Ocean Engineering 116 (2016) 226–235

Contents lists available at ScienceDirect

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

Effect of inclination on oscillation characteristics of an oscillating water

column wave energy converter
Mitsumasa Iino n, Takeaki Miyazaki, Hiroshi Segawa, Makoto Iida
The University of Tokyo, 4-6-1, Komaba, Meguro-ku, Tokyo 153-8904, Japan

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

Article history:
Received 1 November 2015
Oscillating water column (OWC) wave energy converters are adopted in many marine applications such
Accepted 6 March 2016 as buoys. This research investigates how inclination-induced changes in motion affect the oscillation
characteristics of an OWC. As the collector of the OWC wave energy converter is inclined, the direction
of motion of the water column is also inclined because it is constrained by collector walls. To
Keywords:
investigate this idea, we developed a mechanical 1-degree-of-freedom oscillator model of an inclined
Oscillating water column
Wave energy
OWC. The usefulness of the mechanical model was confirmed in a water tank test with OWCs inclined at
Mechanical model 90° (vertical), 45°, and 18.4°. We then investigated the effect of inclination on the resonance period and
Water tank test the optimum power take-off damping of the OWC. The resonance period was mainly prolonged by the
reduced restoring force as the motion direction changed. The optimum power take-off damping of an
inclined OWC equaled the radiation damping, as observed in vertical OWCs. Finally, we varied the air
chamber volume and observed much smaller effects than when varying the inclination. We conclude that
changing the direction of motion affects the oscillation characteristics of OWCs mostly because of the
reduction of gravity restoring force.
& 2016 Published by Elsevier Ltd.

1. Introduction
Previously, the only inclined OWC wave energy converter
subjected to full-scale testing was LIMPET in the United Kingdom.
The oscillating water column (OWC) wave energy converter is
According to the LIMPET report, the inclination of the OWC
among the most reliable wave energy conversion systems. Many
increases the water surface area and hence the resonance period
full-scale prototype tests have been performed on floating offshore
of the OWC. Moreover, the inlet loss is decreased because of the
types (Ishii et al., 1982; Yukihisa et al., 2001) and shoreline fixed
fluid dynamics. Elsewhere, Ram et al. (2010) experimentally
types (Hotta et al., 1986; Pecher et al., 2011). At Port Sakata (Suzuki
compared the efficiencies of OWCs inclined at angles under lim-
et al., 2004), a shoreline fixed OWC wave energy converter has
ited wave conditions. The DTI project in the UK demonstrated the
been continuously tested for 13 years without major accident. In
superior performance of a floating buoy with an inclined OWC and
these experiments, a water column oscillating in the vertical
a backward bent duct buoy OWC (Department of Trade and
motion was constructed as a concrete or steel chamber.
Industry Great Britain, 2004).
Recently, we conducted the Blow-Hole wave energy converter
However, previous studies appear to lack fundamental physical
project in Japan (Miyazaki et al., 2014). In this project, we con-
insight. To elaborate, inclining the OWC changes the direction of
structed a cylindrical OWC collector inclined at 18.4° from the
motion of the water surface. If the direction of motion is inclined
horizontal level at the Fukui test site. Unlike standard artificial
from the vertical axis, the gravitational acceleration slows down,
collectors, our OWC collector is an inclined tunnel drilled into
analogous to a box sliding or water falling down a slope. Moreover,
shore rock. As in traditional OWC devices, power take-off equip-
if the water is allowed to freely oscillate in the water column, the
ment with a turbine generator is attached at the top of the tunnel.
slower acceleration increases the period of the oscillation, similar
The other side of the tunnel is a wave inlet. In this type of device,
to a pendulum swinging on a slope. Thus, the inclination of the
the inclination angle of the OWC is limited by the tunneling cost.
OWC might significantly affect the oscillation characteristics of the
Therefore, inclination effects should be properly considered at the
water column. This type of water motion (with an inclined
preliminary design stage.
direction) is pointed to in the literature of oscillating body type
wave energy converters (Payne et al., 2008). Therefore, this phe-
n
Corresponding author.
nomenon must significantly affect the converter. However, whe-
E-mail address: iino@ilab.eco.rcast.u-tokyo.ac.jp (M. Iino).
ther inclination changes the oscillation or motion characteristics

http://dx.doi.org/10.1016/j.oceaneng.2016.03.014
0029-8018/& 2016 Published by Elsevier Ltd.
was not considered in previous literature on OWC type wave
energy converter.
The collector utilizes the oscillation phenomena of the water
column when transferring the power of the wave to the air
chamber. Therefore, the power transfer is most effective around
the resonance period of the water column. This characteristic has
been confirmed in real sea tests (Osawa and Miyazaki, 2004),
water tank tests (Sheng and Flannery, 2012), and fluid dynamics
simulations (Suzuki, 2005). Alternatively, the oscillations in the
water column can be simply and effectively explored by
mechanical modeling.
Unlike computational fluid dynamics, mechanical modeling
treats the water column as a rigid body, and its motion as a 1-
degree-of-freedom (DOF) oscillation. An example (Brendmo, 1997)
of a mechanical model that ignores the air compressibility of the
air chamber is presented in Fig. 2. Clearly, this model determines
the resonance period or optimum power take-off damping with-
out a time series calculation. Although parameters must be
properly determined to obtain precise characteristics, the reso-
nance period is obtainable from the mass of the water column m,
which is easily computed from the water column geometry
(Osawa and Miyazaki, 2005; Webb et al., 2005). Moreover, if we
can determine the radiation damping C1, we can analytically
derive the power take-off damping C2 that maximizes the effi-
ciency. Combined with a nonlinear model of the air chamber, this
model can be iterated through a time series at a less computa-
tional cost than a fluid dynamics simulation. Thus, mechanical
models are useful for obtaining characteristic values such as the
resonance period and optimum damping parameters. Recently,
such models have been used to investigate air chamber com-
pressibility and turbine characteristics (Folley and Whittaker,
2005) to optimize the damping of a floating OWC (Stappenbelt and
Cooper, 2010) and to establish coupled simulation codes of
floating OWC wave energy converters (Bayoumi et al., 2014).
The present research investigates how changes in the motion
direction affect the water column in an inclined OWC. We first
deduce qualitative characteristics by physical analogy to a fluid or
rigid body on a slope. We then propose a mechanical model of the
inclined OWC that predicts the quantitative characteristics of the
system. The appropriateness of the model is validated in water
tank tests. The time series, sensitivity of the power take-off
damping, and OWC efficiency are determined for different incli-
nation angles. Finally, using the inclined OWC mechanical model,
we discuss the fundamental characteristics of the OWC, i.e., the
resonance period, optimum power take-off damping, and air
compressibility effects.
2. Materials and methods
This research studies the change in the motion or oscillation
characteristics of an OWC caused by inclination. Previous studies
Fig. 1. Vertical OWC (left) and inclined OWC (right).
have focused on the fluid properties, such as inlet resistance, of
inclined OWCs. Here, we clarify the effect of inclination on the
oscillation characteristics of the OWC using a mechanical model.
Section 2.1 presents the details of the model, including the air
chamber model discussed in previous literature. Section 2.2
explains the water tank tests for verifying the model and their
parameters. Section 2.3 describes the efficiency calculation of
the OWC.
2.1. Model of the inclined OWC system
2.1.1. Equation of motion of an inclined water column
In the mechanical model of the inclined OWC (Fig. 1, right), the
water column is assumed as a rigid body. The equation of motion
of the inclined water column is then given by
mðyÞy€ þ C 1 ðωÞy_ þ ρw Aygsinθ ¼ F 1 — F 2 ;
mðyÞ ¼ Ayρw þm0 þ ma ðωÞ; ð1Þ
where y is the displacement of the water surface along the
direction of motion, mðyÞ is the mass of the water column, m0
is the geometrically determined mass of a still water column (i.e.,
a
volume surrounded by inlet, chamber wall, and water surface),
and ma is an additional mass that moves with the water column;
C1 is the radiation damping coefficient, ρw is the water density, A is
the cross-sectional area of the water column along the direction of
motion, and gis the gravitational acceleration; F1 is the wave
exciting force and F2 is the force from the air chamber pressure;
and θ is the inclination angle of the water column from the hor-
izontal level (θ ¼ 901 denotes a vertical water column).
According to Eq. (1), an inclined water column moves along the
inclination direction at a specified slope angle and is constrained
by the surrounding chamber walls. If the water level increases
from the still water level (y40), the water column mass increases
by Ayρw . The vertical gravitational force on the increased mass is
Ayρwg. Because the motion is inclined at angle θ, the gravitational
force is reduced by sin θ along the y direction. If the water level
decreases from the still water level (y o0), the mechanics are
reversed because the gravitational force acts to restore the still
level. As the water level changes, the incoming or outgoing water
induces a wave, giving rise to the radiation damping force C 1 ðωÞy_
, which resists the water column motion.
If θ ¼ 901, the mechanical model reduces to that of the vertical
OWC as follows (Kinoshita et al., 1984):
mðyÞy€ þ C 1 ðωÞy_ þ ρw Ayg ¼ F 1 — F 2 : ð2Þ
Clearly, we cannot retrieve Eq. (1) by tuning any of C1, ma, or
m0, i.e., by adjusting the fluid loss or OWC geometry (such as the
water surface area and the inlet shape). Therefore, the fluid loss
and geometry (which have mainly featured in previous literature)
do not completely explain different oscillational phenomena in the
inclined and traditional vertical OWCs.
2.1.2. Thermodynamics of the air chamber
Importantly, Eq. (1) states that inclining the OWC reduces the
gravity restoring force. However, to properly characterize the
OWC, we require an appropriate model of the air chamber, which
is related to the calculation of F2 in Eq. (1). The air chamber in an
OWC system has been previously modeled as an isentropic ther-
modynamic system (Falcã o and Justino, 1999). In the thermo-
dynamic model, the air in the chamber is assumed to exist in a
uniform thermodynamic state; that is, it undergoes no thermal
exchange through the chamber walls. Given the change in air
density between the inhaling and exhaling processes, the ther-
modynamic system of the air chamber is expressed as ( Sheng
et al., 2013)
8
dV dp
>dt< γ pþdt Vc þ Q pto ¼ 0 ðinhaleÞ
>:
1 : ð3Þ
dV þV dp þp0 γQ
dt γ p dt c pto ¼ 0 ðexhale Þ
pc

Here V is the chamber air volume, γ is the specific heat ratio of

air, p is the air chamber gauge pressure, p0 and pc are the absolute
pressures of the atmosphere and chamber air, respectively, and
Qpto is the air exchange through the power take-off system or
Fig. 2. Mechanical model for oscillating water column wave energy converter
control valve (if it exists). Because air is much less dense than
without air compressibility and with nonlinear turbine P–Q characteristics. F is the
water, the gravitational force acting on air is negligible, as evident wave exciting force, k is the gravity restoring force, m is the mass of the water
from Eq. (3). Therefore, the changing direction of the chamber air column, and C1 and C2 are the radiation and power take off damping coefficients,
flow does not affect the behavior of the air chamber. respectively.
Eqs. (1)–(3) are related as follows:
V ¼ V 0 — Ay; ð4Þ

dV
¼ — Ay_; ð5Þ
dt

F2 ¼ Ap: ð6Þ
Here; V 0 is the initial volume of the air chamber at y ¼ 0. Cou-
pling Eqs. (1) and (3)–(6), we obtain the nonlinear ordinary dif-
ferential equation for the inclined OWC as follows:
mðyÞ ¼ Ayρw þ m0 þ ma ðωÞ; ð7Þ
0 1
0 1 0 10 1 0
y_€ 0 1 0 y C
y 0 y_ F1
θ
A B B
— —
C C1
m @ C C ;
BA ¼ A A
þ B@
dt
@
m —

pair gsin
ð8Þ
γ ðp0 þ pair Þ dV Fig. 3. Overview of the OWC water tank test model.
p_ 0 0 0 V — —
αQ pto
8
air
< ð9Þ
α ¼ : 1ðp air 40Þ
1
γ ðpair o0Þ ;
p 0 p0 þ
pair

where m0 and A are determined from the collector shape. The

parameters F1, C1, ma; and Qpto depend on the collector shape,
water depth, power take-off characteristics, shoreline shape, and
other factors. We configure the values of these parameters in
Section 2.2.

2.2. Water tank testing from the horizontal level by 90°, 45°, and 18.4° by pivoting around the
top of the inlet, as shown in Fig. 4. The top of the inlet was
To confirm the motion of the inclined OWC, we conducted
water tank tests at the National Maritime Research Institute
(Tokyo, Japan). The tank was 50 m long, 8 m wide, and 4.5 m deep.
This section describes the test conditions and parameters of the
numerical model.

2.2.1. OWC model geometry

The test model was the inclinable cylindrical OWC presented in
Fig. 3. The cylinder diameter was 0.3 m. The model was inclined
 Fig. 4. Schematic of the inclinable OWC model.

0.2 m deep. The length L0 was 3.5 m; at 18.4° inclination, a

longer cylinder (L0¼ 11.5 m) was also tested to investigate the
effect of the air chamber volume. The input wave at the side of
the OWC inlet and the interior water level of the center line of
each cylinder were measured by a capacitive wave height
meter.

2.2.2. Power take-off model

To represent the pressure drop of the power take-off
system, an orifice plate was attached at the top of each test
cylinder (Fig. 5). A 0.047-m-diameter hole was made in the
center of the orifice plate.
 Fig. 6. Geometry of cylindrical OWC model for the water tank test.

Table 2
Added mass and radiation damping in the numerical model of the inclined OWC.

Inclination angle (deg) ma C1

Fig. 5. Outlet nozzle (orifice) and pressure tap. 90 0 47

45 0 15
18.4 0 37
Table 1
Tested wave conditions.

Inclination angle θ [deg] Input wave period T0 [s] 5,000

45, 90 1, 2, 3 4,000
18.4 2, 2.5, 3, 3.5
1/140

(Pa)
1/120
3,000
1/100
The area of this hole was 1/40 times the cross-sectional area of the 2,000 1/80
cylinder. A pressure sensor was installed on the orifice plate. To 1/60
1,000
test the air damping of the power take-off system, we varied the 1/40
orifice diameter, creating holes with 1/60, 1/80, 1/100, 1/120, 0
0 0.05 0.1 0.15
and 1/140 times the area of the cylinder cross-section. In this
manner, we tested 6 nozzle area ratios: 1/40, 1/60, 1/80, 1/100, Qpto(m3/s)
1/120, and 1/140. Fig. 7. P–Q characteristics of nozzles with different nozzle area ratios.

2.2.3. Wave conditions depth h. When calculating the water pressure, we applied the
Regular waves with a height of 0.1 m were tested. The wave small-amplitude wave theory.
conditions are shown in Table 1. In this study, ma and C1 are assumed as constants. Their values
are listed in Table 2. Note that these parameters are frequency
2.2.4. Numerical model parameters for the water tank test dependent and can be included in a time series calculation using
As described in Section 2.1, the parameters m0, A, F1, C1, ma, the impulse response function. However, because of time
and Qpto in Eqs. (7) and (8) need to be determined. This section restraints, we could not execute the required identification
describes each of these parameters under the testing conditions. experiments for the impulse response function. These values were
m0 and A are determined from the model geometry as descri- tuned for the case which showed maximum efficiency of the OWC
bed below. in the water tank test at each inclination angle.
m0 ¼ hin πD2 Finally Qpto is calculated from the pressure and theoretical P–Q
sin θ ρw ; characteristics of the orifice, as explained below:
4
D
hin ¼ htop þ cos θ; ð10Þ Q sffi ffiffi ffiffi ffiffi ffiffi ffi ffiffi ffiffi
2 pto ¼ A o pto 2 ρj p :p t o j Þ
C A sign p a 12
where hin is the depth of the center of the OWC inlet, D is the
cylinder diameter, θ is the inclination angle, and htop ¼ 0:2ðmÞ is In Eq. (12), Ao is the area of the orifice hole,ρa is the air density
the depth of the top of the inlet, which is also the pivot point of through the orifice, and CA is the flow rate factor. Based on a
inclination. A geometric view of these parameters is presented in previous report, we set CA¼0.65 (Sheng and Flannery, 2012). The
Fig. 6. orifice P–Q characteristics for a single cylinder are plotted in Fig. 7.
The wave exciting force F1 is obtained by integrating the water Clearly, the orifice damping characteristics are nonlinear with
pressure at the inlet. The integral is straightforward because the respect to the chamber air pressure. Similar characteristics were
inlet shape is simple and no obstacles exist around it. observed in an impulse turbine operated at a fixed speed and a
Wells turbine operated at maximum power (Ceballos et al., 2013).
F 1 ðtÞ ¼ In addition, we correct ppto for the pipe friction loss of the
pw ðx; hÞdA;
Z Ai
n
cylinder:

H0 }
¼∬ ρ cosh kðhd — hÞ tÞ— ρ gh wdhdw: ppto ¼ pair — ploss; ð13Þ
g
ω cos ðkx —
in w
2 w
coshðkhd Þ
ð11Þ L ρ c y_ 2
ploss ¼ signðy_ Þ; ð14Þ
Here pw is the water pressure at the longitudinal position x and λ
D 2
depth h, Ain is the cylinder inlet area, H0 is the wave height, hd is the depth of the water tank, k is the wave number, ωin is the angular
frequency of the input wave, and w is the inlet width at where λ is pipe friction loss factor, L is the length of the air
chamber, and ρc is the chamber air density. Note that ploss is less
than 0.1% of ppto in the present geometry and can be ignored in
this
study. However, we retain the loss term because it becomes sig-
0.1 500
nificant if the air chamber has a thin duct.
0.08 400

chamber air pressure p1(Pa)

0.06 300
2.3. Calculation of efficiency

water level h1(m)

0.04 200
0.02 100
In both numerical and experimental results, the efficiency was h1_exp
0 0
calculated as outlined below. h1_sim p1_exp p1_sim
-0.02 -100
The average efficiency η of the OWC for each wave and geo-
-0.04 -200
metrical setting is calculated by Eq. (15):
-0.06 -300
R
Pdt -0.08 -400
T
η ¼Tinin Pin ; 15 -0.1 -500
0 2 4
ð Þ 6
where Tin is the input wave period, P is the output power at power
Time(s)
take-off, and Pin is the average power of the input wave. Fig. 9. Time series of water level and chamber air pressure for θ ¼ 451; Tin ¼ 2:0;
H0 ¼ 0:1.

P ¼ jppto
sffi ffiffi ffiffi ffi
jCAA ffi ffiffi 2
ffi ffiffi ffi p ; ð16Þ 0.1 500
o j
p to j 0.08 400

chamber air pressure p1(Pa)

0.06 300
H2ρwρ
gωa in 2khd 0.04 200

water level h1(m)

0
Pin ¼ D 1þ : ð17Þ
16k sinh2khd 0.02 100

As shown in Eq. (16), the output power at power take-off is the 0 0

product of the pressure difference and volume flow rate. Eq. (17) -0.02 -100
is derived from the small-amplitude wave theory. -0.04 h1_sim -200
-0.06 h1_exp -300

-0.08 p1_exp -400

3. Results
This section presents the results of time series simulations
using Eqs. (7)–(9). The appropriateness of the numerical model
and the characteristics of the inclined OWC are revealed by the
experiment. First, the appropriateness of modeling the inclined
OWC as a 1-DOF oscillator model is confirmed by comparing the
results of experimental and numerical time series. This test con-
firms that the only frequency in the system is the input wave
frequency. Next, to confirm a proper modeling of the take-off
damping, we compare the experimental and numerical sensitiv-
ities of pressure and water surface amplitude to the nozzle area
ratio. This test is performed because the power take-off damping
significantly affects OWC characteristics; therefore, an excessively
small or large power take-off damping coefficient might not attain
the general characteristics of the OWC. After characterizing the
oscillation phenomenon, we investigate the effect of inclination on
the frequency characteristics of the OWC performance by relating
the efficiency to the input wave frequency.
0.1
0.08
-0.1 p1_sim -500
0 5 10
Time(s)
Fig. 10. Time series of water level and chamber air pressure for θ ¼ 18:41; Tin ¼ 3:5;
H0 ¼ 0:1.
3.1. Time series of water level
Figs. 8–10 confirm that the 1-DOF oscillator model appro-
priately reproduces the motion of the water level even when the
water column is inclined. In these figures, level h1 denotes the
vertical displacement of the water column, and the chamber air
pressure p1 is the measured pressure at the side of the outlet
nozzle in the water tank test or ppto in the numerical simulation. In
the numerical simulation, the water level h1 is calculated as fol-
lows (because the cylinder is inclined from vertical axis).
h1 ¼ y sin θ: ð18Þ
In the later figures of this paper, the numerical and experi-
mental results will be distinguished by sim and exp, respectively.
In all presented results, the input wave and water level have the
same frequency; no other significant frequency component was
500 observed. Because the oscillations can be considered as forced
400 oscillations, the excitation force and motion of the water surface
ber air pressure p1(Pa)

0.06 must also have the same frequency if the proposed mechanical
300
model is appropriate. Therefore, the results show that the 1-DOF
water level h1 (m)

0.04 200
0.02 100 mechanical oscillator model properly represents the inclined

0 water column. Moreover, the simulation and experimental results

-0.02 0 appear to favorably agree at all inclination angles.
-0.04 -100

-0.06 -200
-300 3.2. OWC water level and pressure amplitude
h1_exp
h1_sim
ham
c

-0.1
-0.08 p1_exp -500
-400 Besides acting as an oscillator, the OWC wave energy converter
0 1 2 3 damping. Therefore, the power take-off damping must be prop-
Time(s) erly modeled to discuss the characteristics.
Fig. 8. Time series of water level and chamber air pressure for θ ¼ 901; Tin ¼ 1:0; This section investigates how the amplitudes of the basic fre-
H0 ¼ 0:1. quency components of the water level and air pressure respond to
 1.6
1.6
_sim _sim
1.4
_exp 1.4 _exp
1.2 _sim _exp
1.2
_exp _exp
1 1
0.8 0.8

,
,

0.6 0.6
0.4 0.4
0.2 0.2
0 0
20 40 60 80 100 120 140
20 40 60 80 100 120 140
Inverse of nozzle area ratio
Inverse of nozzle area ratio
Fig. 13. Nondimensional amplitudes of water level and chamber air pressure for
Fig. 11. Nondimensional amplitudes of water level and chamber air pressure for
θ ¼ 18:41; Tin ¼ 3:5; H0 ¼ 0:1.
θ ¼ 901; Tin ¼ 1:0; H0 ¼ 0:1.

1.6
_sim
1.4 _exp
1.6 _exp
_sim 1.2
1.4 _exp _exp
_sim 1
1.2 _exp 0.8

,
1
0.6
0.8
,

0.4
0.6 0.2
0.4 0
20 40 60 80 100 120 140
0.2
Inverse of nozzle area ratio
0
20 40 60 80 100 120 140 Fig. 14. Nondimensional amplitudes of water level and chamber air pressure for
Inverse of nozzle area ratio θ ¼ 18:41; Tin ¼ 2:0; H0 ¼ 0:1.

Fig. 12. Nondimensional amplitudes of water level and chamber air pressure for
0.5
θ ¼ 451; Tin ¼ 2:0; H0 ¼ 0:1. θ=90° exp θ=90° sim θ=45° exp
0.45

the nozzle area ratio. The amplitude of the basic frequency com- 0.4
θ=45°sim
ponent is calculated by the Fast Fourier Transform in MATLAB 0.35 θ=18.4°exp
OWC efficiency η

θ=18.4°sim
2010b and is expressed in a nondimensional form as follows: 0.3
0.25
ζ1 0.2
ζND1 H0=2 ¼ ;
0.15
0.1
ð19Þ

ζpND1 ζp1
¼ ;
ρw gH0=2
where ζ1and ζp1 denote the half amplitudes of the water surface
and air chamber pressure, respectively, in the OWC.
The half amplitudes ζ1 and ζp1 are plotted as functions of the
nozzle area ratio in Figs. 11–13. The amplitudes were evaluated
under the most efficient wave condition at each inclination angle.
The numerical simulation well agrees with the experimental
results, indicating that the model reproduces the sensitivity of the
nozzle area ratio to the OWC amplitude. Consequently, the model
can be used for optimizing the power take-off damping.
Fig. 14 plots the nondimensional amplitude versus inverse
nozzle area ratio at 18.4°. The period of the input wave is 2 s. In
this case, the experimental amplitude was quite low, indicating
that at frequencies considerably higher than the resonance fre-
quency, the radiation damping and added mass are drastically
changed. To reproduce such conditions in the numerical model,
the added mass and radiation damping should be frequency
dependent. However, we are primarily interested to study the
effect of inclination on the resonance period of the OWC; thus, we
do not elaborate on this discrepancy.
0.05
ð20Þ
0
0 1 2 3 4
Input wave period
Fig. 15. OWC efficiency of input wave period at different inclination angles.
3.3. Effect of wave period on OWC efficiency
Sections 3.1 and 3.2 confirmed that the inclined OWC can be
modeled as a 1-DOF mechanical oscillator. Therefore, its perfor-
mance should be evaluated as a function of input energy
frequency.
Fig. 15 plots the energy conversion efficiency η of the OWC as a
function of the input wave period. Except at 45°, the resonance
period appears to lie outside the experimented range of the wave
period. However, for two reasons, the resonance period is
expected to lengthen as the inclination angle reduces. First,
inclination increases the mass; second, it reduces the effect of
gravity (see Eq. (1)), which is the main interest in this study. Note
that although the system is most efficient at 451 among the tested
angles, this incli- nation may still be suboptimal. The peak
efficiency probably depends on factors other than inclination, such
as inlet shape, surrounding shore shape, and water depth. These
conditions would
affect the wave exciting force and radiation damping terms in Eq.
(1). Therefore, Fig. 15 shows only how the efficiency depends on the
input wave period and should be interpreted only in this regard.
Comparing the results of experiments and numerical simula-
tions, we confirmed that the mechanical model reproduces the
motion and pressure characteristics of an inclined OWC and rea-
sonably captures the damping characteristics. The resonance per-
iod of the OWC was found to be a decreasing function of the
inclination angle.
4. Discussion
This section first investigates the characteristics of the OWC
wave energy converter that are affected by motion direction
changes in an inclined OWC. Next, it discusses the characteristics
that are identical in the vertical and inclined OWCs. The latter
constitutes the optimum damping, which is independent of incli-
nation, and the air chamber volume because the air compressi-
bility has little effect in the water tank test and is not considered in
the simplified discussion. To summarize, we discuss
● How the resonance period of the inclined OWC is affected by
the reduced restoring force
●
The optimum damping characteristics of the inclined OWC
● period Tri is given by
Effect of the chamber air volume on the efficiency of the
inclined OWC
4.1. Effect of reduced restoring force on the resonance period of the
inclined OWC
4.1.1. Similarity to a pendulum on a slope
In Section 1, we compared the inclined OWC to a pendulum
swinging on a slope. We now apply this concept to explain the
effect of inclination on the gravity-related oscillation phenomena
of the OWC.
A pendulum oscillating with a small amplitude on a slope of
angle θ is governed by the equation of motion as follows:
mp l2pφ€p ¼ mp g φp sin θ: ð21Þ
As is evident from Eq. (21), the term of inclination angle ( θ)
could not be substituted for by enlarging or reducing the mass.
This fact demonstrates that the effect of inclination is independent
from the effect of mass.
If θ¼ 90°corresponds to a pendulum with zero slope, the
resonance period is given by
sffiffiffiffiffiffiffi ffiffiffiffiffiffiffi
g sinl θ
Tp 2π : 22
¼ ð Þ
From Eq. (22), we observe that the larger the inclination angle
θ, the longer the resonance period. This occurs because the gravity
restoring force decreases with increasing inclination of the moving
plane of the pendulum. If θ¼ 0°, the restoring force also becomes
zero, and the pendulum cannot oscillate.
4.1.2. Resonance period of inclined OWC
We now discuss how the changing motion direction alters the
undamped resonant period of the OWC, a characteristic parameter
used in preliminary designs of the shape of the water column.
Given the similarity between Eqs. (1) and (22), we infer that
the OWC inclination must affect the resonant period of the water
column. However, the water column will also lengthen its reso-
nant period if its mass increases under the same gravity restoring
force. This is because of the potential of the water column mass
and the restoring force of gravity somehow being tuned to a state
of independence, unlike Eq. (21). In previous works, the increased
resonant period has been attributed to the increased water surface
area and relative decrease of the wave inlet area (Heath et al.,
2002).
Therefore, the resonance period of the water column is gov-
erned by both mass and direction of motion of the water column.
To separate these two factors, we explore the undamped
resonant period Trv of a water column in the vertical motion. Eq.
(23) is based on Eq. (2), which excludes the reduction of the
gravity restoring force caused by inclination.
sffiffiffiffiffi ffiffi ffiffiffiffiffi ffiffiffiffi ffiffi
ρw0 Aþwg m a
T rv 2π ; 23
¼ ð Þ
where Aw is the surface area of the water column. The Aw is
usually determined in the preliminary design stage of an OWC
(Osawa and Miyazaki, 2005; Webb et al., 2005). From the geo-
metry of the tested model (see Fig. 6), the Aw in our inclined OWC
is given by
A
A¼w sin θ : ð24Þ
Eq. (23) gives the resonant period of the inclined OWC under
standard gravitational effects.
If the direction of motion changes, the undamped resonant
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Tri ¼ 2π
: ð Þ
m0 þ ma
25
ρ Agsinθ
Eq. (25)wgives the resonant period of the inclined OWC under
reduced restoring force. Note that the inclination angle θ appears
in both (22) and (25). However, Eq. (25) allows the resonance
period to lengthen with increasing inclination angle even for
constant mass of the water column.
Fig. 16 plots the resonant period calculated by Eqs. (23) and
(25) as functions of the inclination angle. The wave period that
maximized the OWC efficiency in the experiment is also plotted at
each inclination angle. Although both Tri and Trv are prolonged
under mass increase, Tri closely approximates the experimental
results, whereas Trv clearly deviates from these results. Due to the
low resolution of the experimental points, a significant level of
uncertainty in view of the resonance period is inherently at hand.
Therefore, the accuracy of Eq. (25) is not completely verified.
However, it could be said that regarding the expected period via
Eq. (23), the experimental efficiency itself is significantly mini-
mized, but not the associated resonance period. Therefore, the
resonance period expected by Eq. (25) describes the phenomenon
more properly. For the same mass of the water column in an OWC,
accounting for the motion direction of the water column can at
6
Undamped resonance
5
period of OWC [s]
4
Exp
3
2
1
0
10 30 50 70 90 110
Inclination angle θ [deg.]
Fig. 16. Predicted undamped resonant period (ignoring (Trv; green dashed line) and
including (Tri; red solid line) the gravity restoring force) and wave period that
maximized the OWC power in the experiments (blue dots). (For interpretation of
the references to color in this figure, the reader is referred to the web version of this
article.)
most double the resonant period when the column is inclined
0.3
from the vertical position. We emphasize that the tested range radiation
encompasses the inclinations of the LIMPET project (40°) and the damping
0.25
Blow-Hole project (18.4°); thus, the range is representative of real
inclination angles.

OWC efficiency η
0.2
Therefore, the preliminary design of an inclined OWC should
consider the reduced restoring force caused by the directional
change in the water column's motion. Otherwise, the OWC design 0.15
will include significant error in the resonance characteristics,
which invalidate the results of a water tank test or a CFD 0.1
performance test.
0.05
4.2. Optimum damping characteristics of the inclined OWC
0
0 50 100 150
Another important OWC characteristic, i.e., the optimum power linearized power take off damping C2
take-off damping, can be derived from the mechanical model of
Fig. 17. OWC efficiency versus linearized power take-off damping at θ ¼ 90°.
the OWC. The optimum conditions depend on many factors such
as the input wave conditions, turbine hysteresis, and chamber air
compressibility. 0.5
In defining the optimum power take-off damping, the usual 0.45
assumptions are an incompressible air chamber, a regular input 0.4
wave, an undamped natural frequency of the OWC, and no turbine

OWC efficiency η
0.35
hysteresis.
0.3
The model under the incompressible air condition is schema-
tized in Fig. 2. 0.25
When this model oscillates at an undamped natural frequency
pffiffiffiffiffiffiffiffiffi 0.2
ω0 ¼ k=m, the exciting force is given by F ¼ fsinðω0tÞ and the 0.15
amplitude of the consumed power is given by
0.1
radiation damping

W ¼ C 2 x_ C 2 ðf ω0 Þ2 0.05
sin 2ðω0t — βÞ; ð26Þ
2 ¼ 0
ω20
ðC 1 þ C 2 Þ=m 0 100 200 300
2

where β is the phase shift of the oscillator motion caused by the linearized power take off damping C2
exciting force. Note that the power is maximized when C2 ¼ C1;,
Fig. 18. OWC efficiency versus linearized take-off damping at θ ¼ 451.
regardless of k and m. This optimum C2 yields maximum power
not only precisely at ω0 but also in some range around ω0.
In the water tank test described in Sections 2 and 3, the air 0.5
chamber is sufficiently narrow (up to 3.3 m) such that we can 0.45
impose the incompressibility condition. If we modify only the 0.4
restoring forcek in the proposed mechanical model, the nozzle
OWC efficiency η

0.35
damping at the maximum power output should remain quite
0.3
similar.
To confirm this notion, we linearized the nozzle damping 0.25
(Suzuki, 2005). The nonlinear P–Q characteristics of the orifice 0.2
yield an equivalent linear damping factor C2 that maintains the 0.15 radiation
available output power, as shown below: damping
0.1
8 1 0.05
C2 ¼ o
2
C ρajζvj; ð27Þ 0
3π 0 linearized
50 power take
100 off damping
150 C2 200
2
Co ¼ A : ð28Þ
AoCwp
Fig. 19. OWC efficiency versus linearized take-off damping at θ ¼ 18:41.
Here; Co is orifice's loss factor, and ζv is the amplitude of the one nozzle area ratio.
water column velocity along the inclined direction of motion.
Figs. 17–19 plot the relations between C2 and efficiency. The
vertical dotted lines plot the coefficients C1 in Table 2. Each C2
corresponds to a different nozzle area ratio in the water tank test.
Clearly, C2 and C1 approximate each other under highly effi-
cient conditions, implying that the optimum power take-off
damping around the undamped natural frequency approximates
the radiation damping, regardless of inclination. The radiation
damping in the model does not exactly match the optimum nozzle
damping C2 in the experiment, possibly because the frequency of
the water tank test is not precisely set to ω0; moreover, modeling
and experimental uncertainties are inevitable. However, the
radiation damping and optimal damping are resolved to within
 Therefore, a traditional “matching” method for designing the
collector and power take-off equipment is also applicable to an
inclined OWC.

4.3. Effect of chamber air volume on OWC efficiency

In Sections 4.1 and 4.2, we ignored the effect of air compres-

sibility. Although the above considerations yield an adequate
OWC design that performs well in real sea tests, the
performance may be affected by the spring-like effect of the air
chamber. If the air compressibility exerts at least a similar effect
to inclination, it should be considered in the preliminary design.
Regarding to air compressibility, air chamber scale ratio is equal
to square of length
 0.35
18.4°
and Henriques, 2014). With a constant Froude number and a
0.3 geometrically-scaled model (length of the model is 1/ε of full-
18.4°11.5m
OWC efficiency η

0.25 scale), the length scale is thus ultimately defined as ε¼ Lf/Lm. Here,
0.2 m denotes the value of the scaled model and f denotes the value of
the full-scale device. To keep the Froude number constant, the
0.15 p
0.1 time scale is defined as εffiffiffi Tf =Tm. Therefore, within the applic-
¼
0.05 able range of small amplitude wave theory, the motion of the
water column is depicted by the length scale proportional to the
0
geometrical scale, and the time scale is proportional to the square-
0 1 2 3 root of the geometrical scale. Specifically, for the full-scalep ffiffiffimodel,
4 wave height is scaled by ε and wave period is scaled by ε.
Input wave period(s)
Furthermore, the consistencies in the air chamber are domi-
Fig. 20. Experimental results of inclined OWC with different lengths of the air nated by thermodynamic similarities (Falcã o and Henriques,
chamber (θ ¼ 18.4°,H0 ¼ 0:1:). The OWC length was varied as 3.5 m (solid line)
2014). As briefly addressed in Section 4.3, to achieve the same
and
11.5 m (dotted line). The nozzle area ratio is 1/60. effect of air compressibility of chamber air in the 1/ ε scaled model,
the volume of the air chamber (V) must be Vf ¼ ε2Vm, if an iso-
thermal compression process is assumed. Specifically, if the cross-
0.30 section of the model is geometrically scaled, the phenomenon of
the model resultantly corresponds to a full-scale device with the
0.25
OWC efficiency η

0.20
0.15
0.10
5m
0.05 20m
0.00 40m
1.0 2.0 3.0 4.0 5.0
input wave period (s)
Fig. 21. Simulation results of inclined OWC with different lengths of the air
chamber (θ ¼ 18.4°, H0 ¼ 0:1; nozzle area ratio¼ 1/40).
scale ratio of the model, whilst scale ratio of a water column is
cube of length scale ratio. In this respect, the air chamber of the
3.5-m-long cylinder used in the water tank test is so narrow that
spring like effect of air chamber is ignorable. To investigate the
effect of air compressibility, we repeated the experiment on a
cylinder of length 11.5 m.
Fig. 20 plots the results of the water tank test with a nozzle
area ratio of 1/60. Clearly, the cylinder length exerts much smaller
effect on the OWC efficiency than inclination (see Fig. 15).
The same trend emerges in the numerical model. The numer-
ical results for various realistic lengths of the air chamber under
optimum damping conditions (nozzle area ratio¼ 1/40) are
plotted
in Fig. 21. The upper length limit (50 m) was determined from the
geometry of the Blow-Hole project's test site (Miyazaki et al.,
2014). In this figure, we observe that the long chamber achieves a
higher efficiency than the short chamber at periods below the
peak efficiency. Similar length trends are reported in previous
literature (Folley and Whittaker, 2005). As the air chamber enlar-
ges, the efficiency becomes higher and lower at wave periods
below and above the natural frequency, respectively. Comparing
Figs. 15 and 21, we find that the air chamber volume exerts much
less effect than the inclination angle.
4.4. Discussion on similarities between model-scale and large-scale
devices
During the simulation and experimental validation process,
small-scale devices were employed. However, for real sea tests or
commercial operations, larger devices than that used for the
model test must be deployed in order to capture more power from
input waves. In this section, the effect of scale and associated
range are hence addressed relative to this requirement.
First, general similarities between a model-scale device and a
full-scale device are discussed. The similarities in water column
motion are essentially dominated by the Froude number (Falcã o
same length or height. In such cases, the nozzle area ‘ratio’ must
be comparable for both the full-scale and model-scale scenarios.
Furthermore, any similarities related to the Reynolds number
10m (such as from pipe friction losses) are not scalable. However, the
30m effect of pipe friction loss is so small that any potential associated
50m impacts are essentially indiscernible, as further described below.
6.0 Finally, scaling of specific characteristics (introduced in former
sections) is discussed. Firstly, the resonance period is scaled by the
square-root of the length scale. Even if the surrounding geometry
or depth of a seabed changes, the frequency of the wave exciting
force is for the same period of input wave. Therefore, if the mass of
the water column and the added mass are properly defined, the
definition of the undamped resonance period is resultantly valid.
As for the optimum damping factor, the subject factor is to the
2.5th power of length scale ε. Therefore, when setting the power
take-off damping factor to the 2.5th power of the model-scale, the
same efficiency must ultimately be achieved. As for air chamber
volume, the height of the model must be identical to the full-scale
device. In this study, the length of the cylindrical OWC was
modified up to 50 m, including a water column with an 18.4°
inclination. Therefore, the discussion in Section 4.3 is valid for the
air chamber up to about a 49 m length (i.e. 15 m above the still-
water level). As for pipe friction loss and pressure drop in the
power take-off system, the relative importance of these two deltas
is essentially insignificant for two primary reasons: (1) both values
are proportional to the square of air velocity if the nozzle area
ratio
is the same for the model-scale and full-scale devices and (2) the
friction-loss factor (λ) of the pipe varies with Reynolds number. As
such, pipe friction-loss is essentially insignificant if the size of the
device is enlarged.
Hence, all of the subject discussions above were found to
satisfactorily apply within the realm of small-amplitude wave
theory (linear wave theory), with the associated range motion of
the water column assumed equal to 1 DOF oscillation. Whether the
motion could ultimately be assumed as 1 DOF motion or not is
directly related to the inlet width and seabed shape. A previous
report (Webb et al., 2005) points out that the OWC inlet width
must be below 40 m to avoid a non-uniform water elevation sce-
nario. Moreover, the potential effect of seabed shape is pointed to
in the LIMPET test plant report (Heath et al., 2002).
5. Conclusion
This study investigated how the change in the motion direction
of a water column caused by inclining the OWC affects the oscil-
lation characteristics of the water column.
 First, we proposed a mechanical model of the inclined OWC,
which accounts for the reduced restoring force in an inclined
OWC. The model was validated in comparisons with a real inclined
cylindrical OWC tested in a water tank. The resonant period was
longer in the inclined OWC than in the vertical OWC.
Next, analogous to a pendulum swinging on a slope, we eval-
uated the undamped resonant frequency of an inclined OWC
subjected to the reduced restoring force. The simulated resonant
periods closely matched the wave periods that maximized the
efficiency in the water tank, unlike the resonant periods derived
without the gravitational response. Moreover, the optimum
damping at the undamped resonant period of the OWC was
accurately predicted by the mechanical model as well as by the
traditional vertical OWC. The effect of the chamber air volume was
also investigated and exhibited trends similar to previous reports
on vertical OWCs. Specifically, the OWC efficiency was higher
(lower) at wave periods below (above) the natural frequency of
the OWC. However, the effect of the air chamber volume is much
less than the inclination angle effect.
In conclusion, the changing direction of motion in an inclined
OWC significantly affects the oscillation characteristic of the OWC.
In particular, the resonant period (an important parameter in
OWC design) is significantly prolonged by inclination. Therefore,
we must consider the effect of the changing direction of motion
when considering inclined OWC wave energy converters.
Acknowledgment
We would like to thank the National Maritime Research Insti-
tute for the help given in performing water tank tests. This work
was supported by JSPS KAKENHI, Grant number 268635.
References
Bayoumi, S., Incecik, A., El-Gamal, H., 2014. Dynamic modelling of Spar-Buoy
oscillating water column wave energy converter. Ships Offshore Struct. 10, 601–
608. http://dx.doi.org/10.1080/17445302.2014.942086.
Brendmo, A., 1997. Linear modelling of oscillating water columns including viscous
loss. Appl. Ocean Res. 18, 65–75.
Ceballos, S., Rea, J., Lopez, I., 2013. Efficiency optimization in low inertia wells
turbine-oscillating water column devices. IEEE Trans. Energy Convers. 28, 553–
564.
Department of Trade and Industry Great Britain, 2004. Near Shore Floating Oscil-
lating Wave Column: Prototype Development and Evaluation.
Falcã o, A., Henriques, J., 2014. Model-prototype similarity of oscillating-water-
column wave energy converters. Int. J. Mar. Energy 6, 18–34. http://dx.doi.org/
10.1016/j.ijome.2014.05.002.
Falcã o, A., Justino, P.A.P., 1999. OWC wave energy devices with air flow control.
Ocean Eng. 26, 1275–1295. http://dx.doi.org/10.1016/S0029-8018(98)00075-4.
Folley, M., Whittaker, T., 2005. The effect of plenum chamber volume and air tur-
bine hysteresis on the optimal performance of oscillating water columns. In:
Proceedings of the 24th International Conference on Offshore Mechanics and
Arctic Engineering (OMAE), pp. 1–6.
Heath, T., Whittaker, T., Boake, C., 2002. Research into the further development of
the limpet shoreline wave energy plant. DTI Sustainable Energy Programmes.
Hotta, H., Washio, Y., Ishii, S., Masuda, Y., 1986. The operational test on the shore
fixed OWC type wave power generator. In: Proceedings of the Fifth Interna-
tional Offshore Mechanics and Arctic Engineering Symposium vol. 2, 546–552.
Ishii, S., Miyazaki, T., Masuda, Y., Kai, G., 1982. Reports and future plans for the
Kaimei project. In: Proceedings of the 2nd International Wave Energy Utilisa-
tion, 305–321.
Kinoshita, T., Masuda, K., Miyajima, S., Kato, W., 1984. Research on the system
simulation for a fixed O.W.C. type wave energy absorber. J. Soc. Nav. Archit.
Japan, 255–263.
Miyazaki, T., Iida, M., Komiya, T., Iino, M., Chino, H., Takeuchi, T., 2014. Project of the
blow-hole wave power generator system. In: Proceedings of Grand Renewable
Energy, 2014.
Osawa, H., Miyazaki, T., 2005. Technical manual for oscillating water column type
wave power device. In: Proceedings of the Fifteenth International Offshore and
Polar Engineering Conference.
Osawa, H., Miyazaki, T., 2004. Wave-PV hybrid generation system carried in the
offshore floating type wave power device “Mighty Whale. Ocean. ’04 MTS/IEEE
Techno-Ocean'04 4, 1860–1866.
Payne, G.S., Taylor, J.R.M., Bruce, T., Parkin, P., 2008. Assessment of boundary-
element method for modelling a free-floating sloped wave energy device. Part
1: Numerical modelling. Ocean Eng. 35, 333–341.
Pecher, A., Crom, I., Le, Kofoed, J., 2011. Performance assessment of the Pico OWC
power plant following the EquiMar Methodology. In: Proceedings of Interna-
tional Offshore and Polar Engineering Conference 8, 548–556.
Ram, K., Asid Zullah, M., Rafiuddin Ahmed, M., Lee, Y.-H., 2010. Experimental stu-
dies on the flow characteristics in an inclined bend free OWC device with
parallel walls. In: Proceedings of Grand Renewable Energy, 2010.
Sheng, W., Flannery, B., 2012. Experimental studies of a floating cylindrical OWC
WEC. In: Proceedings of OMAE2012, pp. 1–10.
Sheng, W., Thiebaut, F., Babuchon, M., Brooks, J., Lewis, A., Alcorn, R., 2013. Inves-
tigation to air compressibility of oscillating water column wave energy con-
verters. In: Proceedings of the 32nd International Conference on Ocean, Off-
shore and Arctic Engineering, pp. 1–10.
Stappenbelt, B., Cooper, P., 2010. Mechanical model of a floating oscillating water
column wave energy conversion device. 2009 Annual Bulletin of the Australian
Institute of High Energetic Materials 1, 34–45.
Suzuki, M., 2005. Numerical methods to predict characteristics of oscillating
water column for terminator-type wave energy converter. Int. J. Offshore Polar
Eng. 15 5381.
Suzuki, M., Arakawa, C., Takahashi, S., 2004. Performance of wave power generating
system installed in breakwater at Sakata Port in Japan. In: Proceedings of the
Fourteenth International Offshore and Polar Engineering Conference.
Webb, I., Seaman, C., Jackson, G., 2005. Oscillating Water Column Wave Energy
Converter Evaluation Report. Carbon Trust.
Yukihisa, W., Hiroyuki, O., Teruhisa, O., Hiroyuki, N., Shuzou, O., Yoshinori, N., 2001.
A study on characteristics of generated output of the offshore floating type
wave power device “Mighty Whale.”. J. Soc. Nav. Archit. Japan 190, 395–405.