Applied Energy 228 (2018) 82–96

Contents lists available at ScienceDirect

Applied Energy
journal homepage: www.elsevier.com/locate/apenergy

Hydrodynamic performance of single–chamber and dual–chamber T

offshore–stationary Oscillating Water Column devices using CFD
⁎
Ahmed Elhanafia,b, , Gregor Macfarlanea, Dezhi Ningc,d
a
National Centre for Maritime Engineering and Hydrodynamics, Australian Maritime College, University of Tasmania, Launceston, Tasmania 7250, Australia
b
Department of Naval Architecture and Marine Engineering, Alexandria University, Alexandria, Egypt
c
State Key Laboratory of Coastal and Offshore Engineering, Dalian University of Technology, Dalian 116024, China
d
Offshore Renewable Energy Research Center, Dalian University of Technology, Dalian 116024, China

H I GH L IG H T S

• The capture width ratio of single–chamber and dual–chamber OWC devices is compared.
• Different wave conditions, power take–off damping values and chamber geometries are investigated.
• The relevance of 3D effects on device capture width ratio is illustrated.
• The capture width ratio of a single–chamber OWC device is significantly improved using the dual–chamber system.

A R T I C LE I N FO A B S T R A C T

Keywords: The Oscillating Water Column (OWC) is considered to be one of the most promising Wave Energy Converter
Wave energy (WEC) concepts in terms of practicality, survivability and efficiency. To date, most research has focussed on
OWC device single–chamber devices, but it is suggested that significant increases in energy extraction can be achieved from
Dual–chamber OWC device dual–chamber devices. This paper investigates, using well–validated 2D and 3D CFD models based on the
Hydrodynamic performance
Reynolds Averaged Navier–Stokes (RANS) equations and the Volume of Fluid (VOF) method, the hydrodynamic
Capture width ratio
CFD
performance of various dual–chamber offshore–stationary OWC–WECs and compares the results to single–-
chamber OWC devices. The effect of chamber lip draught, chamber length in wave propagation direction and the
power take–off (PTO) damping on the capture width ratio (power extraction efficiency) of each OWC device was
studied over a wide range of wave periods for a constant regular wave height. It was found that all the para-
meters tested were important for the design of efficient OWC devices, and the dual–chamber device provided
superior results to the single–chamber device, especially over the intermediate and long wave periods where the
capture width ratio could be improved by a maximum of about 140%; hence extracting significantly more
energy. The effectiveness of using the dual–chamber system was more obvious when 3D effects were considered.
The findings of this paper contribute towards the design and operation of practical OWC devices for efficiently
utilizing ocean waves to produce electricity.

1. Introduction opening and a power take–off (PTO) system. The principal operation of
a typical OWC device is very simple such that ocean wave oscillations
To cope with the continuous increase in the global energy demand drive the motion of the water column inside the chamber, which in turn
and tackle the environmental problems accompanied with the excessive drives the PTO system (an air turbine with an electric generator). Over
use of fossil fuels, marine renewable energy including wave energy the last few decades different experimental, analytical/theoretical and
resources have drawn the attention to many developers and researchers numerical techniques have been developed and utilized to extensively
worldwide. Among the several techniques available to extract wave investigate and improve the performance of OWC devices. The reader is
energy, the Oscillating Water Column (OWC) device is internationally referred to Ref. [1] for an extensive review of these methods.
recognised to be one of the most leading technologies. An OWC device In comparison to onshore or nearshore OWC devices where chamber
consists of two main parts: a hollow chamber with an underwater rear wall is a part of the shoreline or extending to the seabed,

⁎
Corresponding author at: National Centre for Maritime Engineering and Hydrodynamics, Australian Maritime College, University of Tasmania, Launceston, Tasmania 7250, Australia.
E-mail address: Ahmed.Elhanafi@utas.edu.au (A. Elhanafi).

https://doi.org/10.1016/j.apenergy.2018.06.069
Received 20 March 2018; Received in revised form 3 June 2018; Accepted 16 June 2018
Available online 20 June 2018
0306-2619/ © 2018 Elsevier Ltd. All rights reserved.
A. Elhanafi et al. Applied Energy 228 (2018) 82–96

respectively, offshore OWC devices have all walls detached from the novelty of this paper. The remainder of this paper is organized as fol-
seabed allowing ocean waves to pass underneath and around the de- lows: the numerical modelling and system performance calculations are
vice. Most of the previous researches were focused on single–chamber described in Section 2; the results obtained and their implications are
OWC devices where the performance of a typical device is best at and discussed in Section 3; followed by conclusions and future works in
around chamber resonance. A maximum theoretical efficiency of 1.0 Section 4.
was achieved for onshore/near–shore devices with a rear lip standing
on the seabed, whereas this value decreases to 0.5 for offshore OWC 2. Numerical modelling
devices with symmetrical partially immersed front and rear lips [2].
Using asymmetrical lip draughts (front lip draught is smaller than rear In this study, we utilized 2D and 3D CFD models that were pre-
lip draught), the performance of offshore OWC devices could sig- viously validated in Refs. [4,18,19,24,25] for single–chamber OWC
nificantly be improved, especially for the intermediate– and high–- models. In literature, it was shown that well–validated numerical
frequency wave regimes [3,4]. Several studies agreed on the im- models [5,26] for single OWC devices can be employed to investigate
portance of chamber lip draught, chamber length in wave propagation the performance of dual–chamber OWC devices [9,10,27]. Therefore,
direction and PTO damping on OWC devices [3,5–8]. However, these we assume these CFD models are theoretically adaptable for the
design parameters in combination can be used to tune the device to a dual–chamber OWC models tested in this study. These models assume
range of wave conditions. incompressible fluid in the OWC chamber considering that air com-
In order to enhance device performance and extract more wave pressibility effect is negligible at small scale OWC devices [19] such as
energy over a broader range of wave conditions, the concept of the devices investigated in this study. The continuity and RANS (Rey-
dual–chamber OWC devices for onshore and offshore multipurpose nolds Averaged Navier–Stokes) equations were used to describe the
applications could be beneficial. However, there is a very limited re- flow motion of the incompressible fluid, and the Volume of Fluid (VOF)
search on this subject. For instance, using a 2D fully nonlinear potential method [28] was used to model and track the free surface motion. All
flow solver, Wang et al. [9] studied the effect of chamber length and computations were performed using the RANS–VOF solver in
draught of a dual–chamber onshore OWC system for different wave STAR–CCM + code that uses a finite volume method to discretise the
conditions under a constant PTO damping that was simulated via a integral formulation of the RANS equations and a predictor–corrector
shared orifice between the two chambers. They reported superior re- approach to link the continuity and momentum equations [29]. The
sults of the dual system near the resonant frequency when compared to two–equation shear stress transport (SST) k-ω turbulence model was
a single–chamber device, and highlighted the importance of the barrier implemented to model the Reynolds stresses in RANS equations with 10
wall between both chambers. A complementary study was performed prism layers of 1.5 stretching factor and a y+ value of 1 to model the
by Ning et al. [10] to gain a better understanding of the dual–chamber boundary layer.
system through analyzing the free surface elevation in the chamber as The computational domain (numerical wave tank, NWT) used in
well as the differential air pressure. They stressed again that the use of this study (see Fig. 1) was of length equal to 10 wavelengths (L) in-
barrier wall with a proper horizontal location can improve the hydro- cluding one wavelength assigned ahead of the pressure outlet boundary
dynamic efficiency of the system, and a marginal improvement in (on the right side of the NWT) for the damping zone where a resistance
system efficiency over the short–wave regime can be achieved with term was added to the equation of the vertical velocity component [30].
increasing the barrier wall draught. Aiming at reducing the cost of It is worth noting that the length of the NWT was adjusted for each
energy, He et al. [11] investigated, using small–scale physical model wave period tested in this study to maintain a total length of 10L so that
experiments, the hydrodynamic performance of a floating–moored at least eight wave cycles could be collected before the incoming waves
box–type breakwater integrated with dual pneumatic chambers in a interfere with the waves reflected from the OWC structure and/or the
wave flume for different regular wave conditions. They tested several outlet boundary [31]. The NWT was of 2.5 m height divided into 1 m
designs of symmetrical and asymmetrical chambers with different for the air phase and 1.5 m for the water phase. To be able to ignore
chamber length and draught under a constant PTO damping simulated tank sidewalls effect in the case of 3D modelling, the width of the 3D
with a slot opening, resulting that most of the energy extracted by the NWT was set to be 8.75 times the width of the OWC device, which
system was attributed to the front chamber, whereas the rear chamber fulfilled the requirements provided by Chakrabarti [32] and Elhanafi
only provided a supplement. This could be a result of the lengthy et al. [21]. However, to reduce the required computational resources
box–type breakwater that dissipates most of the energy transmitted and time, only half the 3D NWT was modelled with a symmetry plane,
from the front chamber before entering the rear chamber. which has been proved to be a valid procedure [19]. The OWC model
Although the above–mentioned researches on the dual–chamber was placed at 6.5 L from the inlet boundary (on the left side of the
OWC system provided a marginal improvement in the system effi- NWT) where wave velocity components were prescribed. A wave hy-
ciency, there are important design parameters that should be in- drostatic pressure was assigned to the top outlet boundary of the NWT,
vestigated with a major objective of improving device efficiency; hence whereas the tank bottom was set as a slip wall boundary. The front and
extracting more wave energy. The main contribution of this study lies in back sides of the 2D NWT were set as symmetry planes, but the side of
complementing and filling the gap in previous research by providing the 3D NWT was set as a slip wall boundary. To accommodate the
in–depth understanding of various key design parameters (chamber waves reflected by the OWC model and avoid any numerical damping
length, asymmetric lip draught, PTO damping and 3D modelling) af- in the wave height, a free surface zone of height equal to 2.5 H (H is
fecting the energy extraction performance of a dual–chamber OWC wave height) was used around the free surface interface for the 2D
device. Experimentally validated Computational Fluid Dynamics (CFD) model, whereas a zone height of 1.5 H for the case of 3D modelling was
models can capture detailed physics such as strong nonlinearity, com- found to be reasonable and reduces the computation cost [19].
plex viscous effects, turbulence and vortex shedding. Therefore, they NWT meshing was conducted using the built–in automatic meshing
have been employed by a large number of researchers to numerically technique in STAR–CCM+ of a trimmed cell mesher and surface re-
test single–chamber OWC devices for different objectives. These studies mesher. A base cell size of 400 mm was applied to the whole fluid
include energy balance analysis [12,13], the effect of wave conditions domain that was progressively refined at the free surface zone and the
and PTO damping [14–17], chamber underwater geometry [3,4,6,18], OWC model. At least 16 and 74 cells per wave height (z–direction) and
scaling and air compressibility [19,20] on device performance as well wavelength (x–direction, i.e., wave propagation direction) were con-
as investigating the wave forces, device motion and survivability sidered, respectively, which are close to what is recommended by ITTC
[21–23]. However, none of this research applied CFD to examine and [33] and CD–Adapco [29] of 20 and 80 cells per wave height and
design dual–chamber OWC devices, which represents the second wavelength, respectively. The cell aspect ratio (i.e., the ratio between

83
A. Elhanafi et al. Applied Energy 228 (2018) 82–96

Fig. 1. Computational fluid domain (not to scale) of (top): 2D model and (bottom): 3D model.

the cell size in the longitudinal (Δx) and vertical (Δz) directions) was (slot opening for 2D modelling or orifice for 3D modelling, see Fig. 1c
kept to a maximum of 16 [12]. The time step (Δt) was carefully selected and f) respectively were sufficient to provide the same results of the
for each wave period (T) so that it satisfies the time step requirement most refined mesh tested but with less computation time. For the 3D
(Δt = T/2.4n, where n is number of cells per wavelength) by CD–A- modelling, the use of a symmetry plane has proven to be valid, and a
dapco [29] with a second–order temporal discretization scheme. This cell size of Δy = 100 mm was used in the tank transverse (y) direction
results in a Courant number (C = Cg · Δt/Δx, where Cg is the wave group as recommended by Elhanafi et al. [19]. It is worth noting that these
velocity) less than 0.5; hence, the High–Resolution Interface Capturing setting have proven to provide good agreement when compared to
(HRIC) scheme that is suited for tracking sharp interfaces is guaranteed experimental results of a single–chamber OWC device [22,24,25].
to be used during the simulations [29]. Mesh convergence studies of a The principal dimensions of the 1:50 scale single–chamber and
dual–chamber OWC device in 2D and 3D were conducted where the cell dual–chamber OWC devices tested in this study are illustrated in Fig. 2.
size of OWC and PTO surfaces were refined (see Appendix A). These The 3D device had a total width of a = 224 mm (200 mm for the
studies revealed that surface refinements of 6.25 mm and 0.78125 mm chamber and 12 mm for the side plates). Upon the results discussed by
cell size for the OWC device (see Fig. 1b and e) and the PTO surfaces Elhanafi et al. [4] that show the effectiveness of increasing the rear lip

84
A. Elhanafi et al. Applied Energy 228 (2018) 82–96

Fig. 2. Schematic of the 1:50 offshore–stationary OWC devices. (a) Single–Chamber and (b) Dual–Chamber. [All dimensions are in mm.]

Table 1
Test variables and conditions.
Device d1 [mm] b1 [mm] d2 [mm] b2 [mm] Model ID e1 [%] e2 [%] T [s] H [mm]

Single–Chamber OWC device 25 600 – – S4 0.50–3.00 – 0.9, 1.0, 1.1, 1.2, 1.3, 1.4, 1.5, 1.6, 1.7, 1.8 and 2.0 50
50 600 – – – 1.67 –
100 600 – – – 1.67 –
200 600 – – – 1.67 –
300 600 – – S3 0.50–3.00 –
25 500 – – – 1.67 –
25 400 – – – 1.67 –
25 300 – – S2 0.50–3.00 –
25 200 – – – 1.67 –
25 100 – – – 1.67 –
300 300 – – S1 0.50–3.00 –
Dual–Chamber OWC device 25 300 25 300 D1 1.67 1.67
25 300 50 300 D2 1.67 1.67
25 300 100 300 D3 1.67 1.67
25 300 200 300 D4 1.67 1.67
25 300 300 300 D5 0.50–3.00 0.50–3.00
25 300 300 600 D6 0.50–3.00 0.50–3.00
25 600 300 600 D7 0.50–3.00 0.50–3.00
25 600 300 300 D8 1.67 1.67

draught of an offshore OWC device, the rear draught of the OWC de- where A is incoming wave amplitude, ρ is water density, k is the wave
vices tested in this study was fixed at 300 mm. The device performance ω2
number given by the dispersion relationship g
= k tanh(kh) , L is in-
was investigated over a range of wave periods T from 0.9 s to 2.0 s with 2π
coming wavelength calculated as k = L and h is the still water depth of
0.1 s increment for a constant wave height H = 50 mm and different
1.5 m.
chamber length in wave propagation direction (b1 and b2). The effect of
The airflow rate Q (t ) was directly monitored and recorded by in-
PTO damping on device performance was investigated by changing the
tegrating the air vertical velocity (V) through the slot/orifice opening
opening ratio (e) on the top plating (e is the ratio between the slot/
area using the surface integral report in STAR–CCM+. A negligible
orifice opening area to the chamber waterplane area). Different opening
difference has been reported [31] when this approach was utilized in
ratios of 0.50%, 1.00%, 1.67%, 2.33% and 3.00% were considered in
comparison to the conventional method used in physical experiments of
this study. Table 1 summarizes all the parameters tested in this paper.
small scale OWC models (air compressibility is negligible) where the
The capture width ratio of an OWC device (ζ) is given by Eq. (1)
airflow rate is calculated as the product of chamber waterplane area
[25]. The mean incident wave energy flux (power, PI ) per unit width is
and chamber free surface vertical velocity (derived from chamber free
defined in Eq. (2) as the product of the total (potential and kinetic)
surface oscillation). The differential air pressure ΔP (t ) of the OWC
wave energy per unit ocean surface area and the group velocity (Cg )
chamber was calculated using measurements recorded at four points:
given by Eq. (3). The time–averaged extracted pneumatic power (PEx ) is
two inside the pneumatic chamber (one on each side of the PTO at half
calculated from Eq. (4) where Q (t ) is the airflow rate through the PTO
distance between the center of the slot/orifice opening and chamber
and ΔP (t ) is the chamber differential air pressure. For the dual–chamber
lip) and another two points outside the chamber on the top boundary
system, PEx is the summation of the extracted power from each
domain. The differential air pressure was then calculated as the average
chamber.
difference between the inside and outside monitoring points. Ad-
ζ = PEx /(PI a) (1) ditionally, the free surface elevation inside each chamber (ηOWC) was
monitored at three locations and then averaged.
1
PI = ρgA2 Cg
2 (2)
3. Results and discussion
ω⎛ 2kh ⎞
Cg = 1+ ⎜ ⎟

2k ⎝ sinh(2kh) ⎠ (3) 3.1. Single–chamber OWC device

1 T Elhanafi et al. [4] showed that shortening the front lip and
PEx =
T
∫0 ΔP (t ) Q (t ) dt
(4) lengthening the rear lip of an offshore OWC device could significantly

85
A. Elhanafi et al. Applied Energy 228 (2018) 82–96

Fig. 3. Effects of chamber front lip draught and chamber length on the capture width ratio (ζ) of a single–chamber OWC device with a slot opening ratio e1 = 1.67%.
(a) Effect of chamber front lip draught (d1) and (b) effect of chamber length (b1).

improve device capture width ratio. Also, the results presented and
discussed by Elhanafi et al. [18] demonstrated that a minimum front lip
draught equal to the incoming wave amplitude is necessary to avoid air
leakage that negatively affects device performance. Therefore, the
minimum front lip draught (d1) in this study was set as d1 = H/
2 = 25 mm. Generally, for a typical OWC device, the capture width
ratio increases as wave period decreases (i.e., frequency increases) to a
maximum value at chamber natural (resonant) period and then de-
creases with a further decrease in wave period. The effect of front lip
draught on the capture width ratio of the single–chamber OWC device
considered in this study is shown in Fig. 3a for a constant chamber
length b1 = 600 mm and a constant PTO damping represented by a slot
opening ratio e1 = 1.67%. It can be seen that decreasing the front lip
draught from d1 = 300 mm (6H) to d1 = 25 mm (H/2) improved device
capture width ratio for all wave periods tested, especially for the short
and intermediate wave periods, while the maximum capture width ratio
increased from ζ = 0.32 at T = 1.5 s to ζ = 0.78 at T = 1.3 s. The shift
in the resonant period to a lower value is due to the fact that reducing
the mass of water column inside the chamber by shortening chamber
draught and/or length leads to a reduction in system inertia and natural
period as given by Eq. (5) [34]. For example, using Eq. (5) and as-
suming that for 2D modelling the water column area Swc is equal to the
chamber length (b1) x 1.0, the resonant period of the device with
d1 = 25 mm, b1 = 300 mm is about 1.00 s, which increases to about
1.18 s when chamber length increases to b1 = 600 mm for the same
draught, and it further increases to 1.58 s when the front lip draught
increased to d1 = 300 mm for b1 = 600 mm. These results are in very
good agreement with the resonant period shown in Fig. 3, and any
discrepancy could be attributed to the use of the empirical coefficient of
0.41 in Eq. (5) and/or to the fact that the PTO damping effect was not
Fig. 4. Time–series results illustrate the effect of chamber front lip draught (d1)
considered in this equation
on the performance parameters of single–chamber OWC devices with a
chamber length b1 = 600 mm a slot opening ratio e1 = 1.67% at a wave period
d + 0.41 S wc T = 1.3 s. ηOWC: chamber free surface oscillation, V: airflow velocity through
T0 = 2π
g (5) the slot opening, Q: airflow rate through the slot opening per unit chamber
width, ΔP: chamber differential air pressure and PE: extracted instantaneous
where d is the water column draft in the chamber, Swc is the chamber pneumatic power per unit chamber width.
(water column) waterplane area and g is the gravitational acceleration.
To further clarify the effect of the front lip draught on device per-
formance, the time–series of the performance parameters (i.e., chamber
free surface oscillation ηOWC, airflow velocity through the PTO V, airflow The general trends of the results shown in Fig. 3a are, so far, similar
rate per unit chamber width Q, chamber differential air pressure ΔP, to those reported in Refs. [4,18] for offshore OWC devices with a
extracted instantaneous pneumatic power per unit chamber width PE) chamber length b1 = 300 mm, except for the short–period regime,
are shown in Fig. 4 for two cases of d1 = 25 mm and 300 mm at a con- which can be explained as follows. For onshore OWC devices, the
stant chamber length b1 = 600 mm and a wave period of T = 1.3 s (re- chamber length has been found to have a significant influence on device
sonance at d1 = 25 mm, see Fig. 3a). These results illustrate how short- efficiency [5,35]. To further understand the relevance of chamber
ening the front lip draught tunes the device towards the high–frequency length (b1) to the capture width ratio of an offshore single–chamber
regime and facilitates the incoming waves to smoothly enter the chamber OWC device, six different chamber lengths in the range of b1 = 100 mm
(see the increases in ηOWC) which in turn increases the airflow rate and to 600 mm have been tested for a given front lip draught d1 = 25 mm,
the differential air pressure; hence extracting more energy. and results are presented in Fig. 3b. The results in this figure agree with

86
A. Elhanafi et al. Applied Energy 228 (2018) 82–96

Fig. 5. Time–series results illustrate the effect of chamber length (b1) on the performance parameters of single–chamber OWC devices with a front lip draught
d1 = 25 mm and a slot opening ratio e1 = 1.67% at wave periods (a): T = 1.0 s and (b): T = 1.6 s.

previous experimental [5] and numerical [35] studies (that were based 3.2. Dual–chamber OWC device performance
on testing three different chamber lengths for two different onshore
OWC devices) such that increasing the chamber length results in es- In this section, we investigate the hydrodynamic performance of a
sentially the same effect of the added mass when increasing the lip dual–chamber OWC device with a constant PTO damping simulated
draught causing an increase in the resonant period. This is also ac- with a slot opening ratio e1 = e2 = 1.67%. Recently, He et al. [11]
companied with increasing and decreasing device capture width ratio concluded that the power extracted by the rear chamber of a dual–-
for long and short wave periods respectively. Moving the chamber re- chamber floating breakwater–OWC device was only a supplement due
sonant period to a longer wave period by lengthening the chamber to the small energy transmitted from the front chamber (i.e., there is a
length means that the device became more tuned to the low–frequency small portion of the incoming energy available for the rear chamber). In
waves (i.e., long wavelengths), but it is no longer suitable for efficiently that study, the front and rear chambers were separated by a solid box of
extracting wave energy from short waves where the lengthy chamber length longer than the maximum chamber length. This in turn obstructs
breaks down the rigid–piston like (flat) motion of the free surface inside wave energy to enter the rear chamber. Furthermore, Elhanafi et al. [4]
the chamber which in turn reduces device performance. This is de- showed that increasing the thickness of chamber seaward lip of an OWC
monstrated by the time–series of chamber free surface oscillation and device has a negative influence on device capture width ratio. There-
other performance parameters for the cases of b1 = 300 mm and fore, to reduce this obstruction, the thickness of all lips, including the
600 mm as shown in Fig. 5a and 5b at T = 1.0 s and 1.6 s respectively. It intermediate lip (i.e., the lip shared by both chambers), of the dual–-
is clear that as wave period decreases the distortions in the free surface chamber OWC device tested in this study was set as 12 mm. Based on
oscillation inside the chamber become more pronounced, which are the results discussed in Section 3.1, the draught of both front and rear
also reflected in the other performance parameters. It is worth noting lips of the dual–chamber OWC device was set as 25 mm and 300 mm
that, although the airflow velocity achieved with the longer chamber of respectively (see Fig. 2b).
b1 = 600 mm was smaller than that of b1 = 300 mm, the airflow rate The effect of the intermediate lip draught (d2) and chamber length
with b1 = 600 mm was higher than that of b1 = 300 mm, and this is on the dual–chamber OWC device capture width ratio is shown in
because the larger slot size (two times) of b1 = 600 mm that was used to Fig. 6–left and 6–right respectively. From Fig. 6–left, it can be seen that
maintain the same opening ratio (e1 = 1.67%) of b1 = 300 mm. Al- changing the intermediate lip draught significantly influenced the
though Ning et al. [5] observed negligible impact of chamber length on performance of each chamber such that lengthening and shortening this
the peak capture width ratio of onshore OWC devices, results in Fig. 3b lip improved the capture width ratio of the front and rear chambers
as well as those from Ref. [35] indicate that the peak capture width respectively, which can be explained as follows. For the 2D nature of
ratio decreased as chamber length increased. the CFD model considered in this study (except Section 3.4), wave
Overall, there was no optimum chamber length for the whole wave energy is only allowed to enter the rear chamber from underneath the
period range tested in this study; instead, chambers of different lengths intermediate lip. Therefore, the larger the intermediate lip draught, the
were required to maximize device capture width ratio for a certain smaller portion of the incoming wave energy transmitted into the rear
wave period range. Among the six chamber lengths tested, Fig. 3b chamber, especially for short wave periods where more wave energy is
shows that chamber lengths of b1 = 600 mm and 300 mm could be reflected by the device [4,31,36] allowing more energy to be extracted
selected as optimum chamber sizes for wave periods longer and smaller by the front chamber. It is also expected that both chambers of a
than T = 1.2 s respectively. Therefore, the performance of a dual–- dual–chamber OWC device influence each other. For example, the front
chamber OWC device based on these two chamber lengths is discussed chamber of D5 (d2 = 300 mm) and the rear chamber of D1 (with
in the following section. d2 = 25 mm) have the same geometrical characteristics of the

87
A. Elhanafi et al. Applied Energy 228 (2018) 82–96

Fig. 6. Effect of intermediate lip draught (d2) and chambers length (b1 and b2) on the capture width ratio (ζ) of a dual–chamber OWC device with a slot opening ratio
e1 = e2 = 1.67%. (Left): effect of d2 for d1 = 25 mm and b1 = b2 = 300 mm and (Right): effect of b1 and b2 for d1 = 25 mm and d2 = 300 mm.

single–chamber device of d1 = 25 mm and b1 = 300 mm presented in width ratio for all periods tested with the peak capture width ratio
Fig. 3b. However, only the front chamber of D5 showed a similar trend increasing from 0.78 with d2 = 25 mm to about 0.89 with d2 = 300 mm
to the single chamber while the rear chamber of D1 provided a smaller at the same wave period of T = 1.0 s.
capture width ratio for all periods tested. Furthermore, it is clear that With a deep intermediate lip draught, the front chamber serves as
there is a sudden drop in the capture width ratio of the front chamber at the main wave energy converter for the intermediate– and short–wave
a wave period corresponds to the peak capture width ratio of the rear regimes while both front and rear chambers work together to improve
chamber for only cases of d2 ≤ 100 mm. This indicates that, the deeper device performance in the intermediate– and long–wave regimes.
the intermediate draught, the lesser influence between chambers, Having achieved a higher capture width ratio for the intermediate– and
which can also be attributed to the limited energy available for the rear short–wave regimes with the dual–chamber device of D5, it is im-
chamber with deep d2 as discussed above. portant to investigate other parameters such as chamber length that
The capture width ratio achieved from the front chamber with the could improve device performance for the long–wave regime as dis-
deepest draught tested of d2 = 300 mm was much higher than that of cussed in Section 3.1. Therefore, three different combinations of
the rear chamber with the shortest draught of d2 = 25 mm. The su- chamber length have been tested and their effect on device capture
perior results of the front chamber to the rear chamber is in line with width ratio is shown in Fig. 6–right for a fixed intermediate lip draught
the experimentally–based conclusion of He et al. [11], but it is also d2 = 300 mm. The results illustrate that increasing the length of the
worth mentioning that the effectiveness of using a small intermediate front chamber (D7 and D8) and/or the rear chamber (D6 and D7)
lip thickness (12 mm) was demonstrated by the higher capture width twofold shifted the peak capture width ratio to a longer period, in-
ratio of the rear chamber in this study when compared to that presented creased and decreased the capture width ratio for long and short wave
in Ref. [11]. The overall capture width ratio of the dual–chamber was periods, respectively, which was similar to what was observed for the
found to follow the same trend of the front chamber but with a higher single–chamber device discussed in Section 3.1. Considering that wave
capture width ratio. Also, the dual–chamber capture width ratio was power in short waves is less than that in long waves (group velocity
less sensitive to the variation in the intermediate lip draught up to decreases as wave period decreases), it is important to keep the device
d2 = 100 mm. However, a further increase in d2 increases the capture capture width ratio as high as possible for the intermediate– and

88
A. Elhanafi et al. Applied Energy 228 (2018) 82–96

Fig. 7. PTO damping effect on the capture width ratio (ζ) of Single–Chamber OWC devices with different front lip draughts (d1 = 25 mm and 300 mm) and chamber
lengths (b1 = 300 mm and 600 mm).

short–wave regimes to extract as much of this power as possible. were required to maximize the capture width ratio for the inter-
Therefore, the option of increasing the length of the front chamber mediate–wave period regime, while larger damping values provided
should be eliminated for the present conditions. On the other hand, better performance for the long– and short–wave regimes, which agree
increasing the rear chamber length provided the opportunity to in- with other experimental [24,25] and numerical [31] results. However,
crease the capture width ratio of the rear chamber for the intermediate– the results for OWC devices with a short front lip draught (d1 = 25 mm,
and long–wave regimes, while the reduction in the capture width ratio Fig. 7b and d) illustrated less sensitivity of the optimum PTO damping
of the rear chamber for the intermediate– and short–wave regimes re- to the wave period such that a single optimum PTO damping corre-
sulted in increasing the capture width ratio of the front chamber over sponds to an opening ratio of e1 = 1.00% could be used for all wave
the intermediate–wave regime. Overall results in Fig. 6c–right demon- periods tested, except for the very long wave periods of T > 1.7 s that
strate that the dual–chamber device D6 could be considered as the required a larger PTO damping (e1 = 0.50%).
optimum device for the wave conditions tested in this study under the For the effect of the PTO damping on dual–chamber OWC devices,
PTO damping induced by slot opening ratio e1 = e2 = 1.67%. five different slot opening ratios inducing symmetrical pneumatic
damping for both front and rear chambers as well as one case of un-
3.3. Effect of PTO damping symmetrical damping were considered for three different dual–chamber
geometries of D5, D6 and D7. Results for D5 and D6 were quite similar;
The results discussed in the previous sections were based on a and therefore, for clarity only results for D6 and D7 are presented in
constant PTO damping. In this section, five different PTO damping Fig. 8. The PTO damping effect on the front chamber (Fig. 8a) was
conditions corresponding to slot opening ratios of 0.50%, 1.00%, similar to that observed for single–chamber devices of similar geome-
1.67%, 2.33% and 3.00% are tested and their impact on the capture tries. For example, see Fig. 7b and d for results of single–chamber OWC
width ratio of single–chamber and dual–chamber OWC devices is devices of geometries similar to the front chamber of D6 and D7 re-
shown in Figs. 7 and 8 respectively. The results of the single–chamber spectively. Similarly, the response of the rear chamber to the different
OWC device for two different chamber lengths (b1 = 300 mm and PTO damping values showed a correlation between the induced
600 mm) and two front lip draughts (d1 = 25 mm and 300 mm) showed damping and wave period (Fig. 8b), which is the same conclusion
that, for all PTO damping values tested, devices with a short front lip drawn for the single devices with a deep front lip draught as shown in
draught (d1 = 25 mm, Fig. 7b, d) provided a higher capture width ratio. Fig. 7a and c.
There is no single PTO damping condition that can be used as an op- The interaction between the front and rear chambers seems to alter
timum value for all wave periods, and changing the device geometry the relevance of the induced PTO damping to the rear chamber per-
(chamber draught and length) changes the optimum PTO value for a formance compared to the single–chamber devices, especially around
given wave period. For example, at a wave period T = 1.0 s, the op- the peak capture width ratio, such that for all chamber lengths tested
timum opening ratio (corresponds to optimum PTO damping) was (b2 = 300 mm for D5 and 600 mm for D6 and D7) the opening ratio of
e1 = 0.50% and e1 = 1.00% for devices of d1 = 300 mm and 25 mm e2 = 1.67% provided the optimum damping for a maximum capture
respectively with a constant chamber length b1 = 300 mm. Another width ratio of the rear chamber. However, applying asymmetrical
example is for T = 1.4 s where all devices tested showed an optimum damping (e1 = 1.00% and e2 = 1.67%) that was separately identified
opening ratio of e1 = 1.00%, except the device with d1 = 300 mm, based on the optimum damping for each of the front and rear chambers
b1 = 300 mm that required a larger opening ratio of e1 = 1.67%. to the dual–chamber system (i.e., for D6 use e1 = 1.00% for the front
Overall, medium PTO damping values (i.e., medium opening ratio) chamber based on the results shown in Fig. 8a at e1 = e2 = 1.00% and

89
A. Elhanafi et al. Applied Energy 228 (2018) 82–96

Fig. 8. PTO damping effect on the capture width ratio (ζ) of Dual–Chamber OWC devices with an intermediate lip draught d2 = 300 mm and chamber length (Left):
b1 = 300 mm, b2 = 600 mm and (Right): b1 = 600 mm, b2 = 600 mm.

use e2 = 1.67% for the rear chamber based on the results shown in Following the investigation of the effect of chamber geometry and
Fig. 8b at e1 = e2 = 1.67%) decreased the capture width ratio of both the PTO damping of single–chamber and dual–chamber OWC devices, it
chambers. This indicates that the optimum damping of a dual–chamber is important to compare the best results achieved for both systems to
OWC device should take into account the interactions between both better understand the advantage of the dual–chamber system. Among
chambers, and simply selecting the optimum damping of each chamber the different tested PTO damping conditions, the opening ratio of
in isolated mode (a single–chamber device) might not provide the best e1 = e2 = 1.00% was selected in the comparison shown in Fig. 9 due to
performance of the dual–chamber system. the better performance achieved with this damping for both single–-
As mentioned earlier, the overall performance of the dual–chamber chamber and dual–chamber devices considered in this study. It is clear
OWC device was found to follow the same trends of the front chamber, that the use of the dual–chamber system is more efficient than the
and this observation holds true for all PTO damping values tested in this single–chamber device for the different chamber geometries tested. In
study as shown in Fig. 8c. From this figure, it is clear that a symmetrical addition, the improvement in the capture width ratio was mainly
PTO damping of e1 = e2 = 1.00% can be considered as the optimum achieved for the intermediate and long wave periods due to the rear
value for the whole wave period range tested, whereas the case of the chamber contribution over this wave period range. The maximum
asymmetrical damping showed a lower device performance for all wave (achieved for the longest period tested) and average (for all wave
periods. It is important to note that the lower performance achieved by periods tested) improvements in the capture width ratio were 138% and
the asymmetrical PTO damping does not mean that the asymmetrical 47% when using a dual–chamber D6 instead of a single chamber S2;
PTO damping should not be used, but this highlights the dependency of 70% and 24% when using a dual–chamber D7 instead of a single
each chamber performance not only on the applied damping in that chamber S4.
chamber, but also the damping induced on the other chamber.

90
A. Elhanafi et al. Applied Energy 228 (2018) 82–96

Fig. 9. Capture width ratio (ζ) of Single–Chamber vs. Dual–Chamber OWC

devices with symmetrical PTO damping induced by slot opening ratios of
e1 = e2 = 1.00%

3.4. 3D effect
Fig. 10. 3D effects on the capture width ratio (ζ) of Dual–Chamber (D6) and
Single–Chamber (S2, S3) OWC devices with a symmetrical PTO damping in-
2D simulations can provide a good insight into device performance, duced by orifice opening ratios of e1 = e2 = 1.00%. (Top): 3D effects on the
but to accurately model and predict the performance of offshore OWC front and rear chambers of D6 and (Bottom): 3D effects on Dual–Chamber vs.
devices, it is important to consider the 3D effects [18]. However, these Single–Chamber devices.
merits need a higher computation resources and time. In average, the
computation time required for a 3D case was found to be about 3.25
ratio of S2 and the front chamber of D6 where the short front lip
times that required for a 2D model using the same computational re-
draught (25 mm or H/2) and the 2D constraint concentrate the incident
sources (28 CPUs). Therefore, in this section only a few cases were
energy in front of the device allowing more wave energy to be extracted
selected to investigate the 3D effects on the capture width ratio of both
rather than reflected or scattered. On the other hand, devices/chambers
single–chamber and dual–chamber OWC devices. The dual–chamber
with a deep front lip draught (300 mm or H/6) such as S3 and the rear
OWC device D6 with a symmetrical PTO damping of e1 = e2 = 1.00%
chamber of D6 tend to reflect most of the incoming wave energy in the
was selected for its high performance achieved with the 2D model. In
case of 2D modelling which deducts the available energy to be ex-
addition, two single–chamber OWC devices S2 and S3 with geometries
tracted by the device [4]. However, the 3D modelling of these devices
similar to the front and rear chambers of the dual–chamber device D6
allow the incoming wave energy to be transmitted underneath and
respectively were considered. It is important to note that the PTO
around the device. This allows wave energy to enter the chamber not
damping was simulated using a slot opening for the 2D simulations,
only from underneath the front lip but also from underneath the side
while a circular orifice was used for the 3D simulations. It is crucial to
plating, especially for the intermediate– and long–wave regimes, which
ensure that any impact the 3D modelling might have on device per-
in turn increases the device capability to extract more energy. The
formance is not due to the different ways used in modelling the PTO
time–series of chamber free surface oscillation (ηOWC), airflow rate per
damping, but due to the side wave scattering of the 3D model. There-
unit chamber width (Q), chamber differential air pressure (ΔP ) and
fore, a comparison was performed (see Appendix B) for the results of a
extracted instantaneous pneumatic power per unit chamber width (PE)
dual–chamber OWC device (D6) with e1 = e2 = 1.00% using a 2D
shown in Fig. 11 for single–chamber devices S2 and S3 and Fig. 12 for a
model with a slot opening and a model extending to 100 mm in the
dual–chamber device D6 support this explanation.
y–direction of a NWT with a half width equal to 100 mm (i.e., the OWC
3D effects show more advantages for the rear chamber of D6 than
device occupies the full width of the flume, which is referred here as
the single–chamber device S3; see results of the rear–chamber of D6 in
2.5D) with a symmetry plane and a circular office. The Maximum
Fig. 10a and results of S3 in Fig. 10b. This is because in the case of a
Normalized Root Mean Square Error (NRMSE) of the main variables (air
dual–chamber system, the reduction in the energy extracted by the
pressure and air flowrate) measured between the 2D and 2.5D simu-
front–chamber (see Fig. 12a) can be compensated by the larger energy
lations were about 1.2%, which demonstrate the validity of the 2D
extracted by the rear–chamber (see Fig. 12b), allowing the dual–-
results presented in this paper and clarifies any doubt about the effects
chamber system to extract more wave energy specially for the inter-
of PTO modelling on the 3D results.
mediate– and long–wave regimes (see 2D and 3D results of D6 in
The results in Fig. 10 show that the 3D modelling of offshore OWC
Fig. 10b) where the rear–chamber is more effective. It is also worth
devices might have different impact on device performance such that
noting that there was a shift in the resonant period between 2D and 3D
the capture width ratio of the front chamber of D6 and the single–-
results. This can be explained as discussed above that 3D modelling
chamber device S2 decreased, whereas there was a significant increase
allows water waves to enter the chamber from all around the device.
in the capture width ratio of both the rear chamber of D6 and the
Therefore, it is not only the front lip draught and chamber length (in
single–chamber device S3.
wave propagation direction) that identify the resonant period but also
Elhanafi et al. [18] highlighted that 3D effects could result in a
the side plating draught and chamber breadth, which are 300 mm and
reduction in the capture width ratio of offshore OWC devices, especially
200 mm respectively in this study. For example, it was shown in Section
for high–frequency waves (short wave periods) where energy scattering
3.1 that a 2D single–chamber OWC device with d1 = 300 mm and
is more pronounced. This can explain the drop in the capture width

91
A. Elhanafi et al. Applied Energy 228 (2018) 82–96

Fig. 11. 3D effect on the time–series results of Single–Chamber (S2, S3) OWC devices with an opening ratio of e1 = 1.00%. (a) 3D effect on a Single–Chamber device
(S2) at T = 1.2 s and (b) 3D effect on a Single–Chamber device (S3) at T = 1.4 s.

b1 = 600 mm (S3) has a resonant period of about 1.58 s; however, simulations, the findings confirm that notably more energy can be ex-
considering the water column area of the 3D device tracted from a well–designed dual–chamber OWC device than the
Swc = 0.6 m × 0.2 m results in a resonant period of 1.33 s which is in conventional single–chamber devices. Detailed conclusions are given as
line with the changes in the resonant period shown in Fig. 10b for the follows.
single–chamber OWC device S3. The front lip draught should be minimised as much as practical,
whereas chamber length needs to be optimised to tune the single–-
4. Conclusions chamber device to the target wave conditions. The PTO damping could
also be adjusted to tune the device to a particular wave regime; how-
Experimentally validated 2D and 3D CFD models were utilized in ever, with a front lip draught equal to the incoming wave amplitude
this study to gain a deeper insight into the effect of underwater geo- there was an optimum PTO damping value (opening ratio of 1.00%)
metrical parameters (such as lip draught and chamber length) as well as that maximized the capture width ratio for most of the wave period
the pneumatic damping induced by air turbines on the performance range tested in this study.
(capture width ratio) of single–chamber and dual–chamber OWC de- For a dual–chamber OWC device, the optimized single–chamber
vices. Device performance was assessed over a range of eleven regular OWC device can be used as a front chamber of the dual system, whereas
wave periods for a constant wave height. From a total of 572 CFD the length of the rear chamber needs to be increased to allow better

Fig. 12. 3D effect on the time–series results of a Dual–Chamber (D6) OWC device with symmetrical opening ratios of e1 = e2 = 1.00% at a wave period T = 1.4 s.

92
A. Elhanafi et al. Applied Energy 228 (2018) 82–96

extraction of energy contained in long waves, which can easily be 3D effects showed a significant impact on device performance
transmitted into the rear chamber. The effect of the PTO damping on where the 2D predicted capture width ratio decreased and increased for
the performance of each chamber of the dual system was found to be so the front and rear chamber of the dual system respectively. This further
far similar to that on each chamber separately in the isolated mode highlights the importance of the rear chamber in dual–chamber systems
(modelled as a single–chamber OWC device). A symmetric PTO to overcome the deducted power extracted by the front chamber due to
damping for both the front and rear chambers was found to provide 3D effects, especially for intermediate– and long–wave regimes. The
better results than asymmetric damping values. With a PTO damping of relevance of 3D PTO damping to device performance still needs to be
1.00% for both chambers, the dual–chamber OWC device improved the investigated in detail in future works. In addition, it is important to note
capture width ratio by an average of about 47% and a maximum of that the dual–chamber OWC system uses two turbines; one for each
about 138% compared to the single–chamber OWC device. The im- chamber, which in turn increases the cost of the device. Therefore,
provement in the capture width ratio achieved by the dual–chamber further studies where both chambers share the same turbine is crucial
OWC system over the single–chamber in a wider range of wave periods along with a feasibility study to draw an overall conclusion on the
is promising for increased device performance in more realistic irre- cost–effectiveness of the dual–chamber OWC system and the applic-
gular waves, which deserves further investigations. ability of integrating this system into breakwaters.

Appendix A. Mesh convergence study

Mesh convergence studies for a dual–chamber OWC device (D6) were performed with different refinement levels of the cell size on the OWC
device and the PTO surfaces as summarized in Tables A1 and A2 for the 2D and 3D cases, respectively. The Normalized Root Mean Square Error
(NRMSE) calculated by Eq. (A1) was also included in each table for a single wave condition of height H = 50 mm and period T = 1.2 s with a PTO
opening ratio of e1 = e2 = 1.00%.

Table A1
Mesh study for a 2D Dual–Chamber OWC device D6 with e1 = e2 = 1.00% at T = 1.2 s.
Mesh Cell size [mm] Number of Parameter NRMSE [%]
cells
OWC PTO Front Rear
Chamber Chamber

Effect of OWC cell size

M-A 12.5 0.78125 126,445 ΔP 12.50 4.02
Q 10.49 0.75
M-B 6.25 0.78125 139,590 ΔP 0.42 1.84
Q 0.18 0.27
M-C 3.125 0.78125 175,047 ΔP 0.0 0.0
Q 0.0 0.0
Effect of PTO cell size
M-B 6.25 0.78125 139,590 ΔP 1.69 1.98
Q 0.71 0.86
M-D 6.25 1.5625 128,481 ΔP 2.33 6.58
Q 24.90 7.83
M-E 6.25 0.390625 178,855 ΔP 0.0 0.0
Q 0.0 0.0

Table A2
Mesh study for a 3D Dual–Chamber OWC device D6 with e1 = e2 = 1.00% at T = 1.2 s.
Mesh Cell size [mm] Number of Parameter NRMSE [%]
cells
OWC PTO Front Rear
Chamber Chamber

Effect of OWC cell size

M-A 12.5 0.78125 595,528 ΔP 8.36 5.8
Q 4.30 2.54
M-B 6.25 0.78125 870,682 ΔP 1.60 1.98
Q 1.12 0.83
M-C 3.125 0.78125 1,649,005 ΔP 0.0 0.0
Q 0.0 0.0

Effect of PTO cell size

M-B 6.25 0.78125 870,682 ΔP 1.98 1.28
Q 1.47 1.34
M-D 6.25 1.5625 828,565 ΔP 2.89 2.03
Q 1.85 2.72
M-E 6.25 0.390625 983,437 ΔP 0.0 0.0
Q 0.0 0.0

93
A. Elhanafi et al. Applied Energy 228 (2018) 82–96

1 1 N
NRMSE =
x max −x min N
∑i =1 (xi−yi )2 (A1)
where x i , yi , are the data from the most refined mesh and the other meshes, respectively, N is the number of data point (here, data from five wave
cycles are considered), x max , x min are the maximum and minimum measured values from the most refined mesh, accordingly.
From the mesh study results shown in Tables A1 and A2 and Figs. A1 and A2, mesh M-B was selected with a maximum NRMSE of less than 2% for
both 2D and 3D simulations.

Fig. A1. Mesh study results for a 2D Dual–Chamber OWC device D6 with e1 = e2 = 1.00% at T = 1.2 s.

Fig. A2. Mesh study results for a 3D Dual–Chamber OWC device D6 with e1 = e2 = 1.00% at T = 1.2 s.

94
A. Elhanafi et al. Applied Energy 228 (2018) 82–96

Appendix B. Effect of PTO modelling

It is common to use a slot opening and a circular orifice to model a nonlinear PTO system in 2D and 3D simulations. However, it is necessary to
ensure that both ways do not have much difference when assessing device performance. For a dual–chamber OWC device (D6) subjected to regular
wave of period T = 1.2 s with PTO opening ratios e1 = e2 = 1.00%, the CFD results of chamber differential air pressure (ΔP ), air flowrate (Q ) and
extracted instantaneous pneumatic power (PE ) are compared in Fig. B1 when using a 2D slot versus a 3D orifice. For the case of the 3D orifice, a
narrow numerical wave flume of a half width equal to device half width of 100 mm was used, which is referred in Fig. B1 as 2.5D with a symmetry
plane. The results in Fig. B1 along with the calculated NRMSE of about 1.2% illustrate the very identical results from both ways of modelling the PTO
system.

Fig. B1. Effect of modelling the PTO system in 2D and 3D for a Dual–Chamber OWC device D6 with e1 = e2 = 1.00% at T = 1.2 s.

References [14] Vyzikas T, Deshoulières S, Giroux O, Barton M, Greaves D. Numerical study of fixed
Oscillating Water Column with RANS-type two-phase CFD model. Renew Energy
2017;102:294–305.
[1] Falcão AFO, Henriques JCC. Oscillating-water-column wave energy converters and [15] Luo Y, Nader J-R, Cooper P, Zhu S-P. Nonlinear 2D analysis of the efficiency of fixed
air turbines: a review. Renew Energy 2015;85:1391–424. Oscillating Water Column wave energy converters. Renew Energy 2014;64:255–65.
[2] Sarmento AJNA. Wave flume experiments on two-dimensional oscillating water [16] Iturrioz A, Guanche R, Lara JL, Vidal C, Losada IJ. Validation of OpenFOAM® for
column wave energy devices. Exp Fluids 1992;12:286–92. oscillating water column three-dimensional modeling. Ocean Eng
[3] Simonetti I, Cappietti L, Elsafti H, Oumeraci H. Optimization of the geometry and 2015;107:222–36.
the turbine induced damping for fixed detached and asymmetric OWC devices: a [17] López I, Pereiras B, Castro F, Iglesias G. Optimisation of turbine-induced damping
numerical study. Energy 2017;139:1197–209. for an OWC wave energy converter using a RANS–VOF numerical model. Appl
[4] Elhanafi A, Fleming A, Macfarlane G, Leong Z. Underwater geometrical impact on Energy 2014;127:105–14.
the hydrodynamic performance of an offshore oscillating water column–wave en- [18] Elhanafi A, Macfarlane G, Fleming A, Leong Z. Investigations on 3D effects and
ergy converter. Renew Energy 2017;105:209–31. correlation between wave height and lip submergence of an offshore stationary
[5] Ning D-Z, Wang R-Q, Zou Q-P, Teng B. An experimental investigation of hydro- OWC wave energy converter. Appl Ocean Res 2017;64:203–16.
dynamics of a fixed OWC Wave Energy Converter. Appl Energy 2016;168:636–48. [19] Elhanafi A, Macfarlane G, Fleming A, Leong Z. Scaling and air compressibility ef-
[6] Bouali B, Larbi S. Sequential optimization and performance prediction of an oscil- fects on a three-dimensional offshore stationary OWC wave energy converter. Appl
lating water column wave energy converter. Ocean Eng 2017;131:162–73. Energy 2017;189:1–20.
[7] Gomes MN, Nascimento CD, Bonafini BL, Santos ED, Isoldi LA, Rocha LAO. Two- [20] Simonetti I, Cappietti L, Elsafti H, Oumeraci H. Evaluation of air compressibility
dimensional geometric optimization of an oscillating water column converter in effects on the performance of fixed OWC wave energy converters using CFD mod-
laboratory scale. Therm Eng 2012;11:30–6. elling. Renew Energy 2018;119:741–53.
[8] Vyzikas T, Deshoulières S, Barton M, Giroux O, Greaves D, Simmonds D. [21] Elhanafi A, Macfarlane G, Fleming A, Leong Z. Experimental and numerical mea-
Experimental investigation of different geometries of fixed oscillating water column surements of wave forces on a 3D offshore stationary OWC wave energy converter.
devices. Renew Energy 2017;104:248–58. Ocean Eng 2017;144:98–117.
[9] Wang R, Ning D, Zhang C. Numerical investigation of the hydrodynamic perfor- [22] Elhanafi A, Macfarlane G, Fleming A, Leong Z. Experimental and numerical in-
mance of the dual-chamber oscillating water columns. In: Proceedings of 27th in- vestigations on the intact and damage survivability of a floating–moored oscillating
ternational ocean and polar engineering conference, international society of off- water column device. Appl Ocean Res 2017;68:276–92.
shore and polar engineers (ISOPE), June 25–30, San Francisco, California, USA; [23] Crespo AJC, Altomare C, Domínguez JM, González-Cao J, Gómez-Gesteira M.
2017. Towards simulating floating offshore oscillating water column converters with
[10] Ning D, Wang R, Zhang C. Numerical simulation of a dual-chamber oscillating Smoothed Particle Hydrodynamics. Coastal Eng 2017;126:11–26.
water column wave energy converter. Sustainability 2017;9:1599. [24] Elhanafi A, Kim CJ. Experimental and numerical investigation on wave height and
[11] He F, Leng J, Zhao X. An experimental investigation into the wave power extraction power take–off damping effects on the hydrodynamic performance of an off-
of a floating box-type breakwater with dual pneumatic chambers. Appl Ocean Res shore–stationary OWC wave energy converter. Renew Energy 2018;125:518–28.
2017;67:21–30. [25] Elhanafi A, Macfarlane G, Fleming A, Leong Z. Experimental and numerical in-
[12] Elhanafi A, Fleming A, Macfarlane G, Leong Z. Numerical energy balance analysis vestigations on the hydrodynamic performance of a floating–moored oscillating
for an onshore oscillating water column–wave energy converter. Energy water column wave energy converter. Appl Energy 2017;205:369–90.
2016;116:539–57. [26] Rezanejad K, Bhattacharjee J, Soares CG. Stepped sea bottom effects on the effi-
[13] Simonetti I, Cappietti L. A CFD study on the balance of energy in a fixed bottom- ciency of nearshore oscillating water column device. Ocean Eng 2013;70:25–38.
detached oscillating water column wave energy converter. In: Proceedings of 12 th [27] Rezanejad K, Bhattacharjee J, Soares CG. Analytical and numerical study of dual-
European wave and tidal energy conference (EWTEC), 27 August–September 01, chamber oscillating water columns on stepped bottom. Renew Energy
Cork, Ireland; 2017. 2015;75:272–82.

95
A. Elhanafi et al. Applied Energy 228 (2018) 82–96

[28] Hirt CW, Nichols BD. Volume of fluid (VOF) method for the dynamics of free Applications, ITTC Report: 7.5-03 02-03; 2011.
boundaries. J Comput Phys 1981;39:201–25. [34] Veer RV, Tholen HJ. Added resistance of moonpools in calm water. In: Proceedings
[29] CD-Adapco. User Guide STAR-CCM+ Version 10.02; 2015. of 27th international conference on offshore mechanics and arctic engineering
[30] Choi J, Yoon SB. Numerical simulations using momentum source wave-maker ap- (OMAE-57246), American Society of Mechanical Engineers (ASME), June 15–20,
plied to RANS equation model. Coast Eng 2009;56:1043–60. Estoril, Portugal; 2008.
[31] Elhanafi A, Fleming A, MacFarlane G, Leong Z. Numerical hydrodynamic analysis of [35] Ning D-Z, Shi J, Zou Q-P, Teng B. Investigation of hydrodynamic performance of an
an offshore stationary–floating oscillating water column–wave energy converter OWC (oscillating water column) wave energy device using a fully nonlinear HOBEM
using CFD. Int J Nav Archit Ocean Eng 2017;9:77–99. (higher-order boundary element method). Energy 2015;83:177–88.
[32] Chakrabarti SK. Offshore structure modeling. Singapore: World Scientific [36] He F, Huang Z. Hydrodynamic performance of pile-supported OWC-type structures
Publishing; 1994. as breakwaters: an experimental study. Ocean Eng 2014;88:618–26.
[33] ITTC. Recommended Procedures and Guidelines: Practical Guidelines for Ship CFD

96