Ocean Engineering 51 (2012) 16–27

Contents lists available at SciVerse ScienceDirect

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

Hydrodynamic performance of a rectangular floating breakwater with and

without pneumatic chambers: An experimental study
Fang He a, Zhenhua Huang a,n, Adrian Wing-Keung Law a,b
a
School of Civil and Environmental Engineering, Nanyang Technological University, Singapore 639798, Singapore
b
DHI-NTU Center, Nanyang Technological University, Singapore 639798, Singapore

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

Article history: In this study, we investigated experimentally the hydrodynamic performance of floating breakwaters
Received 18 October 2011 with and without pneumatic chambers. The experimental results showed that the pneumatic chambers
Accepted 12 May 2012 significantly enhanced the wave energy dissipation as well as reduced the wave transmission. With the
Editor-in-Chief: A.I. Incecik
installation of the pneumatic chambers, the water inside the chambers helped to reduce the surge
Available online 5 June 2012
response, while the chamber walls increased the moment of inertia of the breakwater and thus
Keywords: mitigated the pitch response. The air pressure fluctuations inside the pneumatic chambers and the
Floating breakwater effects of draught were also examined in the experiments. The overall results illustrate that the new
Pneumatic chamber design of floating breakwater is more effective for wave protection, and also holds the potential for
Motion response
simultaneous wave energy conversion into electricity by installing Wells turbines to the pneumatic
Oscillating water column
chambers.
Wave energy conversion
& 2012 Elsevier Ltd. All rights reserved.

1. Introduction The general desirable characteristics of a floating breakwater

include high cost-effectiveness, good wave attenuation and
Floating breakwaters are commonly used to protect shorelines, low force requirements on the mooring system. Previously,
marine structures, moored vessels, marinas and harbors from Hales (1981) and McCartney (1985) reviewed comprehensively
wave attacks. Compared with permanently fixed breakwaters, various floating breakwater concepts to evaluate their perfor-
floating breakwaters have superiority in terms of environmental mance and applicability. Since then, floating breakwaters with
friendliness, low cost, flexibility and mobility. They are especially other novel configurations have also been proposed for better
competitive for coastal areas with a high tidal range or deep performance, such as the double Y-frame multifunctional float-
water depth. Moreover, they may even be the only viable option ing breakwater (Murali and Mani, 1997), the spar buoy floating
for locations with poor bottom foundation. breakwater fences (Liang et al., 2004), the P shaped floating
Different from bottom-fixed breakwaters such as rubble breakwater with two additional side-boards (Gesraha, 2006), the
mound breakwaters which intercept all approaching waves, the thin plane board floating breakwater with rows of net under-
hydrodynamic interactions between the incoming waves and the neath (Dong et al., 2008), the horizontally interlaced floating
floating breakwater are complex with the wave energy being pipe breakwater with multi-layers (Hegde et al., 2008), the
partially reflected, partially transmitted beneath the breakwater diamond-shape blocks assembled porous floating breakwater
and partially dissipated. Meanwhile, the incident waves excite the (Wang and Sun, 2010), and the floating breakwater with truss
motion responses of the breakwater, which in turn acts as a wave structures (Uzaki et al., 2011).
generator radiating waves away from the breakwater to both its Hales (1981) pointed out that a floating breakwater should be
seaward and leeward sides. Thus, the total transmitted waves as simple, durable and maintenance-free as possible for long-time
include two components: the transmitted incident waves passing operation in real seas, and that highly complex structures should
underneath and the radiated waves propagating to the leeward be avoided. Floating breakwaters with a rectangular cross-section
side of the breakwater. The wave transmission characteristic is an may thus be most suitable to satisfy these requirements. For
important consideration of the functional role of a breakwater various two-dimensional free or moored floating breakwaters
towards the objective of wave protection. with a rectangular cross-section, extensive theoretical, experi-
mental and numerical research had been reported in the litera-
ture. (Christian, 2000; Drimer et al., 1992; Fugazza and Natale,
n
Corresponding author. Tel.: þ65 6790 4737. 1988; Koutandos et al., 2005; Rahman et al., 2006; Sannasiraj
E-mail address: [email protected] (Z. Huang). et al., 1998; Williams et al., 2000).

0029-8018/$ - see front matter & 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.oceaneng.2012.05.008
 F. He et al. / Ocean Engineering 51 (2012) 16–27 17

In this study, we propose a novel configuration of a pneumatic examined. As far as we are aware, there are no published studies so
floating breakwater for combined wave protection and potential far that examine the pneumatic effects on the hydrodynamic
wave energy capturing. The development of the concept origi- performance of a floating breakwater, including wave reflection,
nates from the oscillating water column (OWC) device commonly transmission, energy dissipation and motion responses.
used in wave energy utilization (Falca~ o, 2010). The configuration One of the aims of the present study is to provide an
consists of the box-type breakwater with a rectangular cross- economical way to improve the performance of box-type floating
section as the base structure, with pneumatic chambers (OWC breakwaters for long waves without significantly increasing its
units) installed on both the front and back sides of the box-type weight and construction cost. In this study, the hydrodynamic
breakwater without modifying the geometry of the original base performance of the proposed floating breakwater under regular
structure. The pneumatic chamber used in this study is primarily waves (monochromatic waves) was investigated experimentally.
of a hollow chamber with a large submerged bottom opening The performance was compared with that of the original box-type
below the water level. Air trapped above the water surface inside floating breakwater without pneumatic chambers, including wave
the chamber is pressured due to the water column oscillation transmission, wave energy dissipation and motion responses, to
inside the chamber, and it can exit the chamber through a small elucidate the functional effects of the pneumatic chambers. Since
opening at the top cover with energy dissipation. Since the energy the dynamic characteristics of the floating structure change with
dissipated by the air flow is not directly related to both the different draughts, three different draughts were tested in the
reflected and transmitted waves, thus better wave attenuation experiments to investigate possible influences of draught. The air
can be potentially achieved with the chamber installation. pressure fluctuations inside the chambers, which inferred the
To be an effective floating breakwater, its movements should extent of water column oscillation, were also measured.
be of small amplitude so that the motion-generated radiated
waves into the protected region will not be large. To achieve this,
either a tensioned mooring system (Elchahal et al., 2008; Rahman 2. Experimental setup and test procedures
et al., 2006; Wang and Sun, 2010; Williams and Abul-Azm, 1997)
or vertical piles (Diamantoulaki et al., 2008; Isaacson et al., 1998; 2.1. Physical model
Kim et al., 1994) were proposed to restrain the motion of the
breakwater in earlier studies. A taut mooring system can effec- The geometric details of the pneumatic floating breakwater
tively restrict the motion amplitude, while piles can effectively and the original rectangular box-type breakwater model used in
restrict the horizontal motion but not the vertical motion. In the experiments are shown in Fig. 1. Pneumatic chambers were
practice, the taut mooring system faces such problems as huge attached to the front and back sides of the original rectangular
impulsive forces, high sensitivity to tidal change and construction box-type structure (shown on the right of Fig. 1) to form a new
difficulties, while the pile-restrained implementation also meets configuration (shown on the left of Fig. 1). The models were made
many problems including large loads on the piles, abrasion of 10-mm thick Perspex sheet; additional steel and Perspex plates
between piles and the breakwater, and infeasibility in deep water were placed inside the breakwater as ballasts to adjust the
or poor foundation conditions where the floating breakwater draught. A narrow slot at the top plate was made as the outlet
should have been competitive (McCartney, 1985). Our proposed for each pneumatic chamber to allow an energy loss induced by
configuration mitigates the movements of the breakwater the air passage in and out of the chamber. For convenience of
through the effects of pneumatic chambers instead. Hence, it description, we designate the proposed breakwater of 0.235 m
should be more economical, and a feasible slack mooring system draught as Model 1, the original rectangular box-type breakwater
can be employed with the reduced motion. without chambers of 0.235 m draught as Model 2, the proposed
There have been few earlier studies on floating breakwaters with breakwater of 0.299 m draught as Model 3 and the proposed
pneumatic effects. Vijayakrishna Rapaka et al. (2004) experimentally breakwater of 0.177 m draught as Model 4. The details of these
studied a floating multi-resonant structure of which the OWC-type four models are summarized in Table 1, and a view of the physical
wave energy devices were embedded into the middle of a floating model in the wave flume is shown in Fig. 2.
breakwater. The dynamic behaviors of the structure were studied
including the motion responses and mooring line forces. Koo (2009) 2.2. Experimental setup
developed a nonlinear numerical wave tank to study the pneumatic
floating breakwater in one individual mode. The effects of pneu- The experiments were conducted in a wave flume at the Hydrau-
matic damping on the body motion and wave transmission were lics Modeling Laboratory of Nanyang Technological University,

Fig. 1. Details of the pneumatic floating breakwater and original rectangular box-type breakwater models.
18 F. He et al. / Ocean Engineering 51 (2012) 16–27

Table 1
Details of the four models examined in the experiments.

Length (mm) Bottom Height (mm) Draught (mm) Chamber Slot opening Mass (Kg) Moment Gravity Center
breadth (mm) breadth (mm) breadth (mm) of inertia (Kg m2) above base (mm)

Model 1 1420 750 400 235 400 5 267 35.7 111.9

Model 2 1420 750 400 235 – – 250 14.4 79.6
Model 3 1420 750 400 299 400 5 339 39.3 95.7
Model 4 1420 750 400 177 400 5 195 31.8 143.5

The target wave height H was fixed at 0.04 m. Since the coastal
mean water level changes with tides, four water depths h were
examined in the present study: 0.90 m, 0.70 m, 0.55 m and
0.45 m. Since the main focus of this study is to improve the
performance of existing box-type breakwaters by installing the
pneumatic chambers, it is natural to compare the modified model
with the original model using a length scale that is common for
both models to normalize the wave length L and to present the
results. In this study, we chose B, the bottom breadth of the
original breakwater, to compare our experimental results. The
normalized B/L varied from 0.18 to 0.45. All the test wave
conditions are listed in Table 2. The target geometry scale factor
in this study is 1:15. Thus, the scale for wave height was 1:15 and
the scale for wave period was 1:3.87. The prototype wave period
ranges from 4.26 s to 7.74 s, which is within the bulk range of
wave periods for wind-generated waves.

Fig. 2. Physical model in the wave flume before running waves.

2.3. Data acquisition system

Eight HR Wallingford wave gauges (WG1–WG8 in Fig. 3) were

Singapore. The dimensions of the wave flume were 45-m long, used to measure the surface elevations; four were placed in front
1.55-m wide and 1.5-m deep. A piston-type wave-maker, equipped of the model for separation of the incident waves from the
with a DHI Active Wave Absorption Control System (AWACS), was reflected waves, and the other four in the leeward side of the
installed at one end of the flume, and a wave-absorbing beach was model for separation of the transmitted waves from the waves
located at the other end to reduce the wave reflection. reflected from the wave absorbing beach. The distances between
Fig. 3 shows a sketch of the experimental setup. The floating the wave gauges are listed in Table 3. The manufacturer-specified
breakwater was slack-moored in its equilibrium position, which was accuracy of the wave gauge is 0.1 mm. Before and after each set of
25 m away from the wave-maker. Each chain mooring line was tests, pre- and post-calibration were carried out to ensure the
fastened to a concrete anchor (shown in Fig. 4). The mooring line quality of the measured surface elevation. The two-point method,
was made of stainless steel and had a length of 3.0 m with a line proposed initially by Goda and Suzuki (1976), was employed to
density of 0.155 kg/m. The concrete anchor had an average weight separate the reflected waves from the incident waves. Different
of 2.265 kg and its small dimensions (0.1 m  0.1 m  0.1 m) would distances between the four wave gauges provide several sets of
not significantly disturb the flow field. Three sets of mooring cables data available for wave separation.
were installed on each side of the floating breakwater. Vijayakrishna An optical tracking system was installed to capture the motion
Rapaka et al. (2004) studied the effects of slack mooring-line scope of the floating breakwater. The system consisted of two ProReflex
(defined as the ratio of length of mooring line to water depth) on the infrared cameras, data acquisition and processing software (Qua-
motion responses of floating structures. They found that the motions lisys Track Manager) and retro-reflective markers. An earth-fixed
in all surge, heave and pitch modes depicted similar behaviors and Cartesian coordinate system can be established through calibra-
the difference was minor although the mooring-line scopes widely tion by using the standard calibration tools provided by the
varied from 4 to 6. Since the effects of slack mooring lines on the manufacturer. Since the established coordinate system is in
motion responses were insignificant, we did not change the posi- reference to the locations of the two cameras, the cameras cannot
tions of the anchors in our study (our mooring line scope varied be moved after calibration. The trajectory of retro-reflective
from 3.33 to 6.67). The positions of the small concrete anchors were markers attaching to the floating breakwater can be tracked in
checked after each test, and it was confirmed that those anchors the calibrated coordinate system by the two cameras, and
were not moved by the breakwaters in our experiments. The main Qualisys Track Manager can calculate the motion responses of
function of the mooring lines actually is to resist the slow drift force the floating breakwater after the center of rotation is specified. In
and hold the floating breakwater in its dynamic equilibrium posi- principle, the minimum number of the markers required for the
tion. Our tests showed that our relatively small concrete anchor was calculation is 3, but more markers can be used to ensure the data
strong enough to resist these mooring forces. quality (in case that any marker is out of the field of view when
As shown in Fig. 5, four ball bearings were installed on each the model moves). Sample temporal data of the measured surge,
lateral side of the model. The ball bearing can rotate in all heave and pitch responses are shown in Fig. A1 in the Appendix.
directions and reduce the friction between the model and the A Kistler pressure sensor was installed on the top of each
walls; they also prevent the model from any possible colliding pneumatic chamber (10 cm away from the slot opening)
with the flume walls. In this manner, the motion of the break- to measure the air pressures inside the chambers (PS1 and PS2
water can be restricted to two-dimensional only. in Fig. 3).
 F. He et al. / Ocean Engineering 51 (2012) 16–27 19

Fig. 3. Sketch of the experimental setup for the breakwater with pneumatic chambers.

Table 2
Experimental test conditions (H ¼0.04 m).

h (m) T (s) L (m) B/L

0.9 1.1–1.7 1.88–4.00 0.187–0.399

0.7 1.1–1.8 1.85–4.03 0.186–0.404
0.55 1.1–1.9 1.81–3.96 0.189–0.415
0.45 1.1–2.0 1.75–3.88 0.193–0.430

Table 3
Distances between wave gauges.

Wave gauges Distances (cm)

Fig. 4. A view of the chain mooring line and the concrete anchor.
WG1 and WG2 20
WG2 and WG3 40
WG3 and WG4 40
WG5 and WG6 40
WG6 and WG7 40
WG7 and WG8 20

The amplitudes of surge translation (Asurge), heave translation

(Aheave) and pitch rotation (Apitch) of the breakwater were
captured by the infrared camera system. We define the surge,
heave and pitch RAOs as Asurge/Ai, Aheave/Ai and Apitch/Ai, respec-
tively. Figs. 9–11 show the variations of the surge, heave and
Fig. 5. Ball bearing structure; the circles indicated the installation of the ball pitch RAOs with B/L, respectively.
bearings. Fig. 12 summarizes the variation with B/L of the normalized
pressure fluctuations (DP/rgHi with Hi being the height of the
3. Results and discussion incident waves) inside the front and back chambers for Model 1, 3
and 4.
The details of the four models examined in the experiments From the results, we found that the effect of water depth was
are listed in Table 1. Model 1 and 2 had the same draught of insignificant. The subsequent focus is thus given to the effects of
0.235 m to understand the effects of the pneumatic chambers. In the pneumatic chambers and the draught. However, all the results
addition, two other draughts, one deeper (0.299 m for Model 3) for the four water depths are presented for completeness.
and another shallower (0.177 m for Model 4) were also examined
to understand the effects of draught on pneumatic floating
3.1. The effects of pneumatic chambers
breakwaters.
The amplitudes of incident waves (Ai) and reflected waves (Ar)
The effects of the pneumatic chambers on the wave reflection
were separated from the measured surface elevations by using a
and transmission, wave energy dissipation and motion responses
two-point method (Goda and Suzuki, 1976). We also separated
are elucidated by comparing the hydrodynamic performances of
the transmitted waves (At) from the waves reflected from the
Model 1 and 2.
beach (Arb) to check the dissipation performance of the beach. For
floating breakwaters, we define the reflection coefficient Cr as
Ar/Ai and the transmission coefficient Ct as At/Ai. Figs. 6 and 7 3.1.1. Wave reflection and transmission coefficients
show the variations of Cr and Ct with B/L, respectively. We denote Referring to Fig. 6, the wave reflection of Model 1 was stronger
Cd as the fraction of incident wave energy dissipated, which can for relatively short period waves but weaker for longer period
be estimated by examining the wave energy balance as follows: waves. The minimum reflection coefficient occurred around
B/L¼0.23. In contrast, the reflection coefficient for Model 2 varied
C 2r þ C 2t þ C d ¼ 1 þC 2rb ð1Þ in a narrow range roughly between 0.2 and 0.5.
As shown in Fig. 7, the wave transmission coefficient was
where Crb ¼ Arb/Ai quantifies the wave energy reflected from the reduced in the whole range of B/L by installing the pneumatic
absorbing beach. Fig. 8 shows the variation of Cd with B/L. chambers. For Model 1, increasing B/L decreased Ct nearly
20 F. He et al. / Ocean Engineering 51 (2012) 16–27

1.0 1.0
0.9 0.9 90cm depth
0.8 0.8 70cm depth
55cm depth
0.7 0.7
45cm depth
0.6 0.6
Cr

Cr
0.5 0.5
0.4 0.4
90cm depth
0.3 0.3
70cm depth
0.2 0.2
55cm depth
0.1 45cm depth 0.1
0.0 0.0
0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.15 0.20 0.25 0.30 0.35 0.40 0.45
B/L B/L

1.0 1.0
0.9 0.9 90cm depth
0.8 0.8 70cm depth
0.7 0.7 55cm depth
45cm depth
0.6 0.6
Cr

Cr
0.5 0.5
0.4 0.4
90cm depth
0.3 0.3
70cm depth
0.2 55cm depth 0.2
0.1 45cm depth 0.1
0.0 0.0
0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.15 0.20 0.25 0.30 0.35 0.40 0.45
B/L B/L

Fig. 6. Variation of reflection coefficient Cr versus B/L under four water depths; (a) Model 1, with chambers, 0.235-m draught; (b) Model 2, without chambers, 0.235-m
draught; (c) Model 3, with chambers, 0.299-m draught; and (d) Model 4, with chambers, 0.177-m draught.

1.0 1.0
0.9 90cm depth 0.9
70cm depth
0.8 0.8
55cm depth
0.7 0.7
45cm depth
0.6 0.6
Ct

Ct

0.5 0.5
0.4 0.4
90cm depth
0.3 0.3
70cm depth
0.2 0.2
55cm depth
0.1 0.1 45cm depth
0.0 0.0
0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.15 0.20 0.25 0.30 0.35 0.40 0.45
B/L B/L

1.0 1.0
0.9 90cm depth 0.9 90cm depth
70cm depth 70cm depth
0.8 0.8
55cm depth 55cm depth
0.7 0.7
45cm depth 45cm depth
0.6 0.6
Ct

Ct

0.5 0.5
0.4 0.4
0.3 0.3
0.2 0.2
0.1 0.1
0.0 0.0
0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.15 0.20 0.25 0.30 0.35 0.40 0.45
B/L B/L

Fig. 7. Variation of transmission coefficient Ct versus B/L under four water depths; (a) Model 1, with chambers, 0.235-m draught; (b) Model 2, without chambers, 0.235-m
draught; (c) Model 3, with chambers, 0.299-m draught; and (d) Model 4, with chambers, 0.177-m draught.
 F. He et al. / Ocean Engineering 51 (2012) 16–27 21

1.0 1.0
0.9 0.9 90cm depth
0.8 0.8 70cm depth

0.7 0.7 55cm depth

45cm depth
0.6 0.6
Cd

0.5 0.5

Cd
0.4 0.4
90cm depth
0.3 0.3
70cm depth
0.2 55cm depth 0.2
0.1 45cm depth 0.1
0.0 0.0
0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.15 0.20 0.25 0.30 0.35 0.40 0.45
B/L B/L

1.0 1.0
0.9 0.9
0.8 0.8
0.7 0.7
0.6 0.6
Cd

Cd
0.5 0.5
0.4 0.4
90cm depth 90cm depth
0.3 0.3
70cm depth 70cm depth
0.2 55cm depth 0.2 55cm depth
0.1 45cm depth 0.1 45cm depth
0.0 0.0
0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.15 0.20 0.25 0.30 0.35 0.40 0.45
B/L B/L

Fig. 8. Variation of energy dissipation coefficient Cd versus B/L under four water depths; (a) Model 1, with chambers, 0.235-m draught; (b) Model 2, without chambers,
0.235-m draught; (c) Model 3, with chambers, 0.299-m draught; and (d) Model 4, with chambers, 0.177-m draught.

2.0 2.0
1.8 90cm depth 1.8 90cm depth
1.6 70cm depth 1.6 70cm depth
Surge RAO [m/m]

Surge RAO [m/m]

55cm depth 55cm depth

1.4 1.4
45cm depth 45cm depth
1.2 1.2
1.0 1.0
0.8 0.8
0.6 0.6
0.4 0.4
0.2 0.2
0.0 0.0
0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.15 0.20 0.25 0.30 0.35 0.40 0.45
B/L B/L

2.0 2.0
1.8 90cm depth 1.8 90cm depth
1.6 70cm depth 1.6 70cm depth
Surge RAO [m/m]

Surge RAO [m/m]

55cm depth 55cm depth

1.4 1.4
45cm depth 45cm depth
1.2 1.2
1.0 1.0
0.8 0.8
0.6 0.6
0.4 0.4
0.2 0.2
0.0 0.0
0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.15 0.20 0.25 0.30 0.35 0.40 0.45
B/L B/L

Fig. 9. Variation of surge RAOs versus B/L under four water depths; (a) Model 1, with chambers, 0.235-m draught; (b) Model 2, without chambers, 0.235-m draught;
(c) Model 3, with chambers, 0.299-m draught; and (d) Model 4, with chambers, 0.177-m draught.
22 F. He et al. / Ocean Engineering 51 (2012) 16–27

2.0 2.0
1.8 90cm depth 1.8 90cm depth

1.6 70cm depth 1.6 70cm depth

Heave RAO [m/m]

Heave RAO [m/m]

55cm depth 55cm depth
1.4 1.4
45cm depth 45cm depth
1.2 1.2
1.0 1.0
0.8 0.8
0.6 0.6
0.4 0.4
0.2 0.2
0.0 0.0
0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.15 0.20 0.25 0.30 0.35 0.40 0.45
B/L B/L

2.0 2.0
1.8 90cm depth 1.8 90cm depth
1.6 70cm depth 1.6 70cm depth
Heave RAO [m/m]

Heave RAO [m/m]

55cm depth 55cm depth
1.4 1.4
45cm depth 45cm depth
1.2 1.2
1.0 1.0
0.8 0.8
0.6 0.6
0.4 0.4
0.2 0.2
0.0 0.0
0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.15 0.20 0.25 0.30 0.35 0.40 0.45
B/L B/L

Fig. 10. Variation of heave RAOs versus B/L under four water depths; (a) Model 1, with chambers, 0.235-m draught; (b) Model 2, without chambers, 0.235-m draught;
(c) Model 3, with chambers, 0.299-m draught; and (d) Model 4, with chambers, 0.177-m draught.

monochromatically from a maximum value of 0.71 to a minimum modest than Model 2 in the whole range of B/L. This was because
value of 0.15. Note that this behavior is quite similar to that of a the water columns inside the chambers also moved back and
fixed box-type breakwater (see Fig. 3 in Drimer et al., 1992). In forth with the structure in surge mode, and accordingly increased
contrast, for Model 2, Ct reached a minimum value of 0.33 around the virtual mass of the breakwater. When B/L40.27, the surge
B/L¼0.29. In terms of maximum Ct, it was as large as 0.96 at RAOs of Model 1 were nearly constant at 0.2. From B/L¼ 0.27,
B/L¼0.19 for long period waves, and 0.63 at B/L¼0.42 for short decreasing B/L increased surge RAOs nearly monochromatically
period waves. Drimer et al. (1992) pointed out that the floating to a maximum value of 1.06. In contrast, the surge RAOs of
breakwater is transparent for very long waves. However, the Model 2 was nearly constant at 0.3 only when B/L40.35; the
additional pneumatic chambers changed the wave scattering and maximum value was up to 1.66, which is much stronger than that
energy dissipation. Our results showed that the breakwater with of Model 1.
pneumatic chambers could still be effective in reducing wave In Fig. 10, a comparison between the heave RAOs for Model
energy transmission even for very long period waves. 1 and 2 shows that the heave motion of Model 1 was less than
that of Model 2 in the whole range of B/L. The heave RAOs had
similar decreasing trends when B/L40.20 for both models, while
3.1.2. Wave energy dissipation
the maximum RAOs were 1.27 and 1.72 for Model 1 and 2,
Fig. 8 shows the calculated energy dissipation coefficient Cd.
respectively. The reason for the similarity in heave RAOs could be
Comparing the measured Cd for Model 1 and 2, it reveals that Model
attributed to the fact that the bottom shapes and masses (267 kg
1 dissipated much more energy for longer period waves when
for Model 1 and 250 kg for Model 2) were similar for both models.
B/Lo0.29. However, there was no noticeable difference in the energy
Despite the similarity, due to the presence of the pneumatic
dissipation for shorter period waves when B/L40.29. The major
chambers, the heave RAOs of Model 1 were somewhat lower than
benefit of using pneumatic chambers to dissipate wave energy is thus
that of Model 2.
primarily for long period waves: for the longest wave in the
Fig. 11 shows that the pitch motion of Model 1 was relatively
experiments, Cd for Model 2 was only 0.05, while Cd was as large as
smaller in the whole range of B/L. The reasons can be attributed to
0.51 for Model 1. The additional energy dissipation for Model 1 came
the larger moment of inertia of Model 1 (almost two and half
from the vortex shedding at the tips of the chamber front walls and
times of Model 2) and the effects of the pneumatic chambers. The
the air flow through the slot openings at the top of the chambers.
pitch RAOs of Model 1 had a decreasing trend from the maximum
value of 3.22 to the minimum value of 0.44 with increasing B/L. In
3.1.3. Motion responses contrast, the pitch RAOs of Model 2 were much higher for short
In Fig. 9, a comparison between the surge RAOs for Model and medium period waves (B/L40.24), with the maximum value
1 and 2 indicates that the surge motion of Model 1 was more being 6.92 at B/L¼ 0.32.
 F. He et al. / Ocean Engineering 51 (2012) 16–27 23

8 8
90cm depth 90cm depth
7 7
70cm depth 70cm depth
Pitch RAO [rad/m]

Pitch RAO [rad/m]

6 55cm depth 6 55cm depth
45cm depth 45cm depth
5 5
4 4
3 3
2 2
1 1
0 0
0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.15 0.20 0.25 0.30 0.35 0.40 0.45
B/L B/L

8 8
90cm depth 90cm depth
7 7
70cm depth 70cm depth
Pitch RAO [rad/m]

Pitch RAO [rad/m]

6 55cm depth 6 55cm depth
45cm depth 45cm depth
5 5
4 4
3 3
2 2
1 1
0 0
0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.15 0.20 0.25 0.30 0.35 0.40 0.45
B/L B/L

Fig. 11. Variation of pitch RAOs versus B/L under four water depths; (a) Model 1, with chambers, 0.235-m draught; (b) Model 2, without chambers, 0.235-m draught;
(c) Model 3, with chambers, 0.299-m draught; and (d) Model 4, with chambers, 0.177-m draught.

In summary, the results showed that the motion responses of Referring to Fig. 7, the measured transmission coefficient (Ct)
the floating breakwater were moderate with the installation of for Model 1, 3 and 4 decreased with B/L in a similar manner. The
pneumatic chambers. transmission coefficient of the model with a shallower draught
was relatively low for medium period waves (B/L varied approxi-
mately from 0.26 to 0.38). For long period waves (B/Lo0.26) and
3.2. The effects of draught
very short period waves (B/L40.38), a deeper draught was more
efficient in reducing the transmitted waves. The maximum
To illustrate the effects of draught, comparisons among the three
transmission coefficient for very long period waves reduced from
models with the same pneumatic chambers but different draughts
0.80 for Model 4 to 0.62 for Model 3. Comparatively, there was no
(Model 1, 3 and 4) were made, including wave reflection and
noticeable difference in the minimum transmission coefficient for
transmission, wave energy dissipation, motion responses and air
very short period waves, which decreased from 0.15 for Model
pressure fluctuations inside the chambers. The draught was adjusted
4 to 0.13 for Model 3.
by extra ballasts: the model with a deeper draught had a larger mass
and thus larger moment of inertia; the model dynamic character-
istics also changed with draught. Meanwhile, deepening the draught 3.2.2. Wave energy dissipation
increased the height of the water column inside the pneumatic In Fig. 8, a comparison of the energy dissipation coefficients
chamber and thus increased its natural period accordingly. (Cd) for Model 1, 3 and 4 indicates that Model 4 and 3 dissipated
the most and least wave energy, respectively, for both short and
medium period waves. For long period waves, however, the
3.2.1. Wave reflection and transmission coefficients
results were opposite. This was because that a deeper draught
Referring to Fig. 6, a comparison of the reflection coefficients
increased the height of the water column inside the chamber and
(Cr) for Model 1, 3 and 4 shows that the wave reflection was the
increased its natural period accordingly. The maximum energy
strongest for Model 3 and the weakest for Model 4 in the whole
dissipation of each model occurred at a different values of B/L: the
range of B/L. This is expected as deepening the draught reduces
peak values of 0.80 (Model 3), 0.87 (Model 1) and 0.90 (Model 4)
the wave transmission beneath the breakwater, and accordingly
occurred at B/L¼0.22, 0.24 and 0.28, respectively.
increases the wave reflection. The measured reflection coefficient
shows a similar variation with B/L for Model 1, 3 and 4. With the
dynamic characteristics of the breakwater model changing with 3.2.3. Motion responses
its weight, the minimum reflection coefficient occurred at slightly A comparison of the surge RAOs for Model 1, 3 and 4 given in
different values of B/L for different models. Fig. 9 shows that the surge motion was similar for the three
24 F. He et al. / Ocean Engineering 51 (2012) 16–27

0.30 0.30
90cm depth 90cm depth
0.25 70cm depth 0.25 70cm depth
55cm depth 55cm depth
0.20 45cm depth 0.20 45cm depth
ΔP/ρgHi

ΔP/ρgHi
0.15 0.15

0.10 0.10

0.05 0.05

0.00 0.00
0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.15 0.20 0.25 0.30 0.35 0.40 0.45
B/L B/L

0.30 0.30
90cm depth 90cm depth
0.25 70cm depth 0.25 70cm depth
55cm depth 55cm depth
0.20 45cm depth 0.20 45cm depth

ΔP/ρgHi
ΔP/ρgHi

0.15 0.15

0.10 0.10

0.05 0.05

0.00 0.00
0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.15 0.20 0.25 0.30 0.35 0.40 0.45
B/L B/L

0.30 0.30
90cm depth
0.25 0.25 70cm depth
55cm depth
0.20 0.20 45cm depth
ΔP/ρgHi

ΔP/ρgHi

0.15 0.15

0.10 90cm depth 0.10

70cm depth
0.05 55cm depth 0.05
45cm depth
0.00 0.00
0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.15 0.20 0.25 0.30 0.35 0.40 0.45
B/L B/L

Fig. 12. Variation of pressure fluctuations versus B/L under four water depths; (a) front chamber of Model 1, 0.235-m draught; (b) back chamber of Model 1, 0.235-m
draught; (c) front chamber of Model 3, 0.299-m draught; (d) back chamber of Model 3, 0.299-m draught; (e) front chamber of Model 4, 0.177-m draught; and (f) back
chamber of Model 4, 0.177-m draught.

models despite the different draughts. This is expected because the wave radiation due to the draught, which was primarily
the draught is proportional to the model mass, hence the water caused by the surge motion, was not significant.
resistance is also proportional to the model mass. When
B/L40.25 (Model 3), 0.27 (Model 1) and 0.34 (Model 4), the
surge RAOs were almost constant at 0.2. 3.2.4. Air pressure fluctuations inside the pneumatic chambers
In Fig. 10, the comparison of heave RAOs for the models shows Fig. 12 compares the pressure fluctuations inside the front
that deepening the draught reduced the heave motion slightly. chambers of Model 1, 3 and 4. When the draught is shallow, wave
With increasing the draught, the maximum values of heave RAO energy can be transmitted more easily through the lip of the front
decreased from 1.28 for Model 4 to 1.16 for Model 3. chamber, and stronger water column oscillations are induced.
A comparison of the pitch RAOs for Model 1, 3 and 4 is given As shown in Fig. 12, the model with shallower draughts had
in Fig. 11, where there was no noticeable difference in the larger pressure fluctuations in the whole range of B/L, and the
pitch motions with different draughts. The pitch RAOs of the peaks of pressure fluctuation were 0.127, 0.204 and 0.268 for
models with deeper draughts were slightly lower than that with Model 3, 1 and 4, respectively. Deepening the draught (increasing
shallower draughts. the weight) also increased the natural period of water column,
In general, deepening the draught reduced the surge, heave which caused the peaks of pressure fluctuation to occur at smaller
and pitch motions, but not very much. As a result, the change in values of B/L.
 F. He et al. / Ocean Engineering 51 (2012) 16–27 25

Compared with the front chamber, the air pressure fluctuation vortex shedding at the tips of the chamber front walls and the air
inside the back chamber was generally weaker, especially for a flow through the slot openings). The motion responses of Model
shallower draught. The difference in the pressure fluctuations 1 were in general smaller than those of Model 2. In particular, the
between the two chambers was significant for Model 4, but not so installation of the two pneumatic chambers significantly reduced
for Model 3. The pressure fluctuation inside a pneumatic chamber the surge motion for the long and medium period waves (B/Lo0.35)
was primarily caused by the relative motion between the floating and the pitch motion for the short and medium period waves (B/
breakwater and the water column oscillation inside the chamber. L40.24), while the heave motion slightly was reduced throughout
However, since the motion of floating breakwater was symmetric the whole range of B/L. The smaller motion responses of Model
about the transverse axis through the center of rotation, the 1 reduced the motion-generated radiated waves in the leeward side
difference in water column oscillation should be the main factor of the model. Therefore, the wave transmission was effectively
that caused the difference in pressure fluctuation. reduced for all wave periods.
For very short waves, the blockage of waves by the floating The addition of the two plates to form the pneumatic chambers
breakwater can be effective (Drimer et al., 1992), thus the wave should have also contributed to the wave scattering and the
energy cannot be transmitted easily beneath the breakwater to reduction of the transmission coefficient, as can be qualitatively
the back chamber. For waves of periods close to the natural period inferred from Kagemoto (2011), who studied theoretically the wave
of the water column, a significant portion of incoming wave transmission and reflection due to two vertical surface-piercing
energy was dissipated by the large oscillation of the water column plates fixed in regular waves. His results showed that the transmis-
inside the front chamber, so only a small portion of the incoming sion coefficient could be nearly zero if the ratio of spacing between
wave energy was transmitted to the back chamber. For very long two vertical plates to wave length satisfied certain conditions. Even
period waves, waves were easily transmitted through the floating though his model was fixed and there were no structural members
breakwater to the back chamber; however, since their periods between the two plates, his results still illustrated the importance of
differed significantly from the designed natural period of water wave scattering. If motion responses of the two plates are allowed,
column, both chambers did not function effectively. This explains they will inevitably change the performance of the breakwater
the observation that the pressure fluctuation inside the back studied by Kagemoto (2011). However, since there is no theory
chamber was typically weak. available currently for either the twin-plate floating breakwaters or
Finally, we note that the geometry of the two chambers was the box-type floating breakwaters with pneumatic chambers, it is
identical with the same designed natural period. Thus, strong difficult to quantify in this study the contributions of the vertical
water column oscillations inside the front and back chambers plates, which were used to form the pneumatic chambers, to the
could have been equally triggered by incoming waves with a transmission coefficients.
period close to the natural period. Therefore, it was the balance A deeper draught typically causes the breakwater to reflect more
between the energy dissipation (non-linear processes) and the wave energy, especially for short period waves. In the experiments,
energy input from the waves that determined the magnitude of when the draught was increased, lesser wave energy was transmitted
the air pressure fluctuations inside a chamber. through the lip of the front chamber, thus the energy dissipated
by the pneumatic chambers was reduced. The maximum energy
3.3. Discussion dissipation of Model 3 (with a larger draught) was lower than that of
Model 4 (with a smaller draught). Despite the different draughts, the
The significance of the pneumatic chambers can be examined maximum energy dissipation of the three models occurred around
directly by comparing the results for Model 1 (with pneumatic the wave period at which the maximum pressure fluctuation inside
chambers) and Model 2 (without pneumatic chambers). Since the the front chamber also occurred, suggesting that the energy dissipa-
wave reflection for Model 2 was not sensitive to the change of wave tion was caused mainly by the pneumatic chambers. Since the natural
period, the wave transmission was controlled mainly by the energy period of the pneumatic chamber increased with increasing draught,
dissipation. The energy dissipation of Model 2 can only be related to the pneumatic chambers of Model 3 and 4 functioned better for long
the frictional and flow-separation effects, which occurred mainly at and short/medium period waves, respectively. For the long and very
the sharp edges of the breakwater, so its large motion responses to short period waves, Model 3 (with larger draught) was the most
the incident waves caused the large energy dissipation, resulting in efficient in terms of reducing transmitted waves.
relatively small transmission coefficients: the maximum energy Vijayakrishna Rapaka et al. (2004) experimentally studied a
dissipation and minimum wave transmission occurred at a narrow rectangular floating breakwater with two OWC (Oscillating Water
range of wave periods corresponding to B/L¼0.29–0.33, which was Column) units embedded into its middle section. The surge and
also close to the range of wave period in which the breakwater had heave responses of our models were similar to theirs, but the pitch
its the maximum pitch RAO (occurred near B/L¼0.32). responses of our model were much smaller. The bulk density of our
Clearly, the installation of the two pneumatic chambers model was much smaller than theirs, but our surge RAOs were of
improved the hydrodynamic performance of Model 1, especially little difference from theirs. This was because the larger size of our
the transmission coefficient and the motion responses. Moreover, pneumatic chambers significantly increased the virtual mass of the
Model 1 might dissipate additional energy by the air flow through breakwater (when the breakwater surges with waves, it needs to
the opening on the top of each pneumatic chamber besides through move the water inside the pneumatic chambers). The installation of
the friction and flow separation. Both the maximum energy dissipa- pneumatic chambers on both sides of the breakwater significantly
tion and maximum air-pressure fluctuation occurred at a wave increased the moment of inertia, so the pitch responses our model
period corresponding to B/L¼0.24; while the maximum pitch RAO were effectively reduced; however, the motion responses of our
also occurred around B/L¼0.24. In principle, decreasing the wave models were relatively insensitive to the change in wave period.
period weakens the interaction of the incident waves with the tips Uzaki et al. (2011) examined a floating breakwater model with
of the chamber walls (when one half of the wave length is smaller a truss structure attached to both the front and back sides of a
than the draught, there will be no such interaction). Hence, the two box-type floating breakwater to increase the energy dissipation by
pneumatic chambers did not increase the energy dissipation for extra wave breaking. In terms of transmission coefficient, their
short period waves (see Fig. 8), but rather helped dissipate more design improved the performance mainly for short waves, while
energy for long period waves with additional energy dissipation our design improved the performance over a wide range of wave
from the motion of the water inside the chamber (i.e., the additional frequencies, especially for long waves. Our design dissipated energy
26 F. He et al. / Ocean Engineering 51 (2012) 16–27

more efficiently for long waves, while their design dissipated energy 2) The pneumatic chambers also effectively reduced the wave
more effectively for short waves even though the energy dissipation transmission for all wave periods. In addition to the wave
for their model was found over a wide range of wave frequencies. scattering associated with the two plates forming the pneu-
We remark here that, rather than just dissipating wave energy, our matic chambers, the reduction in wave transmission also came
model has a potential to convert wave energy into electricity by from two other sources: (a) the motion-generated radiated
installing Wells turbines to the pneumatic chambers, although this waves into the leeward side of the breakwater were reduced,
application potential needs to be further explored by examining the and (b) extra energy was dissipated by the pneumatic cham-
air-pressure change caused by the installation of turbines. bers, thus lesser wave energy was reflected or transmitted.
In the present study, the pressure fluctuations inside the back 3) Increasing the draught of the floating breakwater reduced the
chamber were always weak due to the limitation of the model size. surge, heave and pitch motions, but not very much. The air
However, even though the back chamber did not function as pressure fluctuations inside the front chambers decreased
effectively as the front in terms of dissipating energy, its effects on with increasing draught. For the long as well as very short
reducing the wave transmission coefficient were still significant. A period waves, the breakwater with a deeper draught was more
key benefit of the present design is to dissipate more energy of the effective in reducing the transmitted waves.
long period waves by the installation of the pneumatic chambers. We 4) Given the same geometry of the two pneumatic chambers, the
are currently conducting a follow-up study whereby the front and back chamber did not function as efficiently as the front chamber
back chambers have different geometries, with the natural period of in terms of extracting wave energy. This may be improved in the
the water column inside the back chamber specifically designed for future by varying the geometry of the back chamber.
longer period waves. In doing so, when the long period waves are
transmitted beneath the bottom of the floating structure, the back Overall, the results in the present study show that the
chamber may be more effective in dissipating the transmitted waves. installation of pneumatic chambers to a floating breakwater can
be an effective way to improve its hydrodynamic performance for
coastal protection. Moreover, pneumatic chambers can poten-
4. Conclusions tially be turned into devices converting wave energy to electricity
by installing Wells turbines to the chambers.
A new design of floating breakwater equipped with pneumatic
chambers is introduced in this study. The main findings from this
experimental study are the following: Acknowledgments

1) With the pneumatic chambers, the responses of the floating The authors acknowledge the financial support from the Ministry
breakwater to regular waves were mitigated by the water mass of Education, Singapore, through the AcRF Tier 2 Grant no.
inside the chambers and the increase in momentum of inertia. MOE2008-T2-070 and the NTU Start-Up Grant no. CEE SUG 3/07.

1200
Surge Translation

1100
(mm)

1000
900
800
0 20 40 60 80 100 120 140 160 180
Time Series (s)
Heave Translation

-800
-810
(mm)

-820
-830
-840
0 20 40 60 80 100 120 140 160 180
Time Series (s)

4
Pitch Rotation

2
(degree)

0
-2
-4
0 20 40 60 80 100 120 140 160 180
Time Series (s)

Fig. A1. Sample temporal data of motions including surge, heave and pitch; the experimental test conditions are: Model 1, wave height ¼ 0.04 m, water depth ¼0.9 m and
wave period ¼1.4 s.
 F. He et al. / Ocean Engineering 51 (2012) 16–27 27

Appendix Hegde, A.V., Kamath, K., Deepak, J.C., 2008. Mooring forces in horizontal interlaced
moored floating pipe breakwater with three layers. Ocean Eng. 35 (1),
165–173.
Sample temporal data of motions including surge, heave and Isaacson, M., Baldwin, J., Bhat, S., 1998. Wave propagation past a pile-restrained
pitch are shown in Fig. A1. The experimental test conditions are: floating breakwater. Int. J. Offshore Polar Eng. 8 (4), 265–269.
Model 1, wave height ¼0.04 m, water depth¼0.9 m and wave Kagemoto, H., 2011. Revisiting the complete wave transmission and reflection due
period¼1.4 s. The amplitude of motion was fitted by Matlab code to an array of 2-D surface-piercing truncated vertical plates fixed in regular
incident waves. Ocean Eng. 38 (8–9), 976–982.
and used to calculate the RAOs. Slow-drift motion can be clearly Kim, H.-T., Sawaragi, T., Aoki, S.-I., 1994. Wave control by pile-supported floating
seen in the recorded surge response. For this set of data, the breakwater. In: Proceedings of the International Offshore and Polar Engineer-
analysis of RAOs was done using the records from 110 s to 120 s. ing Conference. pp. 545–549.
Koo, W., 2009. Nonlinear time-domain analysis of motion-restrained pneumatic
floating breakwater. Ocean Eng. 36 (9–10), 723–731.
References Koutandos, E., Prinos, P., Gironella, X., 2005. Floating breakwaters under regular
and irregular wave forcing: reflection and transmission characteristics.
Christian, C.D., 2000. Floating breakwaters for small boat marina protection. In: J. Hydraul. Res. 43 (2), 174–188.
Proceedings of the 27th International Conference on Coastal Engineering ASCE. Liang, N.-K., Huang, J.-S., Li, C.-F., 2004. A study of spar buoy floating breakwater.
pp. 2268–2277. Ocean Eng. 31 (1), 43–60.
Diamantoulaki, I., Angelides, D.C., Manolis, G.D., 2008. Performance of pile- McCartney, B.L., 1985. Floating Breakwater Design. J. Waterw. Port Coastal Ocean
restrained flexible floating breakwaters. Appl. Ocean Res. 30 (4), 243–255. Eng. 111 (2), 304–318.
Dong, G.H., Zheng, Y.N., Li, Y.C., Teng, B., Guan, C.T., Lin, D.F., 2008. Experiments on Murali, K., Mani, J.S., 1997. Performance of cage floating breakwater. J. Waterw.
wave transmission coefficients of floating breakwaters. Ocean Eng. 35 (8–9), Port Coastal Ocean Eng. 123 (4), 172–179.
931–938. Rahman, M.A., Mizutani, N., Kawasaki, K., 2006. Numerical modeling of dynamic
Drimer, N., Agnon, Y., Stiassnie, M., 1992. A simplified analytical model for a responses and mooring forces of submerged floating breakwater. Coastal Eng.
floating breakwater in water of finite depth. Appl. Ocean Res. 14 (1), 33–41. 53 (10), 799–815.
Elchahal, G., Younes, R., Lafon, P., 2008. The effects of reflection coefficient of the Sannasiraj, S.A., Sundar, V., Sundaravadivelu, R., 1998. Mooring forces and motion
harbour sidewall on the performance of floating breakwaters. Ocean Eng. responses of pontoon-type floating breakwaters. Ocean Eng. 25 (1), 27–48.
35 (11–12), 1102–1112. Uzaki, K.-i., Ikehata, Y., Matsunaga, N., 2011. Performance of the wave energy
Falca~ o, A.F.d.O., 2010. Wave energy utilization: A review of the technologies. dissipation of a floating breakwater with truss structures and the quantifica-
Renewable Sustainable Energy Rev. 14 (3), 899–918. tion of transmission coefficients. J. Coastal Res. 27 (4), 687–697.
Fugazza, M., Natale, L., 1988. Energy Losses and Floating Breakwater Response. Vijayakrishna Rapaka, E., Natarajan, R., Neelamani, S., 2004. Experimental inves-
J. Waterw. Port Coastal Ocean Eng. 114 (2), 191–205. tigation on the dynamic response of a moored wave energy device under
Gesraha, M.R., 2006. Analysis of shaped floating breakwater in oblique waves: I. regular sea waves. Ocean Eng. 31 (5–6), 725–743.
Impervious rigid wave boards. Appl. Ocean Res. 28 (5), 327–338. Wang, H.Y., Sun, Z.C., 2010. Experimental study of a porous floating breakwater.
Goda, Y., Suzuki, Y., 1976. Estimation of incident and reflected waves in random Ocean Eng. 37 (5-6), 520–527.
wave experiments. In: Proceedings of the 15th International Conference on Williams, A.N., Abul-Azm, A.G., 1997. Dual pontoon floating breakwater. Ocean
Coastal Engineering ASCE. pp. 828–845.
Eng. 24 (5), 465–478.
Hales, L.Z., 1981. Floating breakwaters: state-of-the-art literature review, Techni-
Williams, A.N., Lee, H.S., Huang, Z., 2000. Floating pontoon breakwaters. Ocean
cal Report no. 81-1. US Army Engineer Waterways Experiment Station.
Eng. 27 (3), 221–240.
Vicksburg, Mississippi, USA.