International Journal of Naval Architecture and Ocean Engineering 13 (2021) 86e101

Contents lists available at ScienceDirect

International Journal of Naval Architecture and Ocean Engineering

journal homepage: http://www.journals.elsevier.com/
international-journal-of-naval-architecture-and-ocean-engineering/

The effects of geometrical buoy shape with nonlinear Froude-Krylov

force on a heaving buoy point absorber
Sung-Jae Kim a, b, Weoncheol Koo a, *, Moo-Hyun Kim b
a
Department of Naval Architecture and Ocean Engineering, Inha University, Incheon, Republic of Korea
b
Department of Ocean Engineering, Texas A&M University, College Station, TX, USA

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

Article history: This study examined the effects of buoy shape and Nonlinear Froude-Krylov force (NFK) on a heaving-
Received 4 August 2020 buoy-type Wave Energy Converter (WEC). Based on the Maclaurin expansion, the theoretical solutions
Received in revised form of the NFK were derived for three different buoy shapes; hemispheric buoy, circular vertical cylinder, and
3 December 2020
truncated conical cylinder. A hydraulic power take-off system was adopted, and the latching control
Accepted 27 January 2021
Available online 27 February 2021
strategy was applied to maximize the extracted power from the WEC. The nonlinear effects of the
Froude-Krylov force and restoring force on the heaving point absorber were investigated by comparing
the heave Response Amplitude Operator (RAO) and time-averaged power extraction. The results showed
Keywords:
Nonlinear Froude-Krylov force
that the conventional linear analyses were overestimated by up to 50% under the high amplitude wave
Buoy shape condition. The latching control strategy was the most effective when peak wave period of regular or
Wave energy converter irregular wave was 0.4e0.45 times the heave natural period of the buoy.
Response amplitude operator © 2021 The Society of Naval Architects of Korea. Production and hosting by Elsevier B.V. This is an open
Hydraulic power take-off system access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Latching control strategy

1. Introduction the nonlinear hydrodynamic effects of heaving-buoy-type WECs.

Weakly nonlinear analysis or fully nonlinear analysis was per-
The Wave Energy Converter (WEC) can be classified as formed to consider the nonlinearity of floating body motion. The
moveable-body type, oscillating-water-column type, and over- weakly nonlinear analysis often considered the Nonlinear Froude-
topping/overflow type according to their energy-conversion Krylov force (NFK) in addition to the first-order (linear) hydrody-
methodology. Among these, the movable-body-type WEC gener- namic analysis based on the radiation/diffraction problem. In some
ates powers by converting wave energy to WEC’s kinetic energy. cases, the second-order diffraction force was also adopted. The
WECs can also be classified according to buoy’s mode of motion, nonlinear Froude-Krylov force, which is the force from undisturbed
such as heaving buoy type, pendulum type, oscillating surge-type incident waves, can be obtained by integrating the pressure over
(Oyster type) and so on. The heaving-buoy-type WEC can be the instantaneous wetted buoy surface. The method of calculating
extended easily to a renewable energy farm or a hybrid renewable the NFK may include a remeshing scheme or a theoretical solution.
energy system. The heaving-buoy-type WEC has very large motion The remeshing numerical scheme generates the instantaneous
around its natural frequency because of the resonance phenome- wetted-surface meshes according to the buoy motion and wave
non, but the resonance region is often narrow-banded. Therefore, a elevation around the body (Jang and Kim, 2020). Babarit et al.
control technique is required to increase energy extraction (Folley (2009) and Merigaud et al. (2012) applied the remeshing scheme
et al., 2015). To apply the control technique, the interaction be- to evaluate the nonlinear Froude-Krylov force and the double-
tween the floating body and power take-off (PTO) as well as the frequency diffraction force on a WEC buoy, respectively. Giorgi
range of nonlinear factors due to the application of the control and Ringwood (2017) derived a theoretical solution of the
method, should be considered (Nazari et al., 2013; Kalofotias, 2016). Froude-Krylov force for an axisymmetric heaving buoy using the
Since 2000, numerous studies have been carried out to estimate Maclaurin expansion method under regular wave conditions. They
described the wetted body surface theoretically based on the cy-
lindrical coordinate system. They then simplified the instantaneous
* Corresponding author. water line around the body based on the long wave assumption and
E-mail address: [email protected] (W. Koo). adopted a simple PTO modeling as a linear damper. They
Peer review under responsibility of The Society of Naval Architects of Korea.

https://doi.org/10.1016/j.ijnaoe.2021.01.008
2092-6782/© 2021 The Society of Naval Architects of Korea. Production and hosting by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://
creativecommons.org/licenses/by-nc-nd/4.0/).
S.-J. Kim, W. Koo and M.-H. Kim International Journal of Naval Architecture and Ocean Engineering 13 (2021) 86e101

considered nine types of different numerical approaches, including

ð
∞
Computational Fluid Dynamics (CFD). They reported that the 2
KðtÞ ¼  bðuÞcosðutÞdu
nonlinear Froude-Krylov forces, nonlinear restoring forces, and p
nonlinear viscous forces are very important in applying the control 0
(4)
strategies. Giorgi and Ringwood (2018) derived a theoretical solu- ð
∞
2
tion of the nonlinear Froude-Krylov force for a three-degree-of- ¼ ðMa ðuÞ  Ma ð∞ÞÞu sinðutÞdu
freedom point absorber under regular wave conditions. They p
0
showed that the calculation time using the theoretical solution was
shorter than that using the remeshing technique. Kim et al. (2019b) The frequency-dependent diffraction force can be expressed as
considered the nonlinear Froude-Krylov force and hydraulic PTO Eq. (5).
system under regular wave conditions. Parametric studies on the ð
vfD
hydraulic PTO condition of a hemispherical heaving-buoy WEC FD ðuÞ ¼ r nz dS (5)
were carried out. The capture width of the WEC decreased in the vt
SB
longwave region as the wave amplitude increased.
This study derived the theoretical solutions of nonlinear Froude-
X
nw  
Krylov forces on three different geometries of WEC buoys (a FD ðtÞ ¼ Ai FD i ðuÞcosðui t þ qi þ qr i Þ (6)
hemispherical buoy, a vertical circular cylindrical buoy, and a i¼1
truncated conical buoy) based on Giorgi and Ringwood (2017). The
theoretical solution of the nonlinear Froude-Krylov force was where r, fD , and nz is the water density, diffraction velocity po-
verified by comparison with the results of the remeshing technique. tential, and z-directional component of the normal vector, respec-
Based on them, the hydraulic PTO system and latching control tively. Ai , FD i , ui , qi , and qr i are the incident wave amplitude,
strategy were applied to estimate WEC’s hydrodynamic and power diffraction force in the frequency domain, wave frequency, phase
take-off performances. The effects of the nonlinear Froude-Krylov angle of the diffraction force, and random phase angle of the ith
force were identified, and the relationship between the incident wave, respectively. nw is the number of incident wave components.
wave period (or amplitude) and latching control strategy under Under regular single wave conditions, only one frequency compo-
regular wave conditions was investigated. Finally, based on the nent was applied, and the random phase angle was set to zero.
results of regular wave analysis, the effects of the nonlinear Froude- Under irregular wave conditions, the amplitude of the i-th wave
Krylov force on WEC buoys with latching control under irregular can be obtained (Eq. (7)) using the PM spectrum (Eq. (8)).
wave conditions were estimated, which can be regarded as an qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
extension of the previously developed latching control strategy for Ai ¼ 2Sw ðui Þdu (7)
regular waves.
!
173H2 692
Sw ðuÞ ¼ 4 5s exp (8)
2. Mathematical formulation T1 u T14 u4

2.1. Equation of motion for a rigid body where Sw , Hs , and T1 are PM wave spectrum, significant wave
height, and mean centroid wave period, respectively. The com-
All forces acting on a heaving body type WEC can be expressed mercial hydrodynamic analysis software WAMIT (WAMIT, 2019)
as was used to calculate the hydrodynamic coefficients (added mass,
radiation damping coefficient, and diffraction force). The program
m€x ¼ FD þ FR þ FFK þ Fstatic þ Fvis þ FPTO  W (1) is based on the wave green function and boundary element
method.
where m and €x are the mass of a buoy and acceleration of vertical
motion, respectively. FD , FR , FFK , Fstatic , Fvis , FPTO , and W denote the 2.2. Froude-Krylov force
diffraction force, radiation force, Froude-Krylov force, hydrostatic
force, viscous damping force, PTO force, and gravity force, respec- The Froude-Krylov force is the force acting on the body from an
tively. Among these force components, the radiation force and undisturbed incident wave. This force can be obtained using the
diffraction force can be obtained by solving the radiation problem incident wave velocity potential, as shown in Eq. (9) and Eq. (10).
and diffraction problem, respectively. The radiation force can be
decomposed into two terms: one that is proportional to the ac- !  vf
p x ; t ¼ r I ¼ rgAekz cosðkx cos b þ ky sin b  utÞ (9)
celeration and the other that is proportional to the velocity of a vt
buoy. This force also satisfies the linearity of functions and the
ð ð
Bound Input and Bound Output stability (BIBO), so it can be ! 
FFK ¼ p x ; t nz dS ¼  rgAekz cosðkx cos b þ ky sin b
expressed using the convolution terms.
SB SB (10)

FR ðuÞ ¼ Ma ðuÞ€xðuÞ  bðuÞx_ ðuÞ (2)  utÞ nz dS

ðt !
where x is the position of a collocation point ðx; y; zÞ on the
FR ðtÞ ¼ Ma ð∞ÞxðtÞ  Kðt  tÞx_ ðtÞdt
€ (3) boundary element. fI , g, k, b, and SB are the incident wave velocity
0 potential, gravitational acceleration, wave number, angle of the
incident wave, and body surface, respectively. To express the
where Ma and b are the added mass and radiation damping coef- theoretical solution of nonlinear Froude-Krylov force on an
ficient. KðtÞ is the memory function or impulse response function axisymmetric body, the coordinate system should be transferred
expressed as Eq. (4). from the Cartesian coordinate to the cylindrical coordinate (Giorgi
87
S.-J. Kim, W. Koo and M.-H. Kim International Journal of Naval Architecture and Ocean Engineering 13 (2021) 86e101

and Ringwood, 2017). In this scheme, the equation of the body different, but the upper shape of all floating buoys was fixed in the
radius along the z-axis was derived according to the shape of the form of vertical circular cylinders. Therefore, the effective range of
heaving buoy. The nonlinear Froude-Krylov force is obtained by the z-axis is expressed as Eq. (13).
integrating the incident-wave pressure over the instantaneous

wetted body surface that changes according to the body motions s1 ¼ minðhðtÞ; xÞ
(13)
and incident wave elevations at every time step. In addition, the s2 ¼ x  d
water particle velocity and hydrodynamic pressure above the mean
water level are expressed by vertical stretching. where h and d are wave elevation at the center point of a buoy and
draft of a buoy, respectively. Eq. (12) can be converted to Eq. (14),
2.2.1. Theoretical solution of the Froude-Krylov force for a which can be integrated directly using the Maclaurin expansion
hemispherical buoy method to obtain the Froude-Krylov force.2
Giorgi and Ringwood (2017) transformed the Cartesian co-

2ðp sð1

FFK ¼ rgA cos ut ekz Csphere ðs; qÞðs  xÞdsdq

0 s2 (14)
 qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2  qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 4  qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 6
k R2  ðs  xÞ2 cos q k R2  ðs  xÞ2 cos q k R2  ðs  xÞ2 cos q
Csphere ðs; qÞ ¼ 1  þ  þ/
2! 4! 6!

ordinates, ðx;y; zÞ, into the cylindrical coordinates system ðr; q; sÞ to

derive the theoretical solution of Froude-Krylov force acting on a
hemispherical buoy as Fig. 1(a). The equation of a radius along the 2.2.2. Theoretical solution of the Froude-Krylov force for a vertical
z-axis on a hemisphere can be described using Eq. (11). cylindrical buoy
For a circular cylinder, the bottom surface of the buoy is only
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
affected by the incident wave for the vertical Froude-Krylov force.
r ¼ f ðsÞ ¼ R2  ðs  xÞ2 (11)
Therefore, the coordinate system of the bottom can be expressed as
Eq. (15)(see Fig. 1(b)). The product of dS and nz can be presented as
where R and x are the radius and vertical displacement of the
Eq. (16) according to the coordinate transformation.
hemispherical buoy, respectively. The vertical Froude-Krylov force
can be obtained as Eq. (12).
8
< xðqÞ ¼ r cos q
yðqÞ ¼ r sin q (15)
:
2ðp sð1 z ¼ s2
FFK ¼ rgAekz cosðkf ðsÞcos q  utÞf 0 ðsÞf ðsÞdsdq (12)
0 s2 nz dS ¼ rdrdq (16)
In this study, the lower shape of the WEC floating buoy was Substituting Eqs. (15) and (16) into Eq. (10) and applying the

Fig. 1. Schematic coordinate system of various floating bodies for the Froude-Krylov force.

88
S.-J. Kim, W. Koo and M.-H. Kim International Journal of Naval Architecture and Ocean Engineering 13 (2021) 86e101

cosine law and Maclaurin expansion, the theoretical solution of the ð

Froude-Krylov force for a circular vertical cylindrical buoy is final- Fstatic ¼ ðrgzÞnz dS (22)
ized as Eq. (17).
SB

!
kðxdÞ k2 R4 k4 R6 3 k6 R8 5
FFK ¼ r rgAe cosðutÞ R p  2
pþ p pþ/ (17)
2! 4 4! 6 4 6! 8 8

This can be revised by transforming the coordinate system from

the Cartesian coordinate to the cylindrical coordinate system. First,
the static force for the hemispherical buoy can be expressed as Eq.
2.2.3. Theoretical solution of the Froude-Krylov force for a (23)
truncated conical buoy
sð1 
To calculate the Froude-Krylov force for a truncated conical   1 3 1 2 
buoy, it should be divided into the lateral and bottom parts of a Fstatic ¼  2prg s2  s ds ¼  2prg x  s1 þ d2 x
6 2
truncated cone as Fig. 1(c). For the side part, the equation for the xd
radius along the z-axis can be expressed as Eq. (18). 1 3 
þ s1 þ d3
f ðsÞ ¼ aðs  xÞ þ R (18) 3
(23)
where a is the slope of the side part of a truncated cone.
The static force for the cylindrical buoy can be described using
Substituting Eq. (18) into Eq. (12) and applying the cosine law and
Eq. (24)
Maclaurin expansion, the Froude-Krylov force of the side part can
be obtained as Eq. (19).
Fstatic ¼  rgAw ðx  dÞ (24)
The static force for the truncated conical buoy should be

2ðp sð1

FFK slope ¼ argA cosðutÞ eks Ccone ðs; qÞdsdq

0 s2 (19)
k f ðsÞ cos
2 3 2
q k f ðsÞ cos
4 5 4
q k f ðsÞ cos
6 7 6
q
Ccone ðs; qÞ ¼ f ðsÞ  þ  þ/
2! 4! 6!

At the same time, using Eq. (17), the Froude-Krylov force of the calculated separately for the lateral and bottom parts. For each part,
bottom part of a truncated cone can be expressed as Eq. (20). the formula can be expressed using Eqs. (25) and (26). Each static

!
kðxdÞ k2 R04 k4 R06 3 k6 R08 5
FFK bottom ¼ rgAe cos ut R0 p  2
pþ p pþ/ (20)
2 4 4! 6 4 6! 8 8

force can be combined and expressed as Eq. (27).

where R0 is the radius of the bottom of the truncated conical buoy.
Finally, the total Froude-Krylov force can be obtained.
1 3 R 2 n a 2 
Fstatic slope ¼  2parg
ax  x þ Rd  s1 þ a2
FFK ¼ FFK slope þ FFK bottom (21) 6 2 2
o a   R 
 þ s þ d þ s1  d
3 3 2 2
(25)
3 1 2

2.3. Hydrostatic force

Fstatic bottom ¼  rgðx  dÞR02 p (26)
The static force can be obtained using the hydrostatic pressure
acting on the floating body.
89
S.-J. Kim, W. Koo and M.-H. Kim International Journal of Naval Architecture and Ocean Engineering 13 (2021) 86e101

Table 1
Geometric specifications of the three different buoys.

Item [unit] Hemisphere (Model 1) Circular cylinder (Model 2) Truncated cone (Model 3)

Radius [m] 2.5 2.5 2.5

Draft [m] 2.5 1.667 2.5
Weight [kg] 16,362 16,362 16,362
Water plane area [m2] 19.634 19.634 19.634

2.4. Viscous damping force

1 2 3 n   o To consider the viscous effect, the drag force of the Morrison

Fstatic ¼  prg a x  aRx2 þ 2adR  a2 s21 þ d2 þ R02 x
3 equation was adopted. The drag force on the body can be described
 2    as Eq. (28).
2a
þ s31 þ d3 þ aR s21  d2  dR02
3 1
(27) Fvis ¼  rCd Ap ðx_ ðtÞ  vðtÞ Þjx_ ðtÞ  vðtÞj (28)
2

Fig. 2. Numerical models for three different buoys.

90
S.-J. Kim, W. Koo and M.-H. Kim International Journal of Naval Architecture and Ocean Engineering 13 (2021) 86e101

the drag coefficient of cylindrical buoys from 0.65 to 1.05, and

Sarpkaya (1986) estimated experimentally that the drag coefficient
was between 0.8 and 1.0 for a small Keulegan-Carpenter number.
Guo et al. (2018) reported a drag coefficient of 0.93 through ex-
periments. Based on these studies, the drag coefficients of the
hemispherical floating body and a cylindrical body were set to 0.6
and 0.95±0.05, respectively, in this study. The drag coefficient of a
conical buoy in a few studies was set to 0.8±1.0 (Giorgi et al., 2016).

2.5. PTO force

As a PTO system of a heaving buoy type WEC, a hydraulic system

or a linear generator is generally applied. The hydraulic PTO system
consists of a hydraulic cylinder that converts vertical buoy motion
to fluid flow in a hydraulic circuit, a hydraulic accumulator that
protects the hydraulic system or maintains residual pressure, a
hydraulic motor and a power generator that produces energy from
fluid flow. The hydraulic PTO system can be represented in the form
Fig. 3. Comparison of the heave RAOs on various shapes of a buoy.
of an approximate coulomb damping force according to its char-
acteristics (Falcao, 2008; Babarit et al., 2012; Kim et al., 2019c).

where Cd , Ap , x_ , and v are the drag coefficient, projected area, buoy FPTO ¼ minðGjx_ j; DpSc Þsignðx_ Þ (29)
velocity, and water particle velocity, respectively. To represent the
viscous damping force accurately, it is very important to estimate where Dp and Sc are the pressure difference between High Pressure
the appropriate drag coefficient according to the properties of the (HP) and Low Pressure (LP) and the sectional area of a hydraulic
floating body. For this, the drag coefficients were estimated pri- cylinder, respectively. G is the slope to reduce the numerical error
marily from free decay tests or forced oscillation tests. Recently, due to abrupt changes in the PTO force, and as applied by Babarit
some studies were conducted to estimate the drag coefficients et al. (2012), applied 100 times the added mass value. The instan-
through CFD analysis. In this study, the drag coefficient of the taneous extraction power can be obtained by multiplying the PTO
previous studies was investigated. Based on this, the drag coeffi- force by the velocity of a buoy. Time-averaged extraction power can
cient suitable for this analysis model was selected. Compared to the be expressed as Eq. (30).
wave tank experiment, Zurkinden et al. (2014) calculated the drag
coefficient of a hemisphere, 0.254 m in diameter, to be 0.2, and Lok ðte
1
et al. (2014) estimated the drag coefficient of a hemisphere, 0.17 m Pex ¼ FPTO x_ dt (30)
te  ts
in diameter to 0.175. Giorgi and Ringwood (2017) estimated the ts
drag coefficient of a hemispherical body with a radius of 2.5 m by
0.6 compared to CFD analysis and minimizing the errors. For the where ts and te are the starting and end times to estimate the time-
drag coefficient of a cylindrical buoy, many studies have been averaged extraction power, respectively. Under regular wave con-
conducted since the 1990s. Gudmestad and Meo (1996) estimated dition, te  ts was set to be the same as the wave period. A total of
3 h of data was analyzed for irregular wave analysis.

Fig. 4. Comparison of the heave RAOs on the calculation methods of the Froude-Krylov
force for a hemispherical buoy (Lin: frequency-domain linear analysis, ME: Maclaurin Fig. 5. Comparison of the heave RAOs for various methods of the Froude-Krylov force
expansion for NFK force). Model 2 (Cylindrical buoy) and Model 3 (Truncated conical buoy).

91
S.-J. Kim, W. Koo and M.-H. Kim International Journal of Naval Architecture and Ocean Engineering 13 (2021) 86e101

Fig. 6. Comparison of the Froude-Krylov forces on various analysis approaches according to the buoy displacement (a: Hemispherical buoy, b: Circular cylindrical buoy, c: Truncated
conical buoy, u ¼ 1 rad/s).

2.6. Latching control strategy applied to allow latching control in this study, as expressed in Eq.
(32).
The latching control is a type of discontinuous phase control 
technique. This technique controls the phase of buoy displacement xðtÞ ¼ xðtls Þ; tls < t < tle
(32)
by locking and releasing the movement of the buoy for a certain x_ ðtÞ ¼ 0; otherwise
time when the displacement of the buoy reaches its peak due to
waves. Because the latching duration is the only variable, the where tls and tle denote the starting time and end time of the
latching control technology is very simple, and its effectiveness has latching control, respectively. The effect of the Froude-Krylov force
been proven through various researches on control technologies was estimated by applying latching control with the optimal
(Hal et al., 2011). The latching duration can be determined using the latching duration. The latching control can be applied easily by
relationship with the heave natural period of the body (Sheng et al., locking and releasing the valve in the hydraulic circuit for a certain
2015; Kim et al., 2019a). time when the sign of the hydraulic cylinder stroke speed changes.

Tc  T0 3. Numerical model and results

TL ¼ (31)
2
In this study, numerical modeling with the same radius, weight,
where TL , Tc , and T0 are the latching duration, characteristic wave and waterplane area of a hemispherical buoy, cylindrical buoy, and
period, and the natural period of a buoy, respectively. Under the truncated conical buoy was performed. Table 1 lists the specifica-
regular wave condition, the characteristic period is the wave tions of three different floating bodies. The draft of the truncated
period, and in the irregular wave condition, the characteristic cone was set to be the same as the draft of the hemispherical buoy
period is the same as the energy period, Te (Sheng et al., 2015; Cho, and cylindrical cylinder. The weights of all buoys were the same.
2015). In addition, the motion constraint of the floating body was Fig. 2 shows the numerical model based on Table 1; the number of
92
S.-J. Kim, W. Koo and M.-H. Kim International Journal of Naval Architecture and Ocean Engineering 13 (2021) 86e101

Fig. 7. Comparison of the heave RAOs including the drag forces (a: Hemispherical buoy, b: Circular cylindrical buoy, c: Truncated conical buoy).

nodes was 1200, 1770, and 1576, respectively. The Runge-Kutta 4th- the Maclaurin expansion to the 4th-term was similar to the linear
order time integration method was adopted for time marching, and analysis results. In Fig. 5, convergence tests were also performed for
its time integration step was set to T/200. the number of applied terms in the Maclaurin expansion for Model
Fig. 3 presents the heave RAOs of three different types of buoys 2 (cylindrical buoy) and Model 3 (truncated cone buoy). To obtain
for WEC. In the figure, Z means the amplitude of the heave motion accurate heave RAOs, models 2 and 3 were applied up to the 4th-
displacement. All bodies had the same mass and restoring co- and 6th-terms of the theoretical solution in the Maclaurin
efficients. On the other hand, owing to differences in the hydro- expansion.
dynamic coefficients, the cylindrical buoy with a relatively small Fig. 6 compares the nonlinear Froude-Krylov forces calculated
draft has a heave natural frequency of approximately 1.8 rad/s, and by the theoretical solutions based on a Maclaurin expansion and
the hemispherical and the conical bodies have a heave natural the remeshing scheme according to the buoy displacement. The
frequency of 2.0 rad/s. For the cylindrical buoy, the maximum heave change in the incident wave potential in the z-direction above the
response was approximately 2.8 times the incident wave mean water level was expressed by vertical stretching. For
amplitude. simplicity, it was assumed that the incident wavelength was long
Fig. 4 shows the heave RAOs of a hemispherical buoy from enough to ignore the diffraction effect, and the variation of wave
different methods of the Froude-Krylov force. Two different elevation was the same as the heave response of the buoy. To apply
methods were adopted to evaluate the Froude-Krylov force: con- the remeshing technique, the numerical model used in Fig. 2 for
ventional frequency-domain linear analysis and a method based on frequency-domain analysis was used. The linear Froude-Krylov
the Maclaurin expansion, as shown in chapter 2.2. Convergence force means the force calculated at the mean body posion
tests were performed to investigate how many terms of the theo- without taking into account the instantaneous wetted body sur-
retical solutions in the Maclaurin expansion are needed to calculate face. The results of the theoretical solution agreed well with those
the accurate Froude-Krylov force. The effects of buoy displacement of the remeshing scheme. On the other hand, the linear Froude-
and wave elevation around the buoy were not considered in this Krylov force had a constant value regardless of the buoy displace-
test. In Fig. 4, the heave RAO applying the Froude-Krylov force from ment. Through Fig. 6, as the buoy displacement increases, the

93
S.-J. Kim, W. Koo and M.-H. Kim International Journal of Naval Architecture and Ocean Engineering 13 (2021) 86e101

Fig. 8. Time-averaged extraction power on various PTO conditions (u ¼ 0.8 rad/s, H/l ¼ 1/50, Model 1(Hemispheric buoy)).

magnitude of the Froude-Krylov force increased approximately condition in the wave wave region of 0.25≪u=un ≪0.75. The time-
seven times more than the linear FK force. averaged generated power (P ex ) was divided by the square of
Fig. 7 presents the heave RAOs of the buoy under two different incident wave amplitude (A) and the buoy diameter (B) as P ex =A2 B
wave steepness conditions with various drag coefficients. The re- to estimate and compare the value according to the change in the
sults were calculated through the theoretical solutions with a given incident wave period. The maximum time-averaged extraction
optimal series of a Maclaurin expansion. Even when the drag co- energy occurred when the PTO force was 70% of the vertical exci-
efficients were applied, the sensitivity to the drag coefficient was tation force of the buoy (Fz ¼ FFK þ FD ). In this way, the PTO con-
quite small because the vertical motion of the body did not change ditions for maximum energy extraction under various wave
significantly. In this study, the drag coefficient was 0.6, 0.95, and 0.8 conditions were calculated for each numerical model of the buoy.
for the hemispherical buoy, cylindrical buoy, and truncated conical Based on the previously estimated conditions of the hydraulic
buoy, respectively. The vertical motion decreased in the resonance PTO system as Fig. 8, the results of the weakly nonlinear time-
period when the wave steepness was large (H=l¼1/20), but the domain analysis of a hemispherical buoy for a WEC were
motion increased slightly in the other wave periods. To estimate the compared with the results of linear time-domain analysis (Fig. 9) at
energy extraction of WEC, the hydraulic PTO system was modeled specific wave and PTO condition (u=un ¼0.4, FPTO =Fz ¼0.7). A
numerically and applied. The optimal hydraulic PTO condition was latching duration of 2.356 s, which is the optimal latching time at
determined from the case of a small amplitude wave (H= l¼1/50) the corresponding frequency, was applied. The results of linear
for each numerical model. The hydraulic PTO condition can be analysis and nonlinear analysis agreed well when the incident wave
expressed by the difference between the High Pressure (HP) and height was low at 0.25 m (H=l¼ 1/240). This is because the incident
Low Pressure (LP) from the hydraulic cylinder and the flow rate. wave belongs to the linear wave region. Therefore, the nonlinearity
Fig. 8 presents the time-averaged extraction power under various is small. When the wave amplitude increased, however, the relative
PTO forces applied to three types of buoys for WEC. Various PTO heave response (Fig. 9 (a)) and the time-averaged extraction power
damping values were applied to obtain the optimal PTO force compared to wave amplitude (Fig. 9 (b)) decreased. The static force

94
S.-J. Kim, W. Koo and M.-H. Kim International Journal of Naval Architecture and Ocean Engineering 13 (2021) 86e101

Fig. 9. Time-series of hemispherical buoy displacements, extraction powers, excitation forces and static forces ( u=un ¼ 0.4, FPTO =Fz ¼ 0.7, Lin: Linear analysis, WN: Weakly
nonlinear analysis).

(Fig. 9 (d)) also decreased with increasing wave amplitude. Fig. 9 (c) Fig. 10 compares the heave RAOs and time-averaged extraction
shows the time series of the wave excitation forces, the sum of the powers of three different WEC buoy models under a range of
Froude-Krylov force, and diffraction force under various wave incident wave amplitudes. The analysis was performed only in the
amplitude conditions. Because the diffraction force does not differ longwave condition, where the latching control strategy was
between linear analysis and weakly nonlinear analysis, the differ- known to be effective. For a clear comparison, the heave RAOs and
ence in excitation force was due to the nonlinear Froude-Krylov time average extraction powers were compared when no latching
force. As the wave amplitude increased, the nonlinear effect of control strategy was applied. The time-averaged extraction power
the excitation force increased, as shown in Fig. 9(c). The latching was nondimensionalized by the incident wave flux, which is the
control strategy enables the extraction of large amounts of energy same as the capture width. As shown in Fig. 9, as the amplitude of
with the effect of rapidly moving the body upward or downward. the incident wave increased, the displacement of the buoy
This technique can constrain the body motion during the latching decreased due to the influence of the nonlinear Froude-Krylov
duration, making a 90 difference in phase angle between excita- force, and the dimensionless time-averaged extraction power
tion force and buoy motion. This can be confirmed by comparing decreased. As the incident wave amplitude was increased from
the time when the buoy displacement is zero, and when the 0.25 m to 1.0 m, the extraction power (P=A2 B) decreased from 50%
external force is zero. In Fig. 9 (c), indication (1) is when the buoy to 20%. Based on these results, linear analysis using latching control
moves downward, and (2) is when the buoy moves upward. When techniques provides overestimated results. In addition, when the
the difference was 90 , the excitation force acting on the body was incident wave frequency was 0.4e0.5 times the natural frequency
maximized along the direction of movement of the buoy after of the buoy, the heave RAO and time-averaged extraction power
releasing the buoy motion according to the latching control strat- were maximized due to the latching control technique. Therefore,
egy. As the incident wave amplitudes increased to 0.25 m, 0.5 m, latching control is very effective when the wave height is small, and
and 1.0 m, the magnitude of the buoy motion decreased by 1%, the incident wave frequency is 0.4e0.5 times the heave natural
18.6%, and 32.3%, respectively, compared to the linear analysis re- frequency of the buoy. Table 2 lists the maximum time-averaged
sults. The external force was also reduced at a similar rate. extraction power for the three models under each calculation

95
S.-J. Kim, W. Koo and M.-H. Kim International Journal of Naval Architecture and Ocean Engineering 13 (2021) 86e101

Fig. 10. Comparison of the heave RAOs and the time-averaged extraction powers for three different buoy models.

condition. When the wave amplitude was small (A ¼ 0.25 m), the better power take-off performance than the truncated conical buoy.
effect of latching control was amplified more than twice the Further analysis of a hemispherical buoy under irregular wave
maximum extraction power compared to the case without latching conditions was performed. In the case of irregular wave analysis,
control. When the wave amplitude was large (A ¼ 1.0 m), however, the same drag coefficient as the previous regular wave analysis was
it was amplified by approximately 40%. When applying latching applied. The time interval was fixed to 0.01 s, and the analysis
control, the cylindrical buoy and the spherical buoy have a slightly duration was fixed to the sea state duration of 3 h. Fig. 11 shows the

96
S.-J. Kim, W. Koo and M.-H. Kim International Journal of Naval Architecture and Ocean Engineering 13 (2021) 86e101

Table 2
Maximum time-averaged extraction power.

Model1 [W/m3] Model2 [W/m3] Model3 [W/m3]

WN(A ¼ 0.25 m) 8469 (2626 W) 8236 (2573 W) 7945 (2483 W)

(No latching control)
Lin(A ¼ 0.25 m) 19,866 (6208 W) 19,192 (5997 W) 17,823 (5570 W)
WN(A ¼ 0.25 m) 19,398 (6062 W) 19,296 (6030 W) 16,473 (5148 W)
WN(A ¼ 0.50 m) 15,749 (19,686 W) 15,437 (19,296 W) 13,045 (16,306 W)
WN(A ¼ 1.00 m) 11,900 (59,500 W) 11,487 (57,435 W) 10,123 (50,615 W)

Fig. 11. Comparison of the significant buoy displacement on various significant wave heights.

significant displacements of each numerical model for various incident wave amplitude. In the comparison in Fig. 11, the signifi-
significant wave heights in the absence of a PTO force and the cant displacements of the buoys in all numerical models are slightly
latching control strategy. The zero-crossing method was applied to larger than those of the linear analysis in the WN analysis because
calculate the significant buoy displacement. In the case of irregular the amplitudes of the Froude-Krylov forces by larger displacements
wave analysis, since a random phase angle should be applied to are increased as shown in Fig. 6. However, the differences are small.
each frequency component, the numerical analysis of a single wave Fig. 12 shows the time histories of the heave displacement,
condition (Hs and Tp ) was repeated 10 times, and the average value instantaneous extraction power, static force, and excitation force of
was used. When the latching control strategy and PTO force were a hemispherical buoy type WEC under various significant wave
not applied, the change in the significant buoy displacement was heights with optimal latching control. The PTO force for the
not large. The overall amplitude tended to decrease with increasing maximum time-averaged power was estimated by trial and error
wavelength (frequency decreases) and had a value similar to the (FPTO =Fz ¼ 0.3). The latching duration, the only variable in the
significant wave height. This is because as the wavelength in- latching control strategy, was set as Eq. (31). For this analysis, the
creases, the heave amplitude of the buoy becomes equal to the same random phase angle was applied for each incident wave

97
S.-J. Kim, W. Koo and M.-H. Kim International Journal of Naval Architecture and Ocean Engineering 13 (2021) 86e101

Fig. 12. Time histories of the buoy displacement, instantaneous extraction power, static force, and excitation force on various significant wave heights ( FPTO = Fz ¼ 0.3, uP = u0 ¼ 0.45,
TL ¼ 1.43 s, Model 1).

frequency component. When the amplitude of the buoy displace- 0.3e0.5 for all numerical models unlike the optimal ratio in regular
ment was small (the time range of 30 ≪Time=Tp ≪31.5), all the waves (FPTO =Fz ¼ 0.5e0.7). Based on the results of Figs. 13 and 14
calculated values similar regardless of the amplitude of the incident shows the significant displacement and the time-averaged extrac-
wave. On the other hand, the results of nonlinear analysis varied as tion power under various peak frequencies and significant wave
the displacement of the buoy increased. When the significant wave heights in irregular waves using a latching control strategy. The
height was high, the displacement of the buoy changed rapidly by zero-crossing method was used to calculate the significant
latching control, and the nonlinear static force and excitation force displacement and time-averaged extraction power. The optimal
also changed nonlinearly. PTO conditions and the latching control for the maximum energy
Fig. 13 shows the time-averaged generated power for each nu- extraction were applied. This numerical analysis was performed in
merical model as function of various PTO forces and wave peak the longwave region, where the latching control technique was
frequencies of irregular waves. For this, weakly nonlinear analysis effective. The buoy displacement increased with increasing wave-
was carried out in irregular waves (Hs ¼ 0.5 m) to obtain the length, but the displacement decreased with increasing significant
optimal PTO force ratio (the ratio of PTO force to excitation force at wave height. When the ratio of the peak frequency to body reso-
each peak frequency, FPTO =Fz ). The optimal ratio (FPTO = Fz ) was nant frequency (up =un ) was increased from 0.4 to 0.45, the

98
S.-J. Kim, W. Koo and M.-H. Kim International Journal of Naval Architecture and Ocean Engineering 13 (2021) 86e101

Fig. 13. Comparison of the time-averaged generated power on various PTO forces and wave peak frequencies (Weakly nonlinear analysis, Hs ¼ 0.5 m).

maximum time-averaged extraction power occurred. Therefore, forms (hemispherical, cylindrical, and truncated conical) for a
when designing the heaving hemispherical buoy type WEC to heaving buoy type WEC were derived theoretically using the
which the latching control was applied, it was effective to design Maclaurin expansion method. The theoretical solutions of the
the incident wave frequency at approximately 0.4e0.45 times the nonlinear Froude-Krylov forces were applied to the WEC system
heave natural frequency of buoy. In addition, when the significant using a latching control strategy. The coupled analysis between the
wave height was small (Hs ¼ 0.5 m), the extraction power increased buoy and the PTO system was carried out using a latching control
by 25e50% due to latching control. In contrast, when the significant strategy. To confirm the accuracy of the theoretical solutions, the
wave height was large (Hs ¼ 2.0 m), the latching control effect was heave RAOs of the buoy were compared with the results of linear
increased by only 10e15%. As a result of estimating the influence of analysis and the remeshing numerical scheme. The drag co-
the buoy shape through Fig. 14(b)~(d), the trend of the time- efficients were also applied to each buoy to consider the viscous
averaged generated power was similar regardless of the shape of effect. Based on this, the hydrodynamic performance and extrac-
the buoy. However, the maximum generated power from the tion power under various wave conditions were evaluated for three
truncated conical buoy type WEC was smaller than the others. The types of heaving buoy WECs to which the latching control strategy
cylindrical buoy WEC with high wave amplitude (Hs ¼ 2.0 m) and had been applied, and the following conclusions were drawn.
latching control strategy produced less power than the case
without latching control. Through this, nonlinear analysis is 1. An incident wave period of 0.4e0.45 times the natural period of
required to estimate the power take-off performance of the wave the buoy was the optimal condition to amplify the extraction
energy converter accurately under high wave conditions. power of the WEC buoy through latching control. Under irreg-
ular wave conditions, the peak period of the incident wave was
optimal between 0.4 and 0.45 of the natural period of the buoy.
4. Conclusion

In this study, the nonlinear Froude-Krylov forces for three buoy

99
S.-J. Kim, W. Koo and M.-H. Kim International Journal of Naval Architecture and Ocean Engineering 13 (2021) 86e101

Fig. 14. Comparison of significant displacement of buoy and time-averaged extraction power on various irregular wave conditions (a, b: hemispherical buoy (Model 1), c: cylindrical
buoy (Model 2), d: truncated conical buoy (Model 3)).

2. Among the three types of WEC heaving buoys under latching Korea Institute for Advancement of Technology (KIAT). This
control, the hemispherical buoy showed the maximum energy research was also supported by Basic Science Research Program
extraction performance. through the National Research Foundation of Korea (NRF-
3. When the significant wave height was small (Hs ¼ 0.5 m), the 2018R1D1A1B07040677 and 2018R1A6A3A01013558).
extraction power increased by 25e50% due to latching control,
but the latching control effect was increased by only 10e15%
References
when the significant wave height was large (Hs ¼ 2.0 m).
Through this, nonlinear analysis is required to estimate the Babarit, A., Mouslim, H., Clement, A., Laporte-Weywada, P., 2009. On the numerical
power take-off performance of the wave energy converter modelling of the non linear behavior of a wave energy converter. In: Pro-
accurately under high wave conditions. ceedings of the ASME 2009 28th International Conference on Ocean. Offshore
and Arctic Engineering, Honolulu, Hawaii, USA.
Babarit, A., Hals, J., Muliawan, M.J., Kurniawan, A., Moan, T., Krokstad, J., 2012.
Numerical benchmarking study of a selection of wave energy converters.
Declaration of competing interest
Renew. Energy 41, 44e63.
Cho, I.H., 2015. Latching control technology for improvement of extracted power
The authors declare that they have no known competing from wave energy converter. Journal of the Korean Society for marine Envi-
financial interests or personal relationships that could have ronment and Energy 18 (4), 282e290.
Falcao, A.F.O., 2008. Phase Control through Load control of oscillating-body wave
appeared to influence the work reported in this paper. energy converters with hydraulic PTO system. Ocean. Eng. 35 (3e4), 358e366.
Folley, M., Henryand, A., Whittaker, T., 2015. Contrasting the hydrodynamics of
heaving and surging wave energy converter. In: Proceedings of the 11th Eu-
Acknowledgments ropean Wave and Tidal Energy Conference, Nantes, France..
Giorgi, G., Penalbe, M., Ringwood, J., 2016. Nonlinear hydrodynamic models for
This research was supported by the MOTIE (Ministry of Trade, heaving buoy wave energy converters. In: Proceedings of Asian Wave and Tidal
Energy Conference. AWTEC), Singpore, pp. 144e153.
Industry, and Energy) in Korea, under the Fostering Global Talents Giorgi, G., Ringwood, J.V., 2017. Computationally efficient nonlinear Froude-Krylov
for Innovative Growth Program (P0008750) supervised by the force calculations for heaving axisymmetric wave energy point absorbers.

100
S.-J. Kim, W. Koo and M.-H. Kim International Journal of Naval Architecture and Ocean Engineering 13 (2021) 86e101

Journal of Ocean Engineering and Marine Energy 3 (1), 21e33. International Conference on Ocean. Offshore and Arctic Engineering, Glasgow,
Giorgi, G., Ringwood, J.V., 2018. Analytical representation of nonlinear Froude- Scotland, UK.
Krylov forces for 3-DoF point absorber wave energy devices. Ocean. Eng. 164, Kim, S.J., Koo, W., Shin, M.J., 2019c. Numerical and experimental study on hemi-
749e759. spheric point-absorber-type wave energy converter with a hydraulic power
Gudmestad, O.T., Meo, G., 1996. Hydrodynamic coefficients for calculation of hy- take-off system. Renew. Energy 135, 1260e1269.
drodynamic loads on offshore truss structures. Mar. Struct. 9, 745e758. Lok, K.S., Stallard, T.J., Stansby, P.K., Jenkins, N., 2014. Optimisation of a clutch-
Guo, B., Patton, R.J., Jin, S., Gilbert, J., Parsons, D., 2018. Nonlinear modelling and rectified power take off system for a heaving wave energy device in irregular
verification of a heaving point Absorber for wave energy conversion. IEEE waves with experimental comparison. International Journal of Marine Energy 8,
Transactions of Sustainable Energy 9 (1), 453e461. 1e16.
Hal, J., Falnes, J., Moan, T., 2011. A comparison of selected strategies for adaptive Merigaud, A., Gilloteaux, J.C., Ringwood, J.V., 2012. A nonlinear extension for linear
control of wave energy converters. J. Offshore Mech. Arctic Eng. 133 (3), 031101- boundary element methods in wave energy device modelling. In: Proceedings
1-12. of the ASME 2012 31st International Conference on Ocean. Offshore and Arctic
Jang, H.K., Kim, M.H., 2020. Effects of nonlinear FK and hydrostatic restoring forces Engineering, Rio de Janeiro, Brazil.
on arctic spar motions in waves. International Journal of Naval Architecture and Nazari, M., Ghassemi, H., Ghiasi, M., Sayehbani, M., 2013. Design of the point
Ocean Engineering 12, 297e313. absorber wave energy converter for Assaluyeh port. Iran. J. Energy Environ. 4
Kalofotias, F., 2016. Study for the Hull Shape of a Wave Energy Converter-point (2), 130e135.
Absorber; Design Optimization and Modeling Improvement. Master’s thesis, Sarpkaya, T., 1986. Force on a circular cylinder in viscous oscillatory flow at low
University of Twente. Keulegan-Carpenter numbers. J. Fluid Mech. 165, 61e71.
Kim, J., Cho, I.H., Kim, M.H., 2019a. On numerical calculation and experiment of a Sheng, W., Alcorn, R., Lewis, A., 2015. On improving wave energy conversion, part II:
heaving-buoy wave energy converter with a latching control, Ocean Systems development of latching control technologies. Renew. Energy 75, 935e944.
Engineering. Int. J. 9 (1), 1e19. WAMIT User Manual V. 7.3, 2019. WAMIT, Inc.
Kim, S.J., Koo, W., Jo, C.H., 2019b. Assessment of latching control for the hemispheric Zurkinden, A.S., Ferri, F., Beatty, S., Kofoed, J.P., Kramer, M.M., 2014. Nonlinear nu-
heaving buoy type point absorber with and without nonlinear Froude-Krylov merical modeling and experimental testing of a point absorber wave enregy
force acting on the buoy. In: Proceedings of the ASME 2019 38th converter. Ocean. Eng. 78, 11e21.

101