Wave energy resource characterisation

of the Atlantic Marine Energy Test Site

B.G. Cahill

, T. Lewis
Hydraulics & Maritime Research Centre, University College Cork, Cork, Ireland
a r t i c l e i n f o
Keywords:
Wave energy resource
Test sites
Scalability
a b s t r a c t
The Atlantic Marine Energy Test Site (AMETS), a grid connected test
area for the deployment of full scale Wave Energy Converters
(WECs), is being developed by the Sustainable Energy Authority
of Ireland near Belmullet in Co. Mayo, Ireland. In this paper mea-
sured data provided by two wave buoys, positioned at a deepwater
location (100 m depth) and an offshore location (50 m depth), are
analysed in order to characterise the wave resource at the site. In
particular, a distinction is made between which sea states occur
with the most regularity and which wave conditions are the most
signicant for the capture of power. The spatial variation in the
occurrence of important summary statistics between the deepwa-
ter and offshore sites is examined and the difference in incident
wave power calculated. Finally, this paper compares conditions
at the Belmullet site with those measured at the quarter scale test
site in Galway Bay. An assessment on the degree of scalability
between resource parameters relevant to WEC power production
experienced at the two locations, as recommended by develop-
ment protocols, is made and methods for comparing benign and
exposed sites proposed.
2013 Elsevier Ltd. All rights reserved.
Introduction
In theory, the energy that could be extracted from the oceans is well in excess of any current, or
future, human requirements. While wave energy continues to lag behind conventional, carbon based
sources of power and other renewable sources of energy such as wind and solar, advances continue to
2214-1669/$ - see front matter 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.ijome.2013.05.001

Corresponding author. Tel.: +353 214250013.

E-mail address: [email protected] (B.G. Cahill).
International Journal of Marine Energy 1 (2013) 315
Contents lists available at SciVerse ScienceDirect
International Journal of Marine Energy
j our nal homepage: www. el sevi er . com/ l ocat e/ i j ome
be made. Though the total amount of installed capacity is still insignicant, reported in [1] to be
2 MW, the wave energy industry is beginning to approach the deployment of wave farms, arrays of
full-scale Wave Energy Converters (WECs). Methods to compare and evaluate the energy resource
at different locations, and at different scales, are required in order to inform WEC project developers
and allow them to select the most suitable sites to achieve optimal power capture and economic per-
formance from their installations.
Many assessments of wave energy resource have previously been carried out at regional [2,3], na-
tional [4,5] and global scales [6]. The primary purpose of these studies is to give an indication of the
potential hydrodynamic power that is available at a location of interest, while the classication of sea-
sonal and inter-annual trends in the availability of the resource and the identication of extreme
events which may impact on the operation and survival of WECs are also commonly presented. The
focus of this paper is towards characterising the wave resource with reference to the potential power
available and the performance of typical devices, and also to allow for the comparison of potential
sites for the development of WEC projects. In particular, efforts are made to differentiate between
which sea states are the most common at a location of interest and which sea states contribute most
to the incident wave energy and therefore to the power that can be expected to be produced by a WEC.
This analysis is carried out using measured wave data from buoys located at the Atlantic Marine En-
ergy Test Site (AMETS) that is currently being developed close to the town of Belmullet in Co. Mayo, in
order to describe the conditions that can be expected at exposed sites off the west coast of Ireland. The
spatial variability at this site is also assessed by examining concurrent measurements taken at two
locations with depths of 50 and 100 m.
Finally, these techniques are applied so that a comparison can be made between the AMETS loca-
tion and the quarter scale test site in Galway Bay in order to give an indication of how accurately such
test sites simulate real sea conditions and allow the performance of full-scale WECs to be predicted
from the testing of scaled models.
Wave data
Wave energy test sites
The analysis undertaken for this study relies on point measurements of wave conditions taken at
Irelands two wave energy test sites, whose locations are shown in Fig. 1. Test sites such as these
are important resources for the technical progress of a WEC from an initial concept to the deployment
of commercial wave farms along a staged development programme, as outlined in [7].
The benign, quarter scale, wave energy test site in Galway Bay on the west coast of Ireland provides
a location for developers of oating WECs to deploy and monitor device prototypes in relatively shel-
tered conditions, as suggested in Phase 3 of the development protocols. The site has an area of 37
Fig. 1. The Atlantic Marine Energy Test Site (AMETS) and Irelands wave energy test facilities (Image courtesy of SEAI).
4 B.G. Cahill, T. Lewis / International Journal of Marine Energy 1 (2013) 315
Hectares with water depth of between 21 and 24 m and a tidal range of 4 m and has previously been
used by Ocean Energy Ltd. and Wavebob Ltd. Galway Bay is semi-enclosed as a result of the shelter
afforded by the Aran Islands but does experience swell waves from the West and South-West as well
as the local, fetch limited, wind seas. It will be shown a later section of this paper that these wind seas
are a good representation at quarter scale in combinations of height and period for the Atlantic Ocean.
A grid-connected test site for the appraisal of full scale WECs is currently being developed by the
Sustainable Energy Authority of Ireland (SEAI) off the West Coast, near Belmullet in Co. Mayo and will
be known as the Atlantic Marine Energy Test Site (AMETS). AMETS offers the opportunity for prototype
WECs to be exposed to the type of harsh, energy rich wave climate that can be expected off the west
coast of Ireland and the United Kingdom. It is envisaged that this site will be used by developers for
the nal stages of device testing prior to commercial deployment. Details on the development of this
test site can be found in [8].
Wave data measurement
A Datawell Directional Waverider, a spherical, surface-following measurement buoy, has been
gathering data at the AMETS location since December 2009. This buoy is located at the 50 m depth
contour, designated as Berth B, and transmits a set of spectral and time series readings to shore every
30 min.
In addition, from MayOctober of 2010 a Fugro Wavescan discus shaped buoy with a diameter of
approximately 2.75 m, was positioned at the planned Deep Water Test Area, approximately 10 km
North-East of the Waverider buoy. The spectral wave measurements from this buoy have yet to be
processed by the Marine Institute, the agency tasked with managing the data, therefore only basic
summary statistics have been available for the analysis outlined in this study.
Point measurements of wave conditions have been collected from Galway Bay using individual
Datawell Waverider buoys since 2006. A non-directional Waverider buoy was originally deployed at
the site until it was replaced in the summer of 2008 by a directional Waverider. This buoy is currently
positioned at 53 13.7
0
N, 9 16.13
0
W. The availability of measured data from Galway Bay (GB Wave-
rider) for the year 2010, as well as from the AMETS Waverider (BM Waverider) and Wavescan (BM
Wavescan) buoys, is illustrated in Fig. 2. It is notable that gaps in data exist for both of the Waverider
buoys during the summer months due to scheduled maintenance and cleaning.
Fig. 2. Availability of measured wave data from Irelands wave energy test sites in 2010.
B.G. Cahill, T. Lewis / International Journal of Marine Energy 1 (2013) 315 5
Resource characterisation of the Atlantic Marine Energy Test Centre
Summary statistics and wave power distribution
WEC developers will require a thorough knowledge of the conditions at a potential project location,
or at a test site like AMETS, before they can commit to deploying their devices there. Bi-variate scatter
plots of important summary statistics provide a useful method of ascertaining an overall understand-
ing of the wave resource at an area of interest. In Fig. 3 the sea states experienced at AMETS in 2010
are presented in terms of the signicant wave height, H
m
0
, and the energy period, T
E
. T
E
is dened as
being equivalent to the period of monochromatic wave whose height is equal to H
m
0
, and which has
the same energy as the irregular sea state in question. Both these parameters are calculated from the
moments of measured spectra. H
m
0
is dened as:
H
m
0
4

m
0
p
1
while T
E
is given by
T
E
m
1
=m
0
2
where m
n
is the nth spectral moment. T
02
is the primary wave period parameter provided by the
Wavescan buoy. It is equivalent to the average of the zero-crossing periods and is also denoted by T
z
.
T
02

m
0
=m
2
p
3
Fig. 3 illustrates that although the majority of sea states at AMETS fall within the ranges
1 m < H
m
0
< 3 m and 6 s < T
E
< 11 s, a wide spread of conditions are experienced at the site, including
severe storm conditions where the H
m
0
exceeded 10 m. This is indicative of how well suited the loca-
tion is to its role as a test site and how it will afford developers the opportunity to prove the surviv-
Fig. 3. Bi-variate scatter plot for AMETS showing the percentage occurrence of H
m
0
T
E
for 2010.
6 B.G. Cahill, T. Lewis / International Journal of Marine Energy 1 (2013) 315
ability of their devices during extreme events. However, it should be noted that several years worth of
data will need to be collected before a more complete assessment of the wave climate at the site can
be made.
These plots are not without their limitations. In [9] it has been shown that a signicant amount of
deviation can exist in the spectral shapes of measurements that fall within the same scatter diagram
element, while [10] suggests that a much wider range of parameters are required to fully assess the
resource with reference to the performance of wave energy devices. These plots also fail to fully de-
scribe which sea states contribute most to the wave power experienced at a site. Wave power per unit
crest length, P (kW/m), is calculated from the measured spectra using the formula
P
qg
2
4p
m
1
4
with the assumption of deep water, where q is the density of sea water (1025 kg/m
3
) and g is accel-
eration due to gravity (9.81 m/s
2
). Eq. (3) is more commonly written in terms of H
m
0
and T
E
, giving
P 0:4H
2
m0
T
E
5
The AMETS scatter diagram is redrawn in Fig. 4 with percentage occurrence replaced by percentage
contribution to total annual incident wave energy. When this diagramis compared with Fig. 3 it can be
seen that the cells with the highest contribution are not necessarily those with the highest occurrence.
There is an obvious upward shift in the positions of the most signicant sea states. WEC developers
can take advantage of this by designing their devices to capture as much of the available power as pos-
sible at the sea states that prove to be the most energetic (when considered over the course of a year)
at a potential wave farm site, rather than targeting the conditions which occur most frequently.
It is also notable that a number of very rare sea states that are the product of extreme storm events
have a disproportionate contribution to annual power. For example, the four sea states (0.03% of total
occurrences) that were measured within the range 12.5 m < H
m
0
< 13 m and 13s < T
E
< 17s, were found
Fig. 4. Bi-variate scatter plot for AMETS showing the contribution to annual energy of H
m
0
T
E
for 2010.
B.G. Cahill, T. Lewis / International Journal of Marine Energy 1 (2013) 315 7
to be responsible for nearly 2% of the total gross, power for the year. In reality the performance of a
WEC would be reduced in these conditions due to losses in efciency, or even reduced to zero as
the device enters survival mode. To account for this discrepancy the concept of maximum exploitable
power, P
exp
, was introduced in [11]. P
exp
is dened as being four times the average incident wave
power, written mathematically in Eq. (5) as
P
exp
4
X
N
i1
P
i
N
6
For highly energetic sea states any power above the P
exp
is deemed to be superuous and is dis-
counted. While the value of P
exp
given by Eq. (5) is somewhat arbitrary it nonetheless provides a useful
tool for gaining a more accurate picture of what sea states should be deemed to be most important for
the capture of wave power by a WEC. In practice, the performances of various WEC types in extreme
conditions will differ greatly, but real data from full scale testing will allow for more appropriate
thresholds, tailored to individual device designs, to be determined.
Analysis was also carried out to investigate which sea states contribute most to the performance of
a typical WEC. The publically available power matrix [12], shown in Fig. 5, for an early, 750 kW, ver-
sion of the Pelamis device is combined with the measured values of H
m
0
and T
E
from the AMETS Wave-
rider buoy to replicate the power production of a wave energy device exposed to conditions
experienced at the Belmullet site.
The difference that exists between the occurrence of a particular value of H
m
0
and extent to which
sea states within this range contribute to wave energy is illustrated in Fig. 6. The bar plots indicate the
contribution that each individual H
m
0
bin, with spacings of 0.5 m, makes to the total power from the
year in question. The three measures of wave power dened previously: gross, theoretical power from
Eq. (4), exploitable power (Eq. (5)) and the output power from the Pelamis matrix are plotted. The so-
lid orange line details the percentage occurrence of H
m
0
for each of these bins. The apparent disconnect
between wave height occurrence and contribution to annual power that was suggested previously by
Fig. 4 is easily identied here, with the most signicant inputs to power all occurring less often than
the most commonly experienced value of H
m
0
.
Similar results are observed in Fig. 7 where the occurrence of the average period, T
02
, with 1 s bins,
is compared to the contribution the period bins make to wave power. Shorter period waves can be
seen to contribute proportionally more to Pelamis power than they do to gross and exploitable power
while much of the power from longer period seas appears to be unexploitable.
Fig. 5. Power matrix for a 750 kW Pelamis device, from [12].
8 B.G. Cahill, T. Lewis / International Journal of Marine Energy 1 (2013) 315
Spatial variability
The spatial variability in the wave energy resource at AMETS is investigated through the compar-
ison of measurements collected by the Wavescan buoy positioned at a depth of 100 m and the Wave-
rider deployed at 50 m depth. Only datasets that were measured concurrently at the two buoys were
selected for analysis to remove the bias that seasonality and individual storm event may have on the
results.
Occurrence and exceedance of the summary statistics H
m
0
and T
02
are compared in Fig. 8 for con-
current measurements taken at the 100 and 50 m depths. Both of these gures indicate that in general
the Wavescan buoy experiences waves of slightly greater height and of longer periods. This may be in
part due to the shelter from waves approaching from the South that the 50 m location receives from
the nearby Inishglora and Inishkea islands but further directional analysis will be required once the
supplementary data from the Wavescan buoy becomes available before this conclusion can be accu-
rately made. It is also evident that the distributions of H
m
0
and T
02
occurrence and exceedance at the
two locations follow very similar trends.
The variability of incident wave power at the AMETS location is observed in Fig. 9 through compar-
ison of the measured data from the 100 and 50 m depths. As spectral information was unavailable
from the Wavescan buoy located at the 100 m depth it was not possible to derive the energy period
T
E
and calculate the power per metre crest using Eq. (4). Instead, it was necessary to calculate the inci-
dent wave power using the average period T
02
following the approach described in the Appendix at
the end of this paper. It was also not possible to account for the inuence of nite water depth on
the resulting power so the deep water assumption was followed. To ensure homogeneity the same ap-
proach was followed for the 50 m site, even though values of T
E
could be derived from the data pro-
vided by the Waverider buoy.
It has already been shown that in general the 100 m site experiences waves that are both slightly
higher and longer than those at the 50 m location, therefore it is unsurprising that the average power
Fig. 6. Contribution to total annual power compared to occurrence for 0.5 m bins of H
m
0
.
B.G. Cahill, T. Lewis / International Journal of Marine Energy 1 (2013) 315 9
Fig. 7. Contribution to total annual power compared to occurrence for 1.0 s bins of T
02
.
Fig. 8. Occurrence (bars, left hand axis) and exceedance (solid lines, right hand axis) of H
m
0
(top) and T
02
(bottom) for 100 and
50 m depths at AMETS.
10 B.G. Cahill, T. Lewis / International Journal of Marine Energy 1 (2013) 315
experienced at the 100 m depth is greater than that at the 50 m location. The slope of the trend line
indicates that the power at the 50 m location is less than 80% of what can be expected at the 100 m
point. Introducing the concept of Exploitable Power was found to have little effect on the spatial dis-
tribution of power. This is possibly due to the limited range of concurrent measurements which were
mostly conned to summer months. Were data available from winter months which experienced
storm conditions it is likely that there would be greater variation between gross and exploitable
power, particularly at the Wavescan location.
Fig. 9. Concurrent values of Wave Power at AMETS.
Fig. 10. Bi-variate scatter plot for Galway Bay converted to full-scale, showing the percentage occurrence of H
m
0
T
E
for 2010.
B.G. Cahill, T. Lewis / International Journal of Marine Energy 1 (2013) 315 11
Fig. 11. Occurrence and exceedance of H
m
0
(top) and T
E
(bottom) for AMETS and Galway Bay measurements at full-scale.
Fig. 12. Individual, averaged and theoretical spectra within the range 0.625 m < H
m
0
< 0.75 m and 3.0 s < T
02
< 3.5 s for the
Galway Bay test site.
12 B.G. Cahill, T. Lewis / International Journal of Marine Energy 1 (2013) 315
Resource scalability
It is likely that device developers who will deploy WECs at AMETS will have previously tested ear-
lier versions of their devices at the Galway Bay quarter scale test site, or at other similarly benign loca-
tions. In this section analysis of measured data is used to compare the wave conditions experienced at
the test sites in order to assess whether they are sufciently well matched to allow accurate compar-
isons to be made between performance results derived at the various scales.
Values for the summary statistics H
m
0
and T
E
were calculated from the available measurements
from the Datawell Waverider buoy located in Galway Bay. Froude scaling was then applied to convert
the scatter diagram from Galway Bay for 2010 to full-scale (Fig. 10). Visual comparison with Fig. 3
indicates that while the percentage occurrence of individual sea states may vary between the two
sites, the overall range of conditions that occur at AMETS is fully replicated, at approximately quarter
scale, in Galway Bay. This distinct correlation ensures that WECs deployed at the test site will encoun-
ter a sufciently large spread of wave conditions, and consequently gather ample amounts of opera-
tional data, to allow for accurate predictions of device performance at full scale to be established.
There is a strong similarity in the occurrences of extreme conditions measured at the test sites, par-
ticularly in terms of sea state slopes which follow the constant 1/13 line. This association will provide
developers who deploy in Galway Bay during winter months with the opportunity to prove the sur-
vivability of their devices in scaled conditions equivalent to the storms experienced at exposed loca-
tions off the west coast of Ireland.
It is notable, however, that the Galway Bay scatter diagram displays a greater contribution from
long period sea states with a relatively low signicant wave height. These sea states do not have
any equivalent in the AMETS gure. The inuence of these swell dominated sea states can be seen
in the plots of occurrence and cumulative exceedance for H
m
0
and T
E
displayed in Fig. 11. While the
scaled values of H
m
0
from Galway Bay are generally well matched with those fromAMETS, they exhibit
a much greater occurrence of sea states with H
m
0
< 1 m. Agreement is worse for the T
E
values, as had
Fig. 13. Individual, averaged and theoretical spectra within the range 2.5 m < H
m
0
< 3.0 m and 6 s < T
02
< 7 s for the AMETS
location (50 m depth).
B.G. Cahill, T. Lewis / International Journal of Marine Energy 1 (2013) 315 13
been suggested by Fig. 12. Results gained from testing during these long period sea states T
E
> 13 s
at full scale would provide little relevant knowledge about the performance of a WEC concept in
realistic, full scale, conditions. Device developers should be careful to ensure that these sea states
are excluded from performance analysis, though in general their impact can be considered negligible,
however, as they generally occur when wave energy is low and below the operating threshold of de-
vices; for example, during the CORES deployment in 2011 the OWC turbine did not generate power
until H
m
0
exceeded 0.8 m [13].
The presence of these swell waves can also be detected in the analysis of spectral shapes. In Fig. 12
the measured spectra that fall within the range 0.625 m < H
m
0
< 0.75 m and 3.0 s < T
02
< 3.5 s are plot-
ted together, along with the average of the spectral ordinates and the classical Bretschneider spectrum
for similar summary statistics. Two distinct components are noticeable; the long period swell identi-
ed above along with a wind sea component centred on a frequency of 0.3 Hz which is appropriate for
a one quarter scale seaway. Spectra measured by the AMETS Waverider which fall within the equiv-
alent full scale range (2.5 m < H
m
0
< 3.0 m and 6 s < T
02
< 7 s) are plotted in Fig. 13 and are found to
conform relatively well with what would be expected from an open water site, as evidenced by the
similarity of the average spectral shape to the Bretschneider spectrum. It is evident that the long per-
iod components that appear so distinctly in Fig. 12 are not as inuential here and that it is likely that
sea and swell components from benign sites need to be isolated and dealt with separately in any anal-
ysis prior to conversion to full scale.
Conclusions
This paper has demonstrated that the most prevalent wave conditions experienced at an exposed
ocean location on the west coast of Ireland are not the most energy rich when observed over the
course of a year, both with respect to the theoretical, incident wave energy and in terms of energy cap-
ture from an actual device. While the lack of an archive of measured data spanning a number of years
makes a comprehensive assessment of the wave climate at the AMETS location difcult, these results
have obvious implications for the design of WECs and highlight the range of sea states that they will
need to be tuned to in order to extract the maximum amount of energy from a particular site.
Individual values of incident wave power were found to vary signicantly (20%) between the 100
and 50 m depths at AMETS. However it was also shown that the distributions of occurrences of H
m
0
and T
02
appear to follow similar patterns at both points, indicating that a considerable level of homo-
geneity exists in the wave conditions at the site. Future analysis of spectral and directional data from
the Wavescan buoy will allow for a clearer understanding of the spatial variability at the site.
This work has also highlighted the level of scalability that exists between the wave climates at
AMETS and the quarter scale test site in Galway Bay. It was shown that in general there is excellent
agreement between the sites and that any particular sea state experienced at AMETS is replicated
at scale in Galway Bay. As a result developers who conduct a sufciently long period of testing at
the quarter scale site should gather the necessary performance data to progress to the nal stages
of testing. Additionally, it was demonstrated that the presence of long period swell forms sea states
and spectra that are not suitable for following the same scaling procedures and that care should be
taken when interpreting WEC performance data collected during these conditions.
Acknowledgements
The authors would like to acknowledge the Marine Institute and the Sustainable Energy Authority
of Ireland for providing the data used in this study. This research is funded by the Irish Research Coun-
cil for Science, Engineering and Technology (IRCSET) under the EMBARK Initiative.
Appendix A. Calculation of Power from H
m
0
and T
02
For a generalised PiersonMoskowitz Spectrum for fully developed sea-states, also known as the
Bretschneider Spectrum, let the relationship between the energy period, T
E
, and the average zero-
crossing period, T
02
, be given by
14 B.G. Cahill, T. Lewis / International Journal of Marine Energy 1 (2013) 315
T
E
aT
02
7
where a is a constant.
Rewriting Eq. (6) in terms of spectral moments gives
m
1
m
0
a

m
0
m
2
r
8
From [14] the spectral moment m
n
can be given the general form
m
n

1
4
AB
n
4
1
u 1
n
4
h i
9
allowing Eq. (7) to be rewritten as
0:2266
A
B
5
4
A
4B
a

A
4B
0:443
A

B
p
v
u
u
t
10
This simplies to give a = 1.206, allowing Eq. (4) to be written in terms of T
02
, giving
P 0:59H
2
m
0
T
02
11
In reality there will be some divergence from the Bretschneider spectral shape, and this will be re-
ected in the value of a. Further studies of the measured data [15] have shown that a wave period ra-
tio of approximately 1.35 is more appropriate for the AMETS location.
References
[1] Implementing Agreement on Ocean Energy Systems (OES-IA), Annual Report 2010, 2010.
[2] P. Lenee-Bluhm, R. Paasch, H.T. zkan-Haller, Characterizing the wave energy resource of the US Pacic Northwest,
Renewable Energy 36 (8) (2011) 21062119.
[3] G. Iglesias et al, Wave energy potential in Galicia (NW Spain), Renewable Energy 34 (11) (2009) 23232333.
[4] M.G. Hughes, A.D. Heap, National-scale wave energy resource assessment for Australia, Renewable Energy 35 (8) (2010)
17831791.
[5] ESBI, Accessible Wave Energy Resource Atlas: ESBI Report 4D404A-R2 for Marine Institute (MI) and Sustainable Energy
Ireland (SEI), 2005.
[6] G. Mrk et al., Assessing the global wave energy potential, in: 29th International Conference on Ocean, Offshore Mechanics
and Artic Engineering, Shanghai, China, 2010.
[7] Hydraulics & Maritime Research Centre, Ocean Energy Development and Evaluation Protocol, 2003.
[8] J. Ascoop, M. Frielding, The development of a Wave Energy Test Site in Belmullet (Ireland): experience to date, in: 3rd
International Conference on Ocean Energy, Bilbao, Spain, 2010.
[9] S. Barrett et al., Spatial & spectral variation of seaways, in: Proceedings of the 8th European Wave and Tidal Energy
Conference, Uppsala, Sweden, 2009.
[10] J.-B. Saulnier et al, Wave groupiness and spectral bandwidth as relevant parameters for the performance assessment of
wave energy converters, Ocean Engineering 38 (1) (2011) 130147.
[11] M. Folley, T.J.T. Whittaker, Analysis of the nearshore wave energy resource, Renewable Energy 34 (7) (2009) 17091715.
[12] Pelamis Wave Power Ltd., www.pelamiswave.com.
[13] F. Thiebaut, et al., Testing of a oating OWC device with movable guide vane impulse turbine power take-off, in: 9th
European Wave and Tidal Energy Conference, Southampton, United Kingdom, 2011.
[14] M.J. Tucker, E.G. Pitt, Waves in ocean engineering, in: R. Bhattacharyya, M.E. McCormick (Eds.), Elsevier Ocean Engineering
Book Series, vol. 5, Elsevier, 2001.
[15] B.G. Cahill, Characteristics of the Wave Energy Resource at the Atlantic Marine Energy Test Site. PhD Thesis, University
College Cork, 2013.
B.G. Cahill, T. Lewis / International Journal of Marine Energy 1 (2013) 315 15