR other
Cases
hand,
PEER REVIEW
4, 7, 10,
Cases and
4, 7, 10,11and
were11based
wereonbased
short appendages which smoothly
on short appendages which
5 of 11
dissipate the roll motion. Results of roll motion for Cases 1–11 are shown in Section
smoothly dissipate the roll motion. Results of roll motion for Cases 1–11 are shown in Section 3.1. 3.1.

Figure
Figure 3. 3. Geometry
Geometry of of proposed
proposed W-shape
W-shape hulls.
hulls.

2.3. Workability of Boats by Wake-Wash

In order to investigate the workability of the optimized vessel through the previous process, the
roll motion performance between the original and optimized vessels due to the influence of Kelvin
waves generated by other vessels passing by (called as wake wash [13–15]) were compared. The
portside force of the target boat was applied during the whole simulation to account for the situation
J. Mar. Sci. Eng. 2020, 8, 544 Figure 3. Geometry of proposed W-shape hulls. 5 of 11

2.3. Workability of Boats by Wake-Wash

2.3. Workability of Boats by Wake-Wash
In order to investigate the workability of the optimized vessel through the previous process, the
In order to investigate the workability of the optimized vessel through the previous process, the roll
roll motion performance between the original and optimized vessels due to the influence of Kelvin
motion performance between the original and optimized vessels due to the influence of Kelvin waves
waves generated by other vessels passing by (called as wake wash [13–15]) were compared. The
generated by other vessels passing by (called as wake wash [13–15]) were compared. The portside
portside force of the target boat was applied during the whole simulation to account for the situation
force of the target boat was applied during the whole simulation to account for the situation in which
in which the worker’s weight was biased to one side for the operation. Therefore, the simulation was
the worker’s weight was biased to one side for the operation. Therefore, the simulation was conducted
conducted by inducing an initial roll angle at the beginning of the simulation and compared with the
by inducing an initial roll angle at the beginning of the simulation and compared with the case where
case where the workers were located in the center of the boat and no initial roll angle condition was
the workers were located in the center of the boat and no initial roll angle condition was generated.
generated.
The target ship rotated freely around the x-axis (roll motion) while a passing ship advanced free
The target ship rotated freely around the x-axis (roll motion) while a passing ship advanced free
in the x-direction (surge motion). The overset grid was introduced to allow surge and roll motions in
in the x-direction (surge motion). The overset grid was introduced to allow surge and roll motions in
advancing and target boat, respectively. Figure 4 shows the schematic view of grid system for initial
advancing and target boat, respectively. Figure 4 shows the schematic view of grid system for initial
setup and location of advancing and target boats.
setup and location of advancing and target boats.

4. Schematic
FigureFigure view of
4. Schematic initial
view setup setup
of initial and location of advancing
and location and target
of advancing boats. boats.
and target

Fine mesh was located

Fine mesh around
was located the free
around thesurface and waves
free surface near the
and waves nearLeft
theand
LeftRight Side boundary
and Right Side boundary
conditions
conditions were damped to avoid wave reflection through introducing verticaltoresistance
were damped to avoid wave reflection through introducing vertical resistance the verticalto the
motion. The motion.
vertical advancingTheship accelerated
advancing shipat 25 knots and
accelerated at 25the target
knots andship
therests
targetat ship
zerorests
speedat under
zero speed
working conditions. The original and W-shape hulls were located in turn at the target ship
under working conditions. The original and W-shape hulls were located in turn at the target ship position,
on theposition,
other hand, theother
on the original hull
hand, was
the used as
original thewas
hull advancing
used as ship for both simulations.
the advancing ship for both simulations.

3. Simulation Results
3. Simulation Results
3.1. Roll Decay Motion of Optimized W-Shape Hull
3.1. Roll Decay Motion of Optimized W-Shape Hull
Figure 5 shows the roll motion results of the optimized shapes, hull geometries are shown in
Figure 5 shows the roll motion results of the optimized shapes, hull geometries are shown in
Figure 3. Roll motion of the original hull is presented in Figure 5 in Case 1, while results of investigated
Figure 3. Roll motion of the original hull is presented in Figure 5 in Case 1, while results of
hulls are shown in Cases 2–11. Cases 2 and 5 increase the roll motion and natural period compared with
the original hull. Cases 3, 7–10 showed similar performance to the original hull; however, time series
of Case 3 showed a considerable increment of the natural period. Cases 4, 6, and 11 reduced roll
motion compared with the Case 1. However, the natural period of Cases 4 and 6 are larger than 2 s
which is larger than the original hull. Roll motion of Case 11 is considerably more damped compared
with proposed hulls and original hull (Case 1). Due to natural period and decrement of roll motion,
a 3-dimensional model of Case 11 was performed to study wake-wash effect. The original hull and
optimized W-shape hull are shown in Figure 6, in which appendages at the optimized hull are attached
on the hull bottom to increase the damping dissipation through vortex behavior.
 damped
larger thancompared
2 s which with proposed
is larger hullsoriginal
than the and original hull
hull. Roll (Case of1).Case
motion Due11toisnatural periodmore
considerably and
decrement
damped of roll motion,
compared a 3-dimensional
with proposed hulls andmodel of Case
original hull11(Case
was performed
1). Due to tonatural
study period
wake-wash
and
effect. The original
decrement hull anda optimized
of roll motion, W-shape
3-dimensional modelhull
ofare shown
Case in Figure
11 was 6, in which
performed appendages
to study wake-wash at
the optimized
effect. hull hull
The original are and
attached on theW-shape
optimized hull bottom to increase
hull are shown inthe damping
Figure dissipation
6, in which through
appendages at
vortex
J.the Sci.behavior.
Mar.optimized hull
Eng. 2020, are attached on the hull bottom to increase the damping dissipation through
8, 544 6 of 11
vortex behavior.

Figure 5. Roll decay results of proposed W-shape hulls.

Figure 5. Roll
Figure 5. Roll decay
decay results
results of
of proposed
proposed W-shape
W-shapehulls.
hulls.

(a) (b)
Figure
(a) viewof
Figure6.6.3-dimensional
3-dimensionalview oforiginal
originaland
andoptimized
optimizedhulls.
hulls.(a)
(a)Original
(b) (b)W-shape
Originalboat;
boat;(b) W-shapeboat.
boat
Figure 6. 3-dimensional view of original and optimized hulls. (a) Original boat; (b) W-shape boat
3.2.
3.2. Resistance
ResistanceForce Force and
andEffective Horsepower
Effective HorsepowerCalculation
Calculation
3.2. Resistance
The
Theresistance Force
resistance and Effective
increased
increased duedue Horsepower
to W-shape at theCalculation
to W-shape bottom
at the of the hull.
bottom of Therefore,
the hull. resistance
Therefore,simulations
resistance
at several
simulations incoming
at flow
several velocities
incoming (5–40
flow knots)
velocitieswere performed
(5–40 knots) to
were verify the
performed increment
to verifyof the
the increment
hydraulic
The resistance increased due to W-shape at the bottom of the hull. Therefore, resistance
force
of theathydraulic
the hull. force
Figure 7a theshows a comparison of athe resistant calculation for ballast (the boat is not
simulations at severalat hull.flow
incoming Figure 7a shows
velocities (5–40comparison
knots) were of performed
the resistant tocalculation for ballast
verify the increment
loaded
(the with seaweed) and full loading condition (the boat is loaded with 2 T of seaweed) of the original
of
J.
theboat
Mar.W-shape
is not force
hydraulic
Sci. Eng. 2020,
loaded with
at the
8, Results
x FOR PEER
seaweed)
hull.
REVIEW
and
Figure 7a full loading
shows condition
a comparison of the(the boat iscalculation
resistant loaded with 2 T of
for ballast
7 of 11
and hull. showed that thehull.
force Results
augmented for flow velocities below 20 knots for
because
(the boat is not loaded with seaweed) and full loading condition (the boat is loaded with 2 flow
seaweed) of the original and W-shape showed that the force augmented T of
the W-shape
velocities hull increases
below 20 knots the submerged
because the transversal
W-shape area.
hull However,
increases the above 20 knotstransversal
submerged the boat moved area.
seaweed) of the original and W-shape hull. Results showed that
in Figure 7 and compared with the CFD estimation. Sea trial measurement agrees very well with the force augmented for flow
upwards
However,and transversal
above 20 knots
knotsarea decreased
the boat the inducing
moved a reduced
upwards force.
velocities
simulation below
results. 20 because W-shape hull and transversal
increases area decreased
the submerged inducing
transversal area.a
reduced force.
However, above 20 knots the boat moved upwards and transversal area decreased inducing a
reduced effective horsepower (EHP) was calculated based in the Equation (3) where ux stands for
The force.
Thein
velocity effective horsepower
the x direction F stands
and (EHP) was for
calculated based in
the resistance forcetheasEquation
function(3) of where ux stands
ux . Figure 7b showsfor
velocity inof
the trends the x direction
ballast and
and full F stands
loading for the
condition resistance
of original and force as function
W-shape of uconsumption
hull. Fuel x . Figure 7b shows
is not
seriously
the trendsaffected
of ballastatand
10 full
andloading
15 knots due to hull
condition shape and
of original optimization and aFuel
W-shape hull. motor of 150 KW
consumption is can
not
provide required power for design speed of 15 knots.
seriously affected at 10 and 15 knots due to hull shape optimization and a motor of 150 KW can
provide required power for design speed ofEHP = ux ⋅ F ( ux ) .
15 knots. (3)
EHP = ux ⋅ F ( ux ) . (3)
A full-scale boat, of which the principal dimension was 9.6 2.6 0.65 m, was
built A
offull-scale
aluminum material
boat, basedthe
of which in principal
the W-Type hull and a sea trial was
dimension wastested
9.6 at 2.6
17 knots under
0.65 the
m, was
ballast
built ofcondition.
aluminumEHP was measured
material trough
based in the sea trial
W-Type hull test,
and and
a seathe result
trial was was displayed
tested in aunder
at 17 knots red circle
the
ballast condition. EHP was measured trough sea trial test, and the result was displayed in a red circle

(a) (b)
Figure 7. Resistance
Resistance and
and effective
effective horsepower. (a)
(a) Resistance
Resistance force
force calculation
calculation (N);
(N); (b)
(b) effective
horsepower (KW).

3.3. Workability of Boats by Wake-Wash

Figure 8 shows the Kelvin waves patterns of a passing ship at several time frames for the W-
shape ship located at a target position. The Kelvin waves pull the target ship and rolls in starboard
and portside. The roll motion goes to decay due to the radiation waves produced by the target ship,
vortex generation, and W-shape hull form. The motion is described next. At t = 6 s, the wave trough
pulls target ship into starboard. From t = 7 s to t = 8 s, the following wave crest approaches the rolling
 J. Mar. Sci. Eng. 2020, 8, 544 7 of 11

The effective horsepower (EHP) was calculated based in the Equation (3) where ux stands for
velocity in the x direction and F stands for the resistance force as function of ux . Figure 7b shows the
trends of ballast and full loading condition of original and W-shape hull. Fuel consumption is not
seriously affected at 10 and 15 knots due to hull shape optimization and a motor of 150 KW can provide
required power for design speed of 15 knots.

EHP = ux · F(ux ). (3)

A full-scale boat, of which the principal dimension (L × B × t) was 9.6 × 2.6 × 0.65 m, was built
(a) (b)
of aluminum material based in the W-Type hull and a sea trial was tested at 17 knots under the
ballast7.condition.
Figure EHP
Resistance waseffective
and measured trough sea trial
horsepower. (a) test, and the force
Resistance result calculation
was displayed in a(b)
(N); redeffective
circle
in Figure 7
horsepower (KW).and compared with the CFD estimation. Sea trial measurement agrees very well with
simulation results.

3.3. Workability of Boats

3.3. Workability of Boatsby Wake-Wash
by Wake-Wash
FigureFigure
8 shows8 shows thethe
Kelvin
Kelvinwaves patterns
waves patterns of of a passing
a passing ship
ship at at several
several time
time frames forframes for the W-
the W-shape
shape ship located at a target position. The Kelvin waves pull the target ship and rolls in and
ship located at a target position. The Kelvin waves pull the target ship and rolls in starboard starboard
portside. The roll motion goes to decay due to the radiation waves produced by the
and portside. The roll motion goes to decay due to the radiation waves produced by the target ship, target ship, vortex
and W-shape hull form. The motion is described next. At t = 6 s, the wave trough pulls
vortexgeneration,
generation, and W-shape hull form. The motion is described next. At t = 6 s, the wave trough
target ship into starboard. From t = 7 s to t = 8 s, the following wave crest approaches the rolling ship
pulls target ship into starboard. From t = 7 s to t = 8 s, the following wave crest approaches the rolling
making it roll in the opposite direction (i.e., port side). The target ship rolls into the starboard again
ship making
and the it
rollroll in the
motion opposite direction
is augmented (i.e., portwave
due to the incoming side). The target ship rolls into the starboard
trough.
again and the roll motion is augmented due to the incoming wave trough.

Figure 8. Wave patterns around advancing and target ships.

Figure 8. Wave patterns around advancing and target ships.
J. Mar. Sci. Eng.Sci.
J. Mar. 2020,
Eng.8,2020,
x FOR PEER REVIEW
8, 544 8 of 11 8 of 11

Figure 9a shows the frontal view of the advancing and target boats at t = 9 s where the maximum
Figure 9a shows the frontal view of the advancing and target boats at t = 9 s where the maximum
motionmotion
to the tostarboard can be observed. After t = 9 s the roll motion decays and the boat rolled into
the starboard can be observed. After t = 9 s the roll motion decays and the boat rolled
the portside
into theagain andagain
portside approximately at t = 10
and approximately at ts,=as
10its, is
asshown in Figure
it is shown in Figure9b,9b,the
thefollowing wave crest
following wave
hits thecrest
boat;
hitshowever, the height
the boat; however, theof the wave
height crest crest
of the wave is smaller than
is smaller thetheprevious
than previous wave trough. The
wave trough.
subsequent wave crest and trough do not seem to influence significantly the roll motion of
The subsequent wave crest and trough do not seem to influence significantly the roll motion ofthe
the target
targetthen,
boat. Since boat. roll
Sincemotion
then, roll
ofmotion
W-shapeof W-shape
boat decaysboat decays
until until the ship
the ship rests
rests at zero
at zero rollangle.
roll angle.

(a)

(b)
Figure 9. Influence of wave trough (upper) and crest (lower) on W-shape boat in ballast condition.
Figure 9. Influence of wave trough (upper) and crest (lower) on W-shape boat in ballast condition.
Advancing boat is located at the left side and W-Shape hull is located at the right side of the figure.
Advancing boat
(a) time = 9is located
s; (b) time =at10the
s. left side and W-Shape hull is located at the right side of the figure.
(a) time = 9 s; (b) time = 10 s.
Figure 10 shows the roll motion of the original and W-shape boats under ballast and full loading
conditions
Figure with and
10 shows thewithout an initial
roll motion heeloriginal
of the angle. The
androll motion ofboats
W-shape the original
underand optimized
ballast hulls
and full loading
follows a similar trend when the starboard of hulls is impacted by the crest and trough of Kelvin
conditions with and without an initial heel angle. The roll motion of the original and optimized hulls
waves produced by the passing boat. Figure 10a,c demonstrates the differences between ballast and
followsfulla similar trend when the starboard of hulls is impacted by the crest and trough of Kelvin
loading condition and it can be concluded that the motion is smaller during full loading condition
waves as produced
the centerby of the passing
gravity boat.
is lower Figure
than ballast10a,c demonstrates
condition. the
The largest differences
roll motion canbetween ballast
be observed in and
full loading condition and it can be concluded that the motion is smaller during full
Figure 10b under ballast condition with an initial heel angle. The maximum roll motion of the original loading condition
as the center of gravity
boat is greater than 10 is ◦lower
. On thethan
otherballast condition.
hand, the W-shapeThe boat largest
shows theroll motion motion
maximum can besmaller
observed in
◦
Figure than 10 which
10b under shows
ballast the effectiveness
condition of this heel
with an initial type angle.
of hull The
formmaximum
to reduce the
rollroll motionofduring
motion the original
wake-wash condition.
boat is greater than 10°. On the other hand, the W-shape boat shows the maximum motion smaller
Table 1 shows a comparison of the maximum roll motion for wake-wash simulations. Percentage
than 10° which shows the effectiveness of this type of hull form to reduce the roll motion during
reduction was calculated based on the difference of the original hull with W-Type hull divided by
wake-wash condition.
the maximum roll angle of original hull. It clearly observed a reduction of the maximum roll motion
especially in full loading condition by reduction of 16.31◦ . W-Type hull definitely reduced the roll
motion in wake wash condition.

(a) (b)
full loading condition and it can be concluded that the motion is smaller during full loading condition
as the center of gravity is lower than ballast condition. The largest roll motion can be observed in
Figure 10b under ballast condition with an initial heel angle. The maximum roll motion of the original
boat is greater than 10°. On the other hand, the W-shape boat shows the maximum motion smaller
J. Mar.10°
than Sci. which
Eng. 2020, 8, 544 the effectiveness of this type of hull form to reduce the roll motion during
shows 9 of 11
wake-wash condition.

J. Mar. Sci. Eng. 2020, 8, x FOR PEER REVIEW 9 of 11

(a) (b)

(c) (d)
Figure
Figure10.
10.Roll
Rollmotions
motionsininballast
ballastconditions
conditions(a)
(a)w/o
w/oinitial
initialheel
heelangle;
angle;(b)
(b)w/ initial heel
w/initial heel angle;
angle; and
and in
in
full loading conditions (c) w/o initial heel angle (d) w/ initial heel angle.
full loading conditions (c) w/o initial heel angle (d) w/initial heel angle.

Table 1 shows a comparisonTable

of 1.
theComparison
maximumofrollmaximum
motion roll
for motion.
wake-wash simulations. Percentage
reduction was calculated based on the difference of the original hull with W-Type hull divided by
Max Roll Reduction Reduction
Initial roll
the maximum HeelangleCondition Hull
of original hull. Type observed a reduction of the maximum roll motion
It clearly Motion (Deg) (Deg) (%)
especially in full loading condition by reduction of 16.31°. W-Type hull definitely reduced the roll
Original 10.54
motion inNowake Ballast
wash condition. 1.24 11.76
initial W-Type 9.30
heel angle
Original 7.91
Full loading
Table 1. Comparison of maximum roll motion. 0.99 12.57
W-Type 6.91
Original Max Roll
8.72 Reduction Reduction
InitialInitial
heel Condition
Ballast Hull Type 0.50 5.69
W-Type Motion8.22(Deg) (Deg) (%)
heel angle
Original
Original 10.54
8.11
Full loading
Ballast 1.24
1.32 11.76
16.31
No initial W-Type
W-Type 6.78
9.30
heel angle Original 7.91
Full loading 0.99 12.57
4. Conclusions W-Type 6.91
Original 8.72
The purpose of thisBallast
paper was to optimize a hull to reduce the roll motion0.50 due to the5.69
Kelvin waves
Initial heel W-Type 8.22
produced by another vessel passing by. The hull was modified using a W-shape, based on a standard
angle Original 8.11
(original) hull whichFullis used to harvest seaweed. The present study was performed
loading 1.32 based16.31
on numerical
W-Type 6.78
simulation. Therefore, algorithm (PISO), numerical schemes, and technique (DFBI) were tested and
the results of mentioned numerical schemes were validated with experimental results [3].
4. Conclusions
The numerical method used in the validation case was used to simulate the roll motion of
The purpose
11 modified hulls of this on
based paper
a 2-Dwas to optimize
model. a hull to
The modified reduce
hull the roll
in which motion decay
the motion due towasthe quicker
Kelvin
waves produced
(case 11) was used bytoanother
generatevessel
a 3-Dpassing
model. by.
TheThe
3-D hull
modelwas modified
was used to using a W-shape,
calculate basedof
the increment onthe
a
standard
hydraulic(original)
force duehull which
to the is used to
optimization ofharvest
the hullseaweed.
and EHP. The presentforce
Resistance study was
and EHPperformed based
were compared
on
withnumerical simulation.
a sea trial of a built Therefore,
boat. The algorithm (PISO), numerical
sea trial measurement agreedschemes,
very well and technique
with (DFBI)
the simulation
were tested and the results of mentioned numerical schemes were validated with experimental
results [3].
The numerical method used in the validation case was used to simulate the roll motion of 11
modified hulls based on a 2-D model. The modified hull in which the motion decay was quicker (case
11) was used to generate a 3-D model. The 3-D model was used to calculate the increment of the
J. Mar. Sci. Eng. 2020, 8, 544 10 of 11

results. In addition, the 3-D model was used to study and to compare the roll motion for the original
and optimized seaweed hull under wake wash produced by a passing ship. In addition, ballast
and full loading conditions were compared for the original and the W-shape hulls. The advancing
boat produced Kelvin waves which produced a roll motion in the boat located at the target position.
The Kelvin waves pull the target ship and it rolls in starboard and portside. After the maximum roll
motion toward the starboard was reached (t = 9 s) the roll motion decayed due to the radiated waves
produced by the target ship.
Time series of roll motion was studied for ballast and full loading condition of original and
optimized hull. The weight of the workers was located at the center and towards one side of the boat
during working hours (initial heel angle). Based on the numerical simulation results, roll motion
is smaller for full loading condition as the center of gravity is lower than in the ballast condition.
The larger roll motion was observed under ballast condition with an initial heel angle. The optimized
W-shape reduced the roll motion with an initial heel angle. The original roll angle was greater than 10◦
while the roll motion was below 10◦ for the W-shape. Reduction of roll motion was clearly observed
by comparison of the maximum roll angle. In no initial heel angle case, the maximum reduction
in terms of roll motion reduction was observed during ballast condition with a reduction of 1.24◦ .
The effectiveness of the W-shape is clearly observed under full loading condition with initial heel angle
where a reduction of 1.32◦ (16.31%) was observed in comparison with the original hull. Roll motion of
optimized hull was reduced due to the W-shape demonstrating the effectiveness of the W-shape to
decrease roll motion in ballast and full loading condition considering a lateral change of the center of
gravity due to the worker’s weight.

Author Contributions: Conceptualization, J.-C.P.; Data curation, D.-H.K.; Formal analysis, M.P.E.H.; Funding
acquisition, S.-B.L.; Resources, J.-C.P.; Supervision, J.-C.P.; Writing—original draft, M.P.E.H.; Writing—review and
editing, J.-C.P. All authors have read and agreed to the published version of the manuscript.
Funding: This research was a part of the project titled ‘Development of coastal garbage collection technology in
difficult to access areas’, funded by the Ministry of Oceans and Fisheries, Korea.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Ferdouse, F.; Holdt, F.; Smith, S.L.R.; Murúa, P.; Yang, Z. The Global Status of Seaweed Production, Trade and
Utilization; Food and Agriculture Organization of the United Nations: Rome, Italy, 19 November 2018.
2. Bellafiore, D.; Zaggia, L.; Broglia, R.; Ferrarin, C.; Barbariol, F.; Zaghi, S.; Lorenzetti, G.; Manfe, G.;
De Pascalis, F.; Benetazzo, A. Modeling ship-induced waves in shallow water systems: The Venice experiment.
Ocean. Eng. 2018, 155, 227–239. [CrossRef]
3. Ghani, M.P.A.; Rahim, M.A.R.M.A. The prediction of wake wash in the towing tank. J. Mek. 2008, 26, 129–140.
4. Gu, Y.; Day, S.; Boulougouris, E. A Study on the Effects of Bilge Keels on Roll Damping Coefficient.
In Proceedings of the 12th International Conference on the Stability of Ships and Ocean Vehicles, Glasgow,
UK, 14–19 June 2015.
5. Shi, F.; Malej, M.; Smith, J.M.; Kirby, J.T. Breaking of ship bores in a Boussinesq-type ship-wake model.
Coast. Eng. 2018, 132, 1–12. [CrossRef]
6. Jung, K.H.; Chang, K.A.; Jae, J.H. Viscous Effect on the Roll Motion of a Rectangular Structure. J. Eng. Mech.
2006, 132, 190–200. [CrossRef]
7. SIEMENS. STAR-CCM+ User Guide; Version 11.06; SIEMENS: Munich, Germany, 2016.
8. Menter, F.R.; Kuntz, M.; Langtry, R. Ten Years of Industrial Experience with the SST Turbulence Model.
In Turbulence, Heat and Mass Transfer; Hanjalić, K., Nagano, Y., Tummers, M., Eds.; Begell House Inc.:
New York, NY, USA, 2003; pp. 625–632.
9. Wilcox, D.C. Turbulence Modeling for CFD, 2nd ed.; DCW Industries Inc.: La Canada Flintridge, CA, USA,
1998; pp. 172–180.
10. Caponnetto, M.; Bučan, B.; Pedišić-Buča, M.; Perić, M.; Pettinelli, C. Simulation of Flow and Motion of
High-Speed Vessels. In Proceedings of the 12th International Conference on Fast Sea Transportation,
Amsterdam, The Netherlands, 2 December 2013.
J. Mar. Sci. Eng. 2020, 8, 544 11 of 11

11. Calderer, A.; Kang, S.; Sotiropoulos, F. Level set immersed boundary method for coupled simulation of
air/water interaction with complex floating structures. J. Comput. Phys. 2014, 277, 201–227. [CrossRef]
12. Bhattacharyya, R. Dynamics of Marine Vehicles; Wiley: New York, NY, USA, 1978.
13. Stumbo, S.; Fox, K.; Dvorak, F.; Elliot, L. The Prediction, Measurement, and Analysis of Wake Wash from
Marine Vessels. Mar. Technol. 1999, 36, 248–260.
14. Garel, E.; Fernández, L.L.; Collins, M. Sediment resuspension events induced by the wake wash of deep-draft
vessels. Geo-Mar. Lett. 2008, 28, 205–211. [CrossRef]
15. Osborne, P.D.; McDonald, N.J.; Reynolds, W.J. Response of mixed sediment beaches to wake wash
from passenger only fast ferries: Rich Passage, Washington. In Coastal Engineering 2006; Smith, J.M., Ed.;
World Scientific: Singapore, 2007; Volume 5, pp. 3105–3116.

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).