Professional Documents
Culture Documents
Wave Tank Testing Standard
Wave Tank Testing Standard
Conversion Systems
Foreword
This document has been prepared in consultation with The European Marine Energy
Centre Ltd (EMEC) and with other interested parties in the UK marine energy community.
It is one of twelve publications in the Marine Renewable Energy Guides series, as detailed
in the following figure.
Certification
Guide
Assurance
Device development
Design Basis
(including install,
mooring and
foundations) Tidal Resource Wave Resource
Tidal Tank Test Wave Tank Test Grid Connection
Assessment Assessment
Reliability and
Tidal Wave Maintainability
Performance Performance Site Definition/Selection
Testing Testing
Device Performance
Acknowledgements
This document was written by Brian Holmes, Hydraulics and Maritime Research Centre,
University College Cork under contract from the European Marine Energy Centre Ltd (EMEC).
The author would like to thank: Sean Barrett, for his indispensible assistance in producing
this report; John Griffiths for his patience and understanding; Jamie Grimwade for his
tolerance; Sophie Erskine for her editorial contribution; Andrew Want for his assistance;
Kim Nielsen, David Ingram, Remy Pascal, Gregory Payne, Ken Tease, Stephanie Merry and
Matt Folley for their review comments.
Tank Testing of Wave Energy
Conversion Systems
The information contained in this document is for guidance only and it is not intended, and should not be used,
as a substitute for taking technical advice in any specific situation. Whilst every effort has been made to provide
accurate information on these pages, neither the European Marine Energy Centre (EMEC), nor any of its employees,
make any warranty, expressed or implied, or assume any liability (to the extent permitted by law) or responsibility
for the accuracy or completeness of any information contained in this document. In no event shall EMEC be liable
(to the extent permitted by law) for any special, incidental, indirect or consequential damages or any damages
whatsoever, whether in an action for misrepresentation or of contract, negligence or other delictual action,
arising out of or in connection with, the use of the information available in this document or for omissions or the
continued accuracy of this document.
The right of Brian Holmes to be identified as the author of this Work has been asserted by him in accordance with
sections 77 and 78 of the Copyright, Designs and Patents Act 1998.
EMEC has no responsibility for the persistence or accuracy of URLs for external or third-party internet websites
referred to in this book, and does not guarantee that any content on such websites is, or will remain, accurate or
appropriate.
ISBN 978-0-580-67262-0
Contents
Introduction 1
1 Scope 4
2 Terms, definitions, abbreviations, symbols and units 4
2.1 Terms and definitions 4
2.2 Abbreviations 8
2.3 Symbols and units 9
3 Purpose of testing 10
3.1 General guidance 10
3.2 Use of regular seas 18
3.3 Use of irregular seas 22
4 Recommended wave tests 27
4.1 Energy capture performance 27
4.2 Survival tests 34
5 Test equipment 36
5.1 Model design 36
5.2 Test facilities 45
5.3 Data acquisition system (DAQ) 50
5.4 Calibration 57
6 Calculation procedures 60
6.1 General 60
6.2 Analysis of tank testing results 60
6.3 Extrapolation of results 75
7 Reporting 78
7.1 General 78
7.2 Purpose of tests 79
7.3 Facilities 79
7.4 Model 79
v
Marine Renewable Energy Guides
vi
Tank Testing of Wave Energy
Conversion Systems
Introduction
Within the wave energy community there is, at present, no agreed common approach
for the development and evaluation of energy extraction devices to assist small and
medium-sized enterprises (SMEs), funding bodies, decision makers or other interested
stakeholders. This report attempts to satisfy that requirement by providing guidelines for
a structured programme that will mitigate both the technical and fiscal risks.
The overall programme is based on a five phase approach, which increases in complexity
and required investment as it progresses. Table 1 gives an overview of the phases. This
document deals with the first three only, that is from initial concept review to large scale
(circa λ = ¼) sea trials at benign outdoor test stations.
At the conclusion of each phase, or sub-phase, a set of decision or stage gates are executed
to assist in the evaluation of a device and its potential for successful development. These
are shown in Table 2. Since no devices have progressed to commercialization the actual
numbers to set as limits for each phase are a combination of estimates from the wave
energy community and economics from the off shore wind industry. These will be updated
in future editions of this document as more confirmed realistic evidence is collected.
Please also note that although a standard format can be used to impartially assess the
progress of devices, each test programme will be bespoke in detail. This document,
therefore, describes the minimum test requirements. At the end of each phase, it is
imperative that the evaluation of the status of the device development is conducted in
order to carry out the appropriate planning process for future phases.
This test procedure is to be regarded as codes to follow, based on best practices established
over previous years, rather than rules that have to be accepted verbatim.
It is also stressed that the test programme presented is a basic, practical approach for the
device development not a rigorous, academic thesis.
1
2
Table 1 — Phase overview
Primary scale (λ) λ = 1 : 25 − 100 (\ λt = 1 : 5–10) λ = 1 : 10–25 λ = 1 : 3–10 λ = 1 : 1–2 λ = Full size
Tank 2D flume and 3D basin 3D basin Benign site Exposed site Open location
Marine Renewable Energy Guides
Duration (inc. 1–3 weeks 1–3 months 1–3 months 6–12 months 6–18 months 12–36 months 1–5 years
analysis)
Typical no. tests 50–500 250–500 100–250 100–250 50–250 Continuous Statistical sample
Excitation/Waves Monochromatic Panchromatic waves Deployment: pilot Extended Full scatter diagram for initial
linear waves (20 min full scale) site sea spectra test period evaluation, continuous thereafter
(10–25 ∆ƒ) +15 classical spectra long to ensure
crested head seas Long & short
all seaways
Panchromatic crested classical
included
5 reference seas
Absorbed power
(kW)
Converted power
(kW)
Mass (t)
Manufacturing
cost (€)
3
Tank Testing of Wave Energy Conversion Systems
Marine Renewable Energy Guides
1 Scope
This document specifies a structured development programme and specific test
procedures for the early stages of a wave energy converter development. The format
is based on traditional engineering methods similar to the Technology Readiness Level
(TRL) introduced by National Aeronautics and Space Administration (NASA).
2.1.2
bi-modal spectrum
spectral shape consisting of two distinct concentrations of energy variance due to a
locally driven wind sea in the presence of a swell sea where a valley exists between the
low frequency swell and the high frequency wind sea
2.1.3
Bretschneider spectrum
expansion of the Pierson–Moskowitz (PM) spectrum, allowing for the description of the
spectral shape of seaways other than fully developed by utilizing the variance and peak
frequency as input
4
f 4
5 Hm 0 fp
2
5 p
S( f ) B = exp −
16 f f 4 f
2.1.4
device power
pneumatic or mechanical power produced by a device due to the interaction of the
relative inertias of the device system as excited by the seaway
EXAMPLE Device hull and water column, seabed and float, water reservoir and turbine.
2.1.5
draft
dimension of a floating body below the mean water surface due to a load bearing on
the body
2.1.6
finite waves
waves outside the linear regime that have a finite amplitude
4
Tank Testing of Wave Energy Conversion Systems
NOTE When this is the case the waves are asymmetrical about the mean level, and so it can not be said
that the surface elevation will be Gaussian distributed.
2.1.7
following sea
seaway impacting on a floating body at the stern, and travelling in direction from the
stern to the bow of the body
2.1.8
freeboard
dimension of a floating body above the mean water surface
2.1.9
Froude scaling
scaling method based on the ratio of inertial forces to gravitational forces, used to model
wave and surface behaviour
2.1.10
head sea
seaway impacting on a floating body at the bow, and travelling in direction from the bow
to the stern of the body
2.1.11
heave
translational oscillatory motion parallel to the vertical axis of a floating body caused by
an excitation or a displacement from rest along the vertical axis
2.1.12
Joint North Sea Wave Project (JONSWAP) spectrum
multiplication of the PM spectrum by an enhancement function to facilitate fetch limited
conditions, having three input parameters: peak frequency, peak enhancement factor
and the enhancement shape widths:
( f − fp )
2
f 4 exp− 2σ 2fp2
αg 1
2
5 p
S( f ) J = exp − γ
( )
2π f 4 f
4 5
2.1.13
linear waves
waves based on the assumptions of Airy wave theory
NOTE This is a theory for the propagation of waves on the surface of a potential flow above a
horizontal bottom. The free surface elevation η(x,t) of one wave component is sinusoidal, as a function
of horizontal position x, and time t.
η( x , t ) = a cos( kx − ω t )
The probability distribution of a random sample of the surface elevation of a wavetrain composed of
the superposition of linear waves will yield a Gaussian distribution. This means that there will be no
second or higher order effects.
5
Marine Renewable Energy Guides
2.1.14
monochromatic wave excitation
excitation of a floating body due to a sinusoidal wave composed of a single frequency
2.1.15
panchromatic excitation
excitation by a sea state comprising the superposition of multiple single frequency
sinusoidal components, the distribution of which is dictated by the spectral shape
2.1.16
Pierson–Moskowitz (PM) spectrum
empirical equation defining the one dimensional spectral shape of the distribution of
variance over a frequency range, based on the assumption that the input wind is in
equilibrium with the waves, also known as fully developed
f 4
α g2 1 5 p
S( f )PM = exp −
(2π ) f 4 f
4 5
2.1.17
pitch
rotational oscillatory motion about the horizontal transverse axis of a floating body due
to an excitation or displacement from its equilibrium position
2.1.18
quarter sea
seaway approaching a floating body at an angle between a beam sea and a following sea
head seas
0°
bow
beam seas
midships
270° 90°
following seas
6
Tank Testing of Wave Energy Conversion Systems
2.1.19
roll
rotational oscillatory motion about the horizontal longitudinal axis of a floating body due
to an excitation or displacement from its equilibrium position
2.1.20
significant steepness, Ss
ratio between the significant wave height and the wave length for a panchromatic
time series:
2π Hs
gTz2
Sea states defined by the Pierson–Moskowitz equation have a constant significant steepness
of Ss = 0.0507. Seaways in general do not exceed a significant steepness of Ss = 0.1.
2.1.21
spreading
directional distribution applied to one dimensional spectra to form three dimensional
spectra
( )
NOTE S( f ,θ ) = S( f )D(θ , f ), where D(θ , f ) = F ( s )cos 2 s 21 θ − θ m is the most common assumption with a
peak at the mean direction.
2.1.22
steepness, S
ratio of the wave height, H, to the wave length, λ, of a single frequency sinusoidal wave
NOTE The steepness at which sinusoidal waves in deep water will break is S = 0.142.
2.1.23
surge
translational motion parallel to the horizontal longitudinal axis of a floating body caused
by an excitation or a displacement in the direction of the longitudinal axis from its
equilibrium position
2.1.24
sway
translational motion parallel to the horizontal transverse axis of a floating body caused by
an excitation or a displacement in the direction of the transverse axis from its equilibrium
position
2.1.25
water depth, h
<airy linear wave theory>
shallow water: h/λ ≤ 1⁄20, intermediate water: 1⁄20 < h/λ ≤ ½, deep water: h/λ > ½
NOTE λ is the wavelength defined as the horizontal distance between the consecutive up or down
zero crossing point of a wave measured orthogonally to the wave crest.
2.1.26
wave height, H
measure of the vertical distance from a wave trough to a wave crest
7
Marine Renewable Energy Guides
2.1.27
wave number, k
spatial analogy to frequency
2.1.28
wave period, T
time interval in seconds between the consecutive up or down zero crossing point of a wave
2.1.29
wave power
combined potential and kinematic energy of a sea state being transported with a speed
equal to the group velocity normal to a unit width of wave crest in deep water
ρg2
P= Te Hm2 [kW/m]
64π 0
2.1.30
windage
deflection of a floating body due to the force of the wind acting on the freeboard
2.1.31
yaw
translational motion about the vertical axis of a floating body caused by an excitation or
a displacement from its equilibrium position
2.2 Abbreviations
BDM Bayesian Directional Spectrum Estimation Method
CoB Centre of Buoyancy
CoG Centre of Gravity
DAQ Data Acquisition System
DoF Degrees of Freedom
EU European Union
FFT Fast Fourier Transform
IEA~OES International Energy Agency ~ Ocean Energy Systems
IWS Internal Water Surface
JONSWAP Joint North Sea Wave Project
LWL Load Water Line
NASA National Aeronautics and Space Administration
OWC Oscillating Water Column
PADIWA Package for Directional Wave Analysis
PTO Power Take-Off
RAO Response Amplitude Operator
RMS Root Mean Squared
SS Proprietary Mooring System
8
Tank Testing of Wave Energy Conversion Systems
9
Marine Renewable Energy Guides
h water depth
gT 2 2π h gT 2
λ wavelength, tanh or for deep water, h/λ > 0.5,
2π λ 2π
2π Hs
Ss significant steepness,
gTz2
P energy flux per metre wave crest (kW/m)
M mass of the body (kg)
Ma added mass of the body (kg)
Awp water plane area of the body (m2)
Iyy mass moment of inertia (kg m4)
Iyya added mass moment of inertia (kg m4)
Ic moment of inertia of a plane area (m4)
M + Ma
TH period at which body resonates in heave, 2π
ρgAwp
I yy + I yya
Tθ period at which body resonates in pitch, 2π
ρgIc
TD device natural period, inverse of the frequency of the peak of the power spectrum
kW 1,000 watts, measurement of power
MW 1,000,000 Watts
3 Purpose of testing
3.1.1 General
It is prudent to test wave energy converters (WECs) through increasing model scales since
physical trials provide a mechanism for risk reduction, both technical and financial. This
considered approach is achieved by progressively improving engineering and scientific
knowledge from the initial verification of the concept(s) to the demonstration of the
economics with large scale models. The rationale for this measured approach is to ensure
due diligence can be successfully performed on the technology and business plan whenever
private or public finance is sought to support further development and progression.
Technical output of a well constructed WEC testing programme can provide information
for all of the following objectives if applied in accordance with the recommendations
shown in Table 3:
10
Tank Testing of Wave Energy Conversion Systems
The decision as to which parameters should be addressed at each phase should be based
on practical considerations and previous experience, especially towards the reduction of
uncertainty in later phases.
EXAMPLE Mooring forces can be measured at small scale to produce guidelines on the system’s
capability, but not at the expense of omitting its monitoring during later phases when scale issues are
less of an issue.
Test programmes should, therefore, have overall specific objectives set for each phase,
into which additional investigation can be incorporated as required. Figure 3 shows
an example of a phase 1 performance schedule with the addition of some specific
examinations. These flow diagrams, or similar, should be produced to plan each test
campaign. The mantra for any test engineer should be:
3.1.2.1 General
The following description outlines each of the three phases covered in this document
and the rationale and purpose of the trials to perform in each of them.
This initial phase is subdivided into three stages, each offering a different perspective.
a) Concept
Most new device proposals incorporate some unproven concept or combination of ideas
that should be verified prior to more extensive testing. No performance evaluation is
required at this time so an idealized device configuration can be utilized.
11
12
Table 3 — Technical output of phases
dynamics
Objectives Concept validation Real sea Verify phase I Wave-wire
Performance variables performance performance
RAOs Dir sensitivity Active control
PTO and mooring characteristics Deployment Mooring forces
Diffraction and radiation. seaways Seakeeping
RAOs Survival and
Seakeeping Failure modes seakeeping
(where
applicable)
Measerand Motion, forces, waves, mooring tension Control strategy
Electrical supply
Specialize DoF (heave only) Short crest Storm seas (3 hr) Power take-off Survival forces Quick release Solo or small
seas bench test PTO and Salt corrosion connections array (upgrade to
2D Solo and multi Angled waves Finite regular generator Marine growth generating station)?
hull As required (as required) Permissions Service ops
Maths Methods Hydrodynamic, numerical frequency Finite waves Time domain response model and Array interaction Int. market
(Computer) domain to solve the model Applied control strategy Economic model projection for
Undamped linear equations of damping Naval architects design codes for hull, Electrical stab. devise sales
motion mooring and anchorage system
Multi freq
inputs Economic and business plan
WEC
Phase 1
SET-UP SEAKEEPING
(3 hrs@prototype)
Operation Safety
Wave Web effects Viscous Maths
D#3 D#∞
calibration damping model
Extreme seas
Linear Linear Finite
H = 10 mm SB1-5
F9
F1-f11+3 gain
IWS marker
PERFORMANCE
(40min@prototype)
Hull mass GM
properties PTO
damping
Static Dynamic High Low
Mooring
ZERO D#1 D#2 D#3 D#4 D#5 D#∞
Torus B Torus C
Quarter
(port) Type Attachment
Waves
HMRC SS
Monochromatic Panchromatic
Monochromatic Panchromatic SWL ½ Draft Torus head seas
H = 50 mm 7 Brets’ Bretshneider
Panchromatic Linear Inter Finite
Storm H = 10 mm H = 25 mm H = 50 mm B1-20
Storm
13
Tank Testing of Wave Energy Conversion Systems
Marine Renewable Energy Guides
On successful confirmation that the physical processes work the model should be subject
to a limited number (5) of reference irregular seaways to corroborate operation and gain
insight into the important parameter to vary in proceeding stages, rather than the actual
settings required.
b) Performance
The primary objective of the second section of the test programme is to investigate
the device variables and physical properties that affect the performance, or energy
absorption. They should be regarded as the opportunity to quickly and cost-effectively
establish the working principles of the device and establish the design variables that
affect performance.
Initial testing will be conducted in regular, linear waves but include finite waves to
evaluate higher order influences on the behaviour. These results will also be used for
mathematical model calibration and validation. Depending on the device and the facility
it is possible that these tests can be conducted in a flume. However, blockage effects and
wall reflections should be factored into this decision including how they influence the
post-processing and data analysis.
Fully two dimensional testing should only be run for the mathematical model verification.
Once the underlying important physical variables have been identified a selection of
the key design parameters should be tested in irregular seaways. Since these are still for
comparison purposes classical spectral profiles can be adopted. It is more important to
cover a range of sea states representative of perspective deployment sites scatter diagram
(i.e. open ocean, confined sea, shallow water etc.). For most near-shore applications the
Bretschneider spectrum is recommended.
Diffraction and radiation trials can be included in this stage (see 3.1.4.2).
It is not essential to know the final mooring arrangement at this time. A system that has
minimal influence on the body motion can be considered. Initially this can even be a
securing rig that reduces the degrees of freedom of body motion.
c) Optimization
Based on the previous two stages an optimized version of the device should be
evaluated over a range of extended excitation conditions. The basic design, especially
the hull components, should have been reviewed by a naval architect group or structural
engineers to confirm the design is acceptable, is manufacturable and that it could
be certified as fit for purpose by the appropriate authorities. Any adjustments or refit
resulting from this technical due diligence should be incorporated in the model, even
if this dictates a new model being constructed. The design should first be verified by
rerunning to the previous stage monochromatic tests.
14
Tank Testing of Wave Energy Conversion Systems
• mooring options.
Once the fundamentals have been established in phase 1 a larger model can be used
to investigate other parameters. These trials will be conducted in irregular wave fields,
which can be either generic or site specific. Of particular merit will be the generation
of storm conditions to establish the seakeeping characteristics of the hull or structure.
If an advanced power take-off simulator is fitted, PTO end stop issues and component
survival may also be considered. The finalized mooring should be included by this
time so main failure modes can be experienced. Though not essential a facility offering
current and wind effects could be considered, depending on the expected influence of
these environmental conditions on the device. If a model at the larger end of the scale
range is selected this phase of testing could provide an acceptable platform for control
equipment so different strategies can be investigated.
By phase 2 the model should have advanced from the idealized version of phase 1 into
a unit resembling the wet sections of the proposed WEC in most detail. An important
output from the trials is the verification of the previous phase results and the associated
decisions made from those data.
Ideally phase 2 trials should be conducted in a three dimensional basin but if there is
a requirement for in-depth investigations of specific criteria, then a wide flume could
be selected. An obvious example of this would be an active control system fitted to
the power take-off simulator. Once the fundamental algorithms have been established
and short crested seas are required, even these investigations should be concluded in a
wide basin. Mooring trials, and especially failure modes, should be run in basins that can
accommodate the full mooring line spread footprint.
Typically medium scales also enable improved sensor selection since models are large
enough to accommodate many proprietary transducers. This is particularly the case with
submersible gauges which are often too heavy for phase 1 models.
15
Marine Renewable Energy Guides
Although large models can still be tested in (a limited number of ) flumes or basins it is
anticipated this phase will be conducted at a benign external site. To assist in this process
sea trial locations have now been established in some countries. In Europe, both Denmark
and Ireland have a pre-designated location where deployment permits and licences have
already been obtained. This is a most important phase so should be approached carefully
and extensively. A typical scale for this work would be circa λ = 1 : 4, which for megawatt
size WECs will be a large physical device of several tonnes. As such the model will include
all the equipment necessary to be a fully operational unit, including PTO, generation
and power electronics. The developer will therefore have the opportunity not only to
test the complete system but encounter the difficulties of bringing together a disparate
disciplinary team and multi-technology device for the first time. This factor should not
be underestimated.
Sub-megawatt units (< 500 kW) might not benefit fully from this phase, though evidence
has shown a ½ or 1⁄3 scale units of 10 kW can provide realistic and valuable data prior to
full size prototype build.
Of particular merit for benign site trials is the opportunity to study devices close to shore
and where scaled 50 year storm conditions are Hs ≤ 5 m. It is probable that once deployed
at an open ocean power park site the behaviour of WECs in extreme waves will never be
visually observed since it will be unsafe for humans to be in the vicinity at such times.
3.1.3.1 General
3.1.3.2 to 3.1.3.4 describe some of the factors that particularly require consideration when
deciding a model scale or interpreting the results from tests.
In pneumatic devices the stiffness of the air spring is not scaled correctly by geometry.
For static devices this can be compensated for by the use of an external air reservoir. In
mobile buoyant units this is more difficult due to the practicality of not influencing the
vessel motion. The degree to which this parameter will affect any test result depends on
the level of applied PTO damping.
16
Tank Testing of Wave Energy Conversion Systems
Little empirical evidence yet exists to quantify the difference in results due to this
phenomenon since few full scale devices have been deployed to provide data for scale
comparisons. Theoretical studies suggest that if compressibility is to be an issue it will
only be encountered at prototype (full) scale.
3.1.3.3 Viscosity
Model testing, particularly at a small scale, tends to produce conservative results. Viscosity,
and in particular vortex shedding from edges, does not scale appropriately and can be
overestimated during physical testing. Attention to detail during model construction can
minimize this effect but the best solution is to increase scale when investigating critical
parameters.
This factor often leads to difficulty matching physical and theoretical mathematical model
results (see 3.2.2). Empirical results can be used to calibrate numerical models but, since
this is still an art rather than a science, care and attention to detail should be applied.
Experience in the offshore engineering industry suggests that floating structures exhibit
little viscous effects so coefficients are set to unity in mathematical models. However,
these are usually very large bodies designed to move as little as possible. Conversely
wave energy devices are relatively small (<1,000 t) and experience large excursions in at
least one degree of freedom. They also have the added complexity of a power take-off
system that feeds back into the modes of motion so the unity approximation might not
be valid.
3.1.3.4 Friction
Mechanical friction should also be carefully dealt with during model construction since
it too does not scale when Froude similitude is imposed. This is particularly important
during phase 1 and phase 2 of the programme. Power take-off simulators can be a source
of problems if friction is not minimized. Forces experienced at small scale will only be a
few Newtons and even well engineered PTOs can have resistance such that losses are
proportionally high percentages of absorbed power.
A very careful selection of simulator is required. Low friction dampers can alleviate some
of these uncertainties.
3.1.4.1 General
When developing the device test programme there will be a requirement for specific
tests designed to address other important issues affecting the performance of the WECs.
Two examples of these are given below.
17
Marine Renewable Energy Guides
• Radiation: the device is held by a motor containing a force transducer. In still water
the device is oscillated by the motor at different periods and the corresponding force
measured. The radiated wave might also be monitored for use in the theoretical study.
At what stage this matter should be addressed has not been clearly defined by either
academics or from developers’ previous practical experiences. For multiple small
individual WECs clustered in close proximity to each other (d ≈ D) on a single platform
the situation is clearer. These should be considered as a solo unit with the model scale
adjusted to suit the practicalities of such a single conversion unit.
For device wave parks some fundamental studies are now underway which should
identify the importance of unit interactions in a qualitative way. Based on these studies a
recommendation will be included in future editions of this document.
3.2.1 General
Monochromatic or regular single frequency waves play an important part during the
early stage of a WECs development programme. They are used to provide an indication
of how a device works. The three fundamental applications of results from this sort of
testing are:
2. to observe and monitor device response to regular excitation forces that define
the basic operation of a device;
18
Tank Testing of Wave Energy Conversion Systems
Regular waves are defined by their wave period (wavelength), amplitude (height) and
surface profile. Figure 4 shows these relationships.
Table 4 provides some basic physical definitions for gravity sinusoidal waves based on
linear theory.
Water depth
Parameter Deep – Intermediate – Shallow –
h>λ λ >h<λ h<λ
2 2 20 20
Wavelength, λ gT 2 gT 2
2π 2π
( )
tanh kh
T gh
Wave celerity, c gT gT
2π 2π
( )
tanh kh gh
19
Marine Renewable Energy Guides
Both frequency domain and time domain numerical models require either validation or
calibration by empirical data before they can be confidently applied as a WEC design tool.
This verification procedure can best be achieved by comparison with practical results
produced from scaled physical models and in particular those obtained during regular,
single frequency plane wave trials.
Advanced frequency domain models such as WAMIT, AQWA, Aquadyne etc., take as
their input monochromatic oscillation so for direct response comparisons regular wave
excitations are preferable. Time domain models can operate with regular or irregular input
signals but are best proven in the simpler case of monochromatic waves. Figure 5 shows
a comparison of theoretical and practical data. As expected the mathematical model
predicts higher response around resonance and better agreement away from the natural
period. These results can be used to adjust the hydrodynamic damping coefficients in the
mathematical model. When external damping, such as a PTO, is imposed the peaks are
reduced in both cases such that the results are usually closer without modification.
7
RAO [motion/wave]
0
0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7
Mono wave period (T s)
The physics processes controlling the behaviour and performance of WECs are not always
fully appreciated or initially known. Tests to monitor the device response in regular waves
can reveal many elements of the behaviour and, in particular, a change in performance
due to design variable refits. Regular excitation is also the traditional method that naval
architects employ to produce the response amplitude operators (RAOs) for floating,
buoyant bodies. Figure 6 shows the particularly interesting case when two hulls with
different natural periods are in close proximity. Tests were conducted on solo and
20
Tank Testing of Wave Energy Conversion Systems
combined, but uncoupled hulls (zero). The results indicate the independence of the
system. When connected with infinite damping the system is forced to respond differently
as shown by the black graph line (infinite). At some point between these two extremes
the optimal power extraction will be achieved.
4.5 90
4 80
3 60
2.5 50
2 40
1.5 30
1 20
0.5 10
0 0
0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2
Mono wave period (s)
Figure 7 displays the power performance of a WEC with respect to the PTO damping for
several different device configurations. The optimal level can quickly be located for use in
later, more complex trials.
1.14
D#5
1.23
Wav
D#4
1.33
e per
1.45
D#3
iod,
1.6
T (
D#2
1.78
s)
in
D#1
2
p
am
D
21
Marine Renewable Energy Guides
As previously stated most numerical models, certainly those in the frequency domain, are
based on linear wave theory. Small amplitude waves are therefore used when providing
data for the mathematical model validation. Higher order effects can also be studied in
regular waves by increasing the wave height and monitoring the physical parameters’
corresponding response. Figure 8 shows a typical buoyant body’s heave RAO at different
wave heights.
It should be noted that the graph is the magnification ratio not the absolute motion of
the body which will produce a different relationship.
4
RAO
0
0 10 20 30 40 50 60 70
Wave height
3.3.1 General
Whilst regular, monochromatic trials can be regarded as a method to learn how a device
works, irregular, panchromatic trials begin to reveal how the device will perform.
22
Tank Testing of Wave Energy Conversion Systems
• type of spectrum;
• directionality;
• spreading.
The discussions on these seaway options depends on several factors and in particular
the phase of the device development. The following description includes some
recommendations of best practices that have evolved over recent years.
3.3.2.1 General
The main consideration at an early stage of testing is between the use of generic,
classical spectra or site specific data. Primarily phase 1 trials are designed to compare
the key design variables identified in the monochromatic stage. It is viable and valid
therefore to adopt classical spectra, which most wave generation systems are capable of
producing quickly, accurately and consistently.
It is also probable that the wave energy extraction device’s final deployment site has not
been established at this time. Indeed the unit could be under consideration for several
locations. This more generic approach is required to account for all possibilities. Finally,
site specific records may not be available at all, or limited in duration and mix.
There are several classical spectral equations used to describe the frequency composition
of an irregular water surface elevation time history. Some have their origin based in wind
wave energy transfer processes whilst others depict the relationship with respect to the
summary statistics of a particular seaway (i.e. Hs, Tp, Te, Tz etc.). It is the latter designations
that are of most interest to wave energy device developers since the sea state generation
mechanism is of secondary importance.
23
Marine Renewable Energy Guides
Each of these seaway types are related and all produce the same wave conditions under
specific circumstances. They are back-traceable such that a Bretschneider can describe
a Pierson–Moskowitz and a JONSWAP can represent both a Pierson–Moskowitz and a
Bretschneider.
This would infer that JONSWAP could be used as the base formula to accommodate all
cases of classical testing. Two caveats can prevent this being the case. Firstly, the wave
tank generation system may not be capable of producing sufficient variants of the
spectrum, i.e. it is applied to a Pierson–Moskowitz equation rather than a Bretschneider.
Secondly, insufficient information is available on which to design the spectral shape so
default settings are used which then become Bretschneider spectra anyway. Figure 9
shows the relationship between the three options with the same Tp, but not for the same
energy content so the shapes can be identified.
80
70 Bretschneider
Piersen-Moskowitz
60
Spectral ordinate (m2/Hz)
JONSWAP(PM)
50
40
30
20
10
0
0 0.05 0.1 0.15 0.2 0.25 0.3
Frequency (Hz)
As testing moves into the latter part of phase 1 and certainly by phase 2 it is preferable
and advantageous if actual site data, from at least one, primary deployment location, is
used to evaluate the device performance and seaworthiness in these conditions.
By this time most principle device design options should have been finalized but their
influence under different excitation scenarios need to be established. The importance of
these trials obviously varies, depending on how much any particular station conditions
vary from the classical shapes. In near-shore North Atlantic resource monitoring projects
evidence suggests that by weekly average, and longer time periods, spectra begin
to conform to standard conditions as shown on Figure 10. Individual seaways can be
significantly different but the occurrence should be limited if they do affect the average.
24
Tank Testing of Wave Energy Conversion Systems
40
S(f) (m^2/Hz)
30
20
10
0
0 5 10 15 20 25 30
Period (s)
[email protected] 19/4/09
The industrial standard length of time a seaway should be monitored to obtain accurate
harmonic components of the time series is 20–30 minutes at full scale (i.e. approximately
150–200 waves). This can be adopted as the reference duration to apply for generating
the spectra in tanks. The figure is based on two factors:
These criteria can be applied to most fundamental tank testing investigations considering
a device’s average performance and behaviour. Analysis of wave time histories derived
from this approach should reveal that the correct Rayleigh type distribution of wave
height is present, as shown in Figure 11. Test facilities (including benign sites) should also
verify conformity to the conditions prior to testing.
Recent resource studies have shown that storms can track much faster than this
development criteria so site specific tests might require temporal modification.
25
Marine Renewable Energy Guides
0.7
0.6
0.5
P( 2H/Hs )
0.4
0.3
0.2
0.1
0
0-0.2
0.21-0.4
0.41-0.6
0.61-0.8
0.81-1.0
1.01-1.2
1.21-1.4
1.41-1.6
1.61-1.8
1.81-2.0
2.01-2.2
2.21-2.4
2.41-2.6
2.61-2.8
2.81-3.0
3.01-3.2
3.21-3.4
3.41-3.6
3.61-3.8
3.81-4.0
2H
Hs
Having defined the spectral shape(s) and duration appropriate for a series of tests the
number of seaway to apply should be decided. This will be determined by the stretch
of water in which the device will eventually be deployed (i.e. Atlantic, North Sea, Baltic,
Southern Ocean etc.). The location defines the number of different height and period
combinations that can occur. It is also a function of the scatter table step size used to
produce the bivariate array. Experience has shown that a 1 m by 1 s element is sufficient
to accurately determine the performance characteristic of a wave energy converter up to
phase 3.
3.3.5 Directionality
3.3.5.1 General
There are two considerations regarding the question of wave approach angles and
their importance depends on the sensitivity of individual devices to each. It would be
expected that point absorbing WECs behave independently of both factors but this
might not be the case when clusters or arrays of these devices are considered. Similarly,
mooring effects will be influenced by wave directionality.
26
Tank Testing of Wave Energy Conversion Systems
This is applicable for monochromatic waves and long crested irregular seas. It describes
the mean angle at which the wave front impinges on the WEC. Axi-symmetric point
absorbers (i.e. < 1⁄3 wavelength) would usually remain robust against this parameter but
mooring arrangements will have an influence on the tests. This in turn will affect motions.
Some trials at different mean approach directions should therefore be accommodated
in the programme, especially bow, beam, quarter, following and a special case directly
along a mooring line.
This two dimensional aspect of seaways is less well described than the traditional one
dimensional spectral shape. No agreed standard has yet been specified to cover general
cases, mainly because there have been limited measurements of crest length criteria to date.
As with directionality the effect of this seaway parameter will vary, depending on the
WEC under investigation. The best practice evolved to date is to generate an extreme
spreading, for example a cosine squared (cos2) relationship which produces very short
crest lengths, and monitor the difference in performance indicator between this and the
long crested case.
4.1.1 General
In all the phases of a test programme the device’s ability to capture and convert the wave
energy can be regarded as an important criteria. Programmes should be constructed
around investigating the device design variables to optimize power production. Some
practical aspects of this approach have been described in Clause 3. Manufacturing,
materials, deployment and connectivity matters etc., should also be considered in
economic reviews but these are somewhat more ephemeral criteria that can change over
time whist the fundamental energy conversion will not.
4.1.2.1 General
27
Marine Renewable Energy Guides
In phase 1 and the early part of phase 2, the investigations can be based on monochromatic
wave excitation such that the effect of one individual design parameter can be assessed
relative to itself. For the same input power two variants of the same device might display
very different output power. Figure 12 shows this effect whereby the unit with the increased
heave motion produces much less power, probably due to phase considerations.
OE BUOY PROJECT MARINE INSTITUTE
0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2
In higher waves it is usual that more power will be produced but as shown earlier in 3.2.4
the efficiency of a device will most probably decrease with increasing wave height. This is
shown in Figure 13.
The input power of these monochromatic waves is derived from the equations defined in
Table 4, i.e.
ρg2 H 2T
Power = (W/m) ≈ H2T (kW/m)
32π
The amount of energy flux actually available to a wave energy converter can be expressed
in several ways. Two common and recommended approaches are defined by:
a) Hull width
In this instance the power per unit length of wave front (W/m) is multiplied by an
appropriate length scale taken from the device to define the total power (W) input to
the WEC.
28
Tank Testing of Wave Energy Conversion Systems
b) Capture width
An alternative definition of the power available to drive the wave energy converter
derived as the quotient of the machine power and the wave energy flux.
Device power (W)
CW = (m)
Wave power/unit length (W/m)
A further calculation can be used to define the capture width as a ratio with the device
width perpendicular to the wave front.
CW CW: capture width
ratio =
DL DL: longitudinal dimension
NOTE Since the wave power equation produces the average power/wave cycle it is recommended that
the average converted power is also used.
0.0
Power / wave height (W/mm^2)
0.0
0.0
0.0
0.0
0.0
0.0
0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2
Mono wave period (seconds)
Having defined the device variables of interest the performance trials can progress to
real seaways. The types of spectral shapes to be included have been described in 3.3 and
can be selected from classical or measured data. Whichever the source, the important
factor is to cover a spread of conditions such that the annual power production figure
can be extrapolated.
For energy capture trials it is conventional that the test duration is a scaled equivalent
of 20–30 minutes at prototype scale.
29
Marine Renewable Energy Guides
A similar formula and approach to that used for the regular wave definition can be
implemented to obtain the wave power in real seas. These equations, however, are based
on the measurement of the specific time period above which the stationary summary of
that seaway are obtained, rather than a single cycle.
It is not necessary to test all of the scatter diagram elements and the offshore industry
recommended practice is to select 15–20 of these Hs-Tz combinations such that the
overall behaviour of a structure can be determined. This policy is adopted for evaluating
the performance potential of a WEC across the various seas it will encounter. Device
engineers can then use this information to finalize design parameters to optimize annual
output of a unit. Figure 14 shows an example of tailored seaways for any selected test site
with the appropriate wave combinations marked. It will be noted that lines of constant
significant wave height and average/peak period are included. This allows extrapolation
graphs to be drawn up such that, if required, all elements of a bivariate table can be
computed. An example of how the limited test data can be extended to seaways not
included in the test programme is shown in Figure 15. As can be seen from this graph the
changes between sea states becomes a deterministic relationship with no discontinuities
existing in the curves. This enhances confidence in the extrapolation procedure.
During phase 3, input conditions are no longer controllable and should be accepted
as they occur naturally. This imposes a greater requirement for organization and rigour
when specifying test programmes. It is not sufficient just to obtain the seaway summary
statistics since most WECs are resonant oscillators which respond in the frequency
domain. The full spectrum should be obtained from surface elevation time series such
that performance values can be related to known similar input conditions. Figure 16
shows an example of measured power performance for seas exhibiting the same statistics
but with significantly different spectral shapes. In this case a wave energy device would
not generate similar power levels for this sea state due to resonance characteristics.
30
Tank Testing of Wave Energy Conversion Systems
90
80
70
Average power (kW)
60
50
40
30
20
10
0
0 1 2 3 4 5 6 7 8
Significant wave height (m)
31
Marine Renewable Energy Guides
45000
40000
35000
S(f) (cm^2/Hz)
30000
25000
20000
15000
10000
5000
0
0 2 4 6 8 10 12 14 16 18 20
Period (s)
During phase 1 tank testing PTO simulators are employed. It is sufficient for the mechanism
to be adjustable at stepped values which apply external damping to the relative motion
between the WEC’s inertial masses. When equipped with a force (or pressure) transducer
and the corresponding body motions are measured, the instantaneous absorbed power
can be calculated for each specified damping setting, since for example:
As shown in Figure 17 the two parameters can be graphed to establish that a representative
simulator has been selected. These trials produce the maximum power performance since
the PTO has no constraints, such as end stop, or decoupling, that will be presented in a real
PTO. They are to provide the environmental load data, including extremes, from which the
PTO rating can be derived.
Phase 2 trials are fundamentally to expose the device to sea conditions likely to be
encountered by the prototype machines. However, because the model is physically larger
a more sophisticated PTO simulator can be incorporated. This is not essential at this
time but can be regarded as advantageous to progress. The primary requirement for all
configurations is that the characteristics of the final PTO are reproduced. This statement
assumes the PTO is known, which might not be the case. Experience has shown that
unless a very complex type system is to be used a high quality energy dissipater provides
a generic representation suitable for most options. The exception to this general rule is
if the PTO will include a control feedback option that attempts to tune a device to the
incoming waves as a time domain system. For such experiments a more complex PTO
model is required.
32
Tank Testing of Wave Energy Conversion Systems
Monochromatic
F05 (H = 54.05 mm T = 1.33 s)
Extension vs Force
4
1
Force (N)
-1
-2
-3
-4
-20 -15 -10 -5 0 5 10 15 20
Extension (mm)
Monochromatic
F05 (H = 50mm)
Force vs Velocity
4
1
Force (N)
-1
-2
-3
-4
-100 -75 -50 -25 0 25 50 75 100
Velocity (mm/s)
Work over many years at Edinburgh University has concentrated on developing electro-
magnetic motors that can be adjusted in real time by a predetermined source feedback
signal to facilitate active control. Although they can operate at the small phase 1 scale the
advantages off conducting this work on a medium scale phase 2 model are recommended.
By phase 3 sea trials, it would be expected that an operational PTO and generator
combination is under investigation rather than the secondary conversion system. A scaled
design of either a pneumatic air turbine, hydraulic or direct drive generator should be
33
Marine Renewable Energy Guides
incorporated. This might have to be specifically designed for the large scale application,
rather than direct copies of the final full size prototype PTO, but the characteristics will
be similar. This ensures results from the trials can be confidently extrapolated. Even at
this large scale electricity output levels are small due to the Froude scale factor for power.
This limited level makes connection to the grid a moot point which can be decided by
convenience, cost or public relations issues.
The calculation below is an example of this power scale issue for a 1 : 4 design.
It should not be regarded as a requirement for phase 3 sea trials that a link to the local
distribution network is required. The restricted power rating enables resistive load banks
to be used to dissipate the energy, even the high peaks experienced in irregular seas.
Technically, even if the connection was made, it can not be regarded as addressing
multi-megawatt prototype grid issues. Other power electronic options, however, do
scale sufficiently for the developer to investigate supply matters with the aim to improve
the quality of the electricity output. It would be recommended that this work should
be studied on dry test rigs prior to sea trial deployment if possible. However, such test
programmes are beyond the scope of wave tank trials.
NOTE Established large scale test site owners may be encouraged to provide a power cable since
multiple use could validate the expense of connection.
4.2.1 General
Before undertaking phase 3 sea trials it is important to conduct specific trials to establish
the seaworthiness of the WEC, both for the hull(s) and the mooring. These tests are to
provide the extreme motions and loads exerted on the hull, power take-off, mooring
lines, anchors and foundations for fixed or gravity structures.
Initial investigations can be undertaken at the small scale but it is highly recommended
that the scenarios studied are conducted using a medium scale model in an appropriate
sized basin.
Sections of breakwaters, or other fixed type devices, can still be studied in 2D flumes but
the medium scale would still be recommended.
An important consideration when developing the survival test programme is that most
simulation test facilities have a limitation on the size of waves (sea states) that can be
34
Tank Testing of Wave Energy Conversion Systems
produced. Although the conditions that will combine to produce the extreme loads
are not well defined some trials will be conducted in high energy seas approaching
50–100 year storms. This means the scale of the model used for these tests might require
adjustment from the energy capture model. An alternative can be to a use small scale
model in the medium size facility as a compromise.
Short crested seas and directional wave approach should be included in the overall
programme for each of the test types described below.
The selection of sea states will usually lie between the fully exposed limit (Ss = 1 : 10) and
the fully developed sea line (Ss = 1 : 20).
As stated in 3.1 these trials can be conducted at small scale providing the test facility can
generate high seaways. These need not necessarily be only the highest energy seas but also
steep waves in shorter period seas where the breaking limit is being approached. Devices
should be observed under operation conditions with the PTO working and also failed
scenarios modelled with both zero damping and infinite damping. Individual trials should
be conducted over an extended test period. As suggested above the recommendation is
a scaled equivalent of a 3 hour storm such that the wave train series will contain sufficient
individual waves to produce an acceptable Hm0/Hmax ratio (i.e. 1 : 1.8–2).
It still remains a debated topic as to which wave group combination produce the worst
loading condition on a structure, mooring or foundation. This aspect might also be a
device specific issue so observation made during the seakeeping section, phase 2, should
provide pointers to specific test cases requiring review. For floating devices it is possible
that periods close to one of the resonance responses will create maximum loads in the
hull structure and moorings.
NOTE The recommendation would be that investigations into this area are conducted at the
intermediate scale (circa λ = 1 : 15) though indication can be inferred during small scale trials.
Knowledge of the device fundamentals will assist in the design of tests aimed at
monitoring and measuring response to irregular wave excitation. This series of trials will
be an extension of the seakeeping programme and can be combined with it. Tests can
be constructed along the wave breaking limit (Ss = 1 : 10) and include seaways based on
each of the six degrees of motion (6 degrees of freedom (DoF)) natural periods. Angled
seas should be included or the model (and mooring) rotated to ensure all motions are
observed. Short crested seaways should also be used since, although these conditions
tend not to create maxima, they do excite all the motions simultaneously.
35
Marine Renewable Energy Guides
Identifying the worst seaway in terms of motion and/or loads of the device can only
be achieved by testing across a range of conditions and/or spectral shapes. The
combinations for these are extensive therefore a systematic sweep approach should be
adopted. The details for selecting bins from a wave scatter diagram was covered in the
previous section so in this instance selected Hm0–T02 combinations based on a classical
Bretschneider spectrum should be used to attack the model. From the worst cases a set
of adjusted JONSWAP offering increasing peak factors will reveal the sensitivity of the
device to wave steepness. Twin peaked and bimodal seas can contribute to these and the
previous survival tests.
5 Test equipment
5.1.1 General
As well as being correct in geometric scale, physical models should be exact dynamic
replicas of the prototype device they represent. This is achieved by following fixed scaling
laws based on Froude similitude criteria and similarity conditions.
A full description of the relationships can be found in Clause 6 of this document but
basically all physical properties can be related via inertia scaling such that:
For static, or fixed station WECs, only the device geometry need be scaled. For buoyant,
floating WECs the key parameters are the mass distribution and associated moments.
These dictate the response periods to wave excitation for each of the 6 DoF.
Small and medium sized models (phases 1 and 2) are rarely manufactured from prototype
materials. The customary design principle is therefore to construct a model that is lighter
than the prototype such that lead ballast can be added at strategic locations to produce
the required dynamic characteristics.
A location within the frame of a floating body which represents the centre of mass of the
water displaced by the submerged portion of that body. When mass and hull dimensions
36
Tank Testing of Wave Energy Conversion Systems
are correct the centre of buoyancy (CoB) will automatically adjust. The exception to this,
which is relevant to WECs, is when the body walls are proportionately thick and water is
present on both sides. This produces an increased displaced volume and thus a reduced
draught. Models should be heavier in air to compensate, which is obviously a compromise.
This is a point within the frame of a floating body where the total mass of the body
can be represented by a point mass. The placement of mass about the body structure
influences the position of the centre of gravity (CoG).
This is the distribution of the mass of an object relative to a given axis. It is a measure of
an object’s resistance to changes in its rotational rate, and is calculated as the product of
the distance from the rotational axis squared and its mass. The units are mass × area.
i
I jj = ∑ m jδ 2j (kg m2 )
1
The second moment of area of the water plane is particularly important for floating
bodies. Also known as the moment of inertia of a plane area, this is a measure of the
restoring moment and is given by the area multiplied by the distance to the rotational
axis squared. It has units of metres raised to the fourth power. It is a similar property to
the mass moments but based on distributed area.
i
I j = ∑ a jδ 2j (m4 )
1
This is the area on the water surface broken by a protruding body (m2). If it is assumed that
there is no significant change in the water plane area during heave motion, the ship is said
to be side walled at the load water line (LWL) and this parameter can then be expressed as
a block coefficient, which is the ratio of the actual area and the bounding rectangle.
5.1.2.6 Metacentre
This is the point at which a vertical line through a tilted centre of buoyancy and centre of
gravity crosses the line through the non-tilted conditions.
37
Marine Renewable Energy Guides
This point can also be estimated from the quotient of the second moment of area about
the water plane and the displaced volume through the equation:
Ij
BM = (m)
∇
This is the distance from the centre of gravity to the metacentre. GM represents the static
stability of a floating body and it is depicted schematically in Figure 18.
GM = BM − BG
GM: distance from the CoG to the metacentre along the vertical axis
BM: distance from the CoB to the metacentre along the vertical axis
BG: distance from the CoB to the CoG along the vertical axis
This is the change of hull mass and inertia caused by a volume of water that is required
to move with the hull. These values are frequency dependent and can either be obtained
from mathematical models or empirically by the following series of model validation
procedures.
Following construction the mass properties of the model should be confirmed by various
standard balance and suspension practices.
38
Tank Testing of Wave Energy Conversion Systems
i) Firstly by placing the model on a sharp edge until it is balanced. This is repeated for
each of the axes the model shape accommodates.
ii) For models that will not facilitate this approach a suspension method can be adopted.
For an irregular hull shape if the longitudinal CoG has already been located by the
balance method a plum line dropped from the hull suspension point crosses this plane
at the vertical centre of gravity.
iii) The hull is suspended horizontally from points on a particular plane and a small weight
can be added off centre of that axis to induce a trim or heel. Accompanying equations
then produce the location of the centre of gravity whose axis is perpendicular at the
suspended axis.
The three dependent mass moments of the inertia are verified separately by the bifilar
suspension method. Two long suspension lines are attached to the model about the
axis of interest. The model is then caused to rotate about that axis and the oscillatory
frequency recorded. Associated formulas permit the mass moment of inertia in that plane
to be calculated.
Following on from the dry rig, water trials can be used to further validate the model
configuration. For these an appropriate moving mass is located on, or vertically above,
the horizontal centre of gravity. By displacing this mass along a principle longitudinal or
transverse axis the device is inclined at angles, which are measured. A plot of these settings,
relative to the applied moment, as shown in Figure 19, reveals the metacentric height (GM).
Disp. lightship = kg
Inclination test
Moving mass = g
Displ. X GM = gm
Hull Trim GM = 70.5 mm
GM Lightship = mm
8
6
4
4
3
Angle (deg)
0 0
-2
1
-4 2
-6
-8
-40 -30 -20 -10 0 10 20 30 40
Torque (g m)
39
Marine Renewable Energy Guides
• Heave resonance:
M: mass of the floating body
M + Ma
TH = 2π Ma: the added mass of the floating body
ρgAwp Awp: water plane area
I yy + I yy' I + I'
Tθ = 2π ; Tθ = 2π xx xx
∆ GMg ∆GMg
These calculated natural periods should then be checked dynamically in water. The
hydrodynamically restored degrees of freedom (heave, pitch and roll) are evaluated
directly by both free oscillation and forced oscillation trials.
a) Free oscillation
This reveals the eigen period of the parameter under investigation. The hull is given an
initial displacement and allowed to return naturally to its still water equilibrium position.
A typical trace of the resulting free oscillations is shown in Figure 20. The period of
oscillation for each mode is the resonant frequency for that degree of freedom.
Prototype scale
Pitch free oscillation ∆T = 9.28 s
Oscillation01 Peaks
4
2
y = 4.342e-0.0627x
Displacement (m)
-1
-2
-3
-4
0 5 10 15 20 25 30 35 40 45 50 55 60
Time (s)
These tests also reveal the amount of hydrodynamic damping present. This is shown by
the reduction of the motion with time. The exponential decay is the curve the oscillation
peaks trace, the dotted line in Figure 20. This damping is a combination of the primary
40
Tank Testing of Wave Energy Conversion Systems
component, wave-making, and any viscous forces present. The latter is usually in the
form of vortex shedding and can be proportionally high in poorly constructed models.
b) Forced oscillation
These trials are the results of waves being used to excite the device. Initially monochro-
matic waves are used but panchromatic seas can also be engaged. Usually irregular
waves are for specific purposes or during phase 3 sea trials when there is no alternative.
Primary characteristics investigated by this method produce the RAOs of the device.
Response amplitude operators: RAOs are the response of the body relative to the wave
excitation. A range of single frequency sinusoidal waves are independently impacted
on the device and the responses (motion, force, power, etc.) to each are measured and
plotted as the ratio with the wave, as shown Figure 21. This represents the fundamental
characteristic of the resonator to harmonic excitation, as will be present in ocean surface
waves. From the RAO, the behaviour of the device in any irregular sea can be predicted.
The peak period on the graphs should coincide with the natural period measured by the
previously described methods.
A similar plot can be generated from real seas by spectrally analysing the irregular time
history. A typical example is also shown in Figure 21 where the heave response of a
device to the frequency components is plotted and compared to the regular wave RAO.
One advantage of the irregular wave technique can be that frequencies outside those
tested in the monochromatic trials exist as harmonic components in the spectral analysis
and so the frequency range is extended. However, care should be taken that large peaks
in the RAO in this extended range are not a function of noise contamination and very
small number division.
1.8
1.6
1.4
1.2
Heave RAO (m/m)
0.8
0.6
0.4
0.2
0
4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Period (s)
[email protected] 19/4/09
41
Marine Renewable Energy Guides
The regular wave trials can also be used to produce the phase relation graph, as shown in
Figure 22. This procedure is described in more detail in Clause 6. Resonance occurs when
the phase shift between excitation and response is at 90°.
When the device is moored the missing 3 DoF can be included in the investigation to
reveal the second order motion caused by the restoring of the mooring. This information
is essential for the mooring design consultants.
There are many variables that can be consulted through the RAOs although more
generally these would be the motions.
42
Tank Testing of Wave Energy Conversion Systems
However, other properties of the device can be characterized by this methodology, such
as pneumatic power, force, pressure, etc. By following this approach the characteristics
affecting the device performance can be isolated and optimized design parameters located.
5.1.3 Mooring
It is important that the station keeping system is also scaled correctly. The practicality of
this depends on the complexity of the proposed mooring system. For simple catenary
arrangements it is usually possible to directly scale the mass/unit length of the prototype,
which automatically produces the correct load extension curves for a given design. A
typical catenary characteristic is shown in Figure 23.
OE BUOY PROJECT S.E.I. OE/03/006
50
Mooring force (N)
40
30
20
10
0
0 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400
Displacement (mm)
[email protected] ECN Nantes, September 2004 Printed: 28/5/09 SNB
The fidelity of the modelling should be verified by either hanging weights on the
mooring or, if a force transducer is fitted in the line, displacing the model and recording
the excursion and corresponding line tension.
When the elastic properties of the mooring lines play a roll in the system a trial and error
process of selecting appropriate material should follow. As above, the acceptability of the
solution is verified by an empirical load extension similarity.
If buoys are included in the proposed mooring these too should be scaled accordingly.
For these units it is the hydrodynamic stiffness, rather than the mechanical stiffness that
is important. This can be achieved by geometry and mass similarity.
43
Marine Renewable Energy Guides
Until phase 3, anchors can be functional clump weights. Even at the large scale the
methodology of scaling the overall system remains the same, as described above.
5.1.4.1 General
The materials used for the fabrication of the device will be a function of which phase of
testing that is under consideration.
As mentioned earlier small and medium scale models are rarely constructed of prototype
materials. Because of this components are not failure tested but rather forces that will be
encountered and can be measured for use in full size design integration.
Light alloys, fibreglass and various plastics form the main elements of models and should
be selected for suitability of purpose. Acrylics (perspex) can be of particular interest since
they are close to the density of water and therefore almost neutrally buoyant. They are
also available in transparent forms so useful viewing ports can be incorporated in models
to facilitate observation of inertial behaviour, such as inside an oscillating water column
plenum chamber.
Phase 3 models will probably be in prototype materials scaled accordingly. This is not
a trivial task, however, since geometrical scaling can lead to the material, such as steel,
being too thin for the required strength. Non-scaled materials, however, can produce
heavy models since mass is the cube of the linear scale (λ3).
Large individual components, such as the PTO and power electronics equipment, can be
an even greater problem for design engineers at quarter scale. However, experience at
the Nissum Bredning and Galway Bay benign test sites suggest that elegant solutions can
be achieved with engineering ingenuity. It should not be necessary to resort to exotic
lighter materials although these can always be considered.
44
Tank Testing of Wave Energy Conversion Systems
5.2.1 General
There are several types of hydraulic test tanks but the three primary ones for wave energy
use are:
• flumes (2D);
• basins (3D).
• benign site.
The size of the tank should be selected by the test phase under investigation, paying
particular attention to the water depth and wave generation capability.
These were initially constructed for ship model testing. They are long and narrow with
a movable carriage that can be driven along the tank to propel the model through the
water. Large towing tanks can be a few hundred metres long and although this means
they are several metres wide only two dimensional waves can be generated.
• short tests can be conducted before the wave field reflects back from the
downstream beach.
• side walls are proportionally near so radiated wave reflections result in simulated
array effects;
5.2.3 Flumes
These are similar to towing tanks in that the longitudinal dimension is much greater than
the transverse width. However, they do not have driven carriages so are usually much
shorter than towing tanks. Flumes vary in size from a few metres to several dozen metres
depending on their primary purpose. In general most would conform to being 1–2 short
wavelengths wide (i.e. T = 1.2 s, λ = 2.25 m → width ≈ 2.4 m)
45
Marine Renewable Energy Guides
As with towing tanks flumes have not been constructed for wave energy device testing
so should be selected to fit a specific purpose. A selection of flumes include current flows
which might be important to check mooring forces and hull behaviour under a limited
number of conditions.
Small flumes can be ideal tools for producing two dimensional results for mathematical
models or studying new or unproven concepts.
Irregular seaways will be long crested in nature and side wall effects should be considered.
5.2.4 Basins
These are proportionally wide to their lengths. In general those available internationally
are useful for model scales of λ = 1 : 10–100, and so cover the phases 1 and 2.
Since the side walls are now in the far field, radiated waves present less of a problem.
However, the additional width means more transverse modes of oscillation can occur so
care is required that cross waves do not develop to contaminate results. Most tanks have
specific frequencies to avoid, especially since standing waves are very slow to dissipate as
there is no absorption mechanism on the side walls.
In some facilities paddle banks are fitted along one side and sophisticated programming
moves paddles independently such that spreading seas occur. An alternative to this
approach is to have paddles down two adjacent walls so waves can be propagated from
both directions. The choice of facility can become a question of final deployment site
conditions.
Both solo and arrays of devices can be studied in 3D basins. Mooring arrangements can
usually be accommodated.
The common requirement in each of the tank types is the wave generation system and
the downstream energy absorption beach.
Wave generation systems are usually either mechanically or hydraulically driven and
computer controlled. They should be capable of producing the variety of waves described
in other sections of this report, that is monochromatic and short and long crested
irregular seas. There can be other less important features not covered in this document
that produce rare events or complex frequency mixes.
The range of wave periods is often the same in different sized tanks but the height that
can be generated increases proportionally in the larger, deeper tanks. The functional
components of all generation systems are the paddles.
46
Tank Testing of Wave Energy Conversion Systems
There are several important issues concerning the selection of an outdoor test site.
These are:
• accessibility, a local convenient and available harbour for light service tasks;
• distance to landfall;
Of these the most difficult to achieve is the last, the local wave climate. These should be as
close as possible to the scale conditions found at the site the test station is representing.
There can be more than one representation at different scales.
Wave paddles come in many forms but the principle of operation is similar for each
type. Each paddle has a Biésel transfer function which relates the wave amplitude to the
paddle displacement. The most common paddle types are piston and flap.
Piston paddles stay vertical and oscillate in a horizontal translatory mode. Flap types
are hinged at the bottom so oscillate in a rotational motion. The latter are regarded as
more appropriate to deep water wave production since they cause the water particles to
develop orbital velocities quicker than piston paddles.
In well equipped facilities flap paddles have a mid-depth additional hinge and are
designated second order types. These can be controlled such that they produce waves
with a pure sinusoidal profile whilst single hinge types exhibit higher harmonics and
bound waves.
47
Marine Renewable Energy Guides
It is important to reduce the effect of reflection of the generated waves from the end
wall of flumes and basins, so sloped beaches and/or absorption systems are put in place.
For shallow basins, these beaches generally have a slope of 1 : 20 to half depth and are
combined with porous absorbing material placed on the berm.
For deep water tanks this system is impractical and several alternatives are available. A
ubiquitous system is the surface piercing parabolic beach. These are used to dissipate
the energy in the active wave as it passes over the beach, breaks and permeates
the absorption mesh. In this way, the amount of energy returned by the end wall to
contaminate the measuring site is minimized.
The amount of reflection from the beach is a function of the wave period and, to a lesser
extent, the amplitude. An efficient absorber should have reflection coefficients of less
than 20% by amplitude at the worst wave period.
It is inevitable that some energy will return to the paddles due to the primary reflections
from either the test device or the end wall boundary. This will reduce the accuracy of the
trials as the reflected components will interact with the incident wave field and modify
it. If secondary reflections from the paddles occur the situation becomes unstable as the
incident wave will increase over time. This can be overcome by using a wave generator
with an active wave absorption system that can generate the desired incident wave field
while simultaneously absorbing any waves that are reflected back towards the paddle.
The active wave absorption system maintains the programmed incident wave field at the
test structure while preventing spurious wave energy build-up in the section of the tank
between the wave generator and the test structure.
Two types of absorption systems are available. One uses wave probes measuring the
surface elevation close to the paddle, while the other method uses dynamic force on the
paddle. Either set of readings can be used for the feedback signal into the main servo
control loop for the paddle motion. Any incoming wave train to the paddle is absorbed
while generation of the programmed outgoing wave train continues.
Wave probes are used to give a measure of the elevation of the water surface at any point
in the wave base or flume as a function of time. The following are some of the different
types of wave elevation recorders available:
• resistance;
• capacitance;
• acoustic.
48
Tank Testing of Wave Energy Conversion Systems
Resistive and capacitive probes are twin wire electrical probes that pierce the surface.
Elevation is measured by the change in conductance between two parallel wires. The
conductance and hence immersion can be measured by applying a potential difference
between the wires and measuring the current that flows. This change in voltage is linearly
proportional to the water surface displacement. As the conductance in water can change,
the probe should be calibrated regularly.
Acoustic elevation recorders are less common, however, they are highly accurate
and non-intrusive. Their weakness is in their sensitivity to the media properties, wave
steepness and cost.
5.2.10.1 General
The sources of reflected waves in test tanks have been highlighted and the hardware
that can reduce this unwanted contamination has been described. The options for how
the water surface elevation at the model station is monitored have also been covered.
The final recommendation is that the reflection coefficient of the tank should be
measured and quantified.
There are several methods available to conduct this calibration depending on whether
the waves are regular or irregular. All methods work on a similar principle, which requires
a number of wave probes simultaneously measuring the water surface elevation. These
can be in line or in a 2D array. The standard approaches are:
• Goda and Suzuki’s twin probe moveable carriage (2D, frequency domain);
To present an example of the methodology the Funke and Mansard approach may be
considered for long crested waves.
To achieve the required measurement three longitudinal in-line wave probes should
be positioned at any location in the tank. The distances between the probes, however,
should be exact spacings referenced to the wavelength under review. Each records a set
time history of the water surface which is transposed into associated trigonometrical
equations of the Fourier coefficients.
49
Marine Renewable Energy Guides
From these calculations the incident and reflected component waves can be obtained
and the reflection coefficient, which is the ratio of the amplitudes of the two wave sets:
5.3.1 General
The data acquisition equipment required for any device or series of tests will vary
depending on several factors, such as:
Analogue Digital
Logger
Sensor Multiplex
A/D
Sensor
High speed
Sensor clock
Sensor
50
Tank Testing of Wave Energy Conversion Systems
A specialized sensor is selected for each particular physical parameter being monitored.
Sensor requirements are:
• type;
• range;
• calibration;
• accuracy/resolution/repeatability;
• weight/dimensions;
• water resistance.
The three main requirements during conversion of the analogue signal to digital form are:
• A high speed clock that will scan channels quickly enough that reading can be
regarded as simultaneous. This is not always possible due to the number of input
signals and/or variation of sensors. If a time penalty is found the shift should be
compensated for in the analysis stage. This can be time-consuming and not always
accurate. Recently introduced acquisition systems offer parallel reading so all
channels are instantaneously recorded.
• An analogue to digital converter (A–D) which produces the stepped signal shown
in Figure 24. The specification for this is that the number of bits available should be
sufficiently large for fidelity of the conversion. With modern components 16–32 bits
can be quite common, which is more that adequate for most operations.
5.3.4.1 General
As stated above the quality of a digital signal depends on the hardware producing
it and the sensors generating the original record. Other than these factors the main
requirements are the length of the record and/or number of samples taken, these
obviously being related through the acquisition rate. There are two important
considerations when establishing conversion settings:
51
Marine Renewable Energy Guides
The type of physical property being measured dictates the data acquisition rate required
to produce acceptable fidelity of the time series. If a slowly varying signal is monitored,
which is less than the wave frequency, then the rate can be set coincidental with the
water surface elevation acquisition rate. Examples of such variables are non-extreme
motions, second order mooring loads, PTO forces etc.
If the physical property has a rapidly varying signal then an increased acquisition rate
should be selected and the different time step compensated for during analysis. Examples
of these are snatch loads in the mooring, wave slamming, air turbine pressure differences
for control purposes etc.
This second group of measurements can lead to very large data files so it is usually kept
to a minimum. Early phase 1 and phase 2 tests should have indicated which parameters
might require faster than normal acquisition rates, and under which conditions.
Model testing experience has shown that between 16–32 samples per average cycle
produces an acceptable accuracy for most signals. The actual rate depends on the scale
underway. The equivalent at a benign site for example would be between 2 and 4 Hz.
It can also be beneficial to acquire data at the same frequency the forcing function is
generated, for the reasons described below.
When the fully functional PTO, generator and power electronics are operational on a
phase 3 model task specific rates should be set. This report is concentrating only on the
data gathered for performance evaluation, not control.
The length of the record taken and number of data points in a file is a consequence of
the purpose of the time series. Regular single frequency wave trials in hydraulic facilities
theoretically need only be one cycle but in practice 50–100 oscillations (depending on
period) are usually taken so instabilities or higher order effects become visible.
The acquisition rate can be selected to generate a 2n number of data points, usually
4096 (n = 12), for the 20–30 minute record. The approach is appropriate if monitoring is
conducted in the truly random wind generated sea, where excitation is not controlled
and is not expected to contain periodicity, or in a test facility that uses quasi-random
waves generated by harmonic methods that do produce time histories with periodicy.
In the latter case it is convenient to select a sampling acquisition rate equivalent to the
frequency used in wave spectra generation. This couples the spectral ordinates and
reduces numerical errors, such as leakage between the spectral ordinates.
52
Tank Testing of Wave Energy Conversion Systems
The actual number of data points is set for convenience of Fast Fourier Transform methods
but equally applies for frequency or time domain analysis, see Clause 6. The amount of
data points required is also a factor of the number of degrees of freedom that will be
applied during the analysis and might remove the requirement of a 2n set. For example if
six 512 blocks are used the total will be 3072. However, a single set is still 29.
Although the principles of data acquisition should be regarded as the same for both
laboratory tests and actual sea trials, the logging hardware differs considerably. For
indoor laboratory studies the digital signal usually feeds directly into the computer by
an appropriate connection board. This can result in online analysis, or more often, post-
processing. At outdoor stations, which obviously are at sea, data can be stored on-board
or transmitted to shore for logging in real time. One of the main concerns in all cases is
security of data since loss is irretrievable. The advantage of on-board logging is simplicity
of the equipment, the disadvantage is possible bulk loss of the data even when checked
regularly. The advantages of telemetry are real time monitoring but the disadvantage is
equipment complexity leading to higher probable short term loss.
The type of monitoring required for WEC testing increases in complexity as the scale
increases in accordance with the objective of each phase. The core of each schedule is
shown in Table 3 and a method of detailing each phase to ensure no item is overlooked
by the flow diagram in Figure 3.
• body motions;
• PTO forces;
• mooring forces.
53
Marine Renewable Energy Guides
5.3.7.1 General
A detailed record of the water surface elevation providing the wave energy extraction
excitation forces is essential during all three phases of the development path. The
equipment used differs, but the principle behind the acquisition should be the same.
For fixed devices in long crested waves this position can be athwart of the model and in
the far field to minimize the influences of radiated waves. For moored, mobile devices
only an approximation can be made within the mooring surge range.
These real time measurements of the impinging waves should be included during all the
tests in the programme. This is principally to provide time history information for each
test but also verifies that programmed conditions are supplied. For moored vessels the
wave probes can be attached to a movable carriage so the sensors can be adjusted to the
required surge offset for each different monochromatic wave period under investigation.
In irregular seas, where mooring induced surge is constantly varying, only the average
setting is possible as stated above.
Wave recorders can also be attached to models so they move around with the device.
Since this means they are quite close the influence of radiated waves, especially when
out of phase with the incident waves, they should be considered. This approach should
be regarded as an approximation method but can still provide useful data during the
signal analysis stage.
In short crested seas the situation is more akin to outdoor sea trials in that only a
near field approximation can be measured by a 2D system. The pre-calibration station
measured records can be referenced in parallel with the real time surface record to
ensure there are no major deviations in the summary statistics.
Wave measurements during sea trials are also essential if full value is to be taken of the
deployment. There are various acceptable types of instruments available, such as surface
54
Tank Testing of Wave Energy Conversion Systems
buoys. An alternative to this is the bottom mounted acoustic Doppler type. Originally
introduced to measure current profiles a variant of the ADCP can now furnish the water
surface time history similar to the more conventional buoys. The location of the wave
recorder is somewhat arbitrary but a station upstream of the device(s) is often favoured.
The dispersive nature of water waves dictates that only spectral data can be used reliably,
and not the time histories which will differ between the buoy and the device.
Until recently for early small scale laboratory trials device hulls had to be connected
to rotary or linear potentiometers by physical means, such as wires or drive cords. The
body motion was then measured by the position sensor. To facilitate small amplitude
assumptions the connection line had to be long and in large motions, especially when
combined with second order drift displacement, it was common for the device cord to
fall off the pulley. This inevitably occurred towards the end of an experiment and lead to
the test having to be repeated.
For medium scale trials accelerometers could be used but were not very reliable or
convenient to deploy and required double integration to obtain displacement.
Although improved versions of these sensor arrangements are still used today they have
in general been replaced at both scales by non-contact motion monitoring systems. These
are based on active digital cameras which are in fixed locations around the test site and
passive (or active) reflective markers which are attached to the model.
If a passive marker system is used the active camera sends out an electromagnetic
radiation pulse in the infrared band which is reflected back by the markers and their
position in 2D space is recorded. If more than one camera is operational the system
software converts each camera reading to a 3D point in space. A continuous time history
of each reflective marker’s x, y and z coordinates are generated in the output file matrix. If
more than one marker is used (two for 2D and three for 3D) further transforms should be
used converting the Cartesian coordinates into the required degrees of freedom motion
for the hull heave, pitch, surge, roll, sway and yaw.
Because the system is non-contact the WEC is allowed to move freely on a mooring system
providing the second order excursion remains within a pre-calibrated measurement zone.
Although a considerable improvement over the previous methods the system is not
without flaws. Extreme accelerations are difficult to track and merging markers (when
two reflective markers cross in a camera view) cannot be distinguished and lead to
erroneous data errors (see Clause 6).
For large scale, phase 3 sea trials, on-board sensors should still be relied upon. There are
several systems commercially available based on accelerometers which usually require an
axis reference stabilizer, such as a gyroscope. Pitch and roll stabilizers should be of a high
quality or signal integrated noise problems are inevitable.
55
Marine Renewable Energy Guides
Once the body motions are monitored the forces developed between the two inertial
masses of the WEC should be measured. From these two parameters the instantaneous
power can be calculated since:
For pneumatic devices this requires the monitoring of the pressure difference across the
turbine simulating orifice. If mechanical PTO simulators are used the force transmitted
through them should be recorded. It is not actually necessary in either case to know the
level of applied damping but recommended that the settings are calibrated.
Force transducers can be deployed as in-line measurement sensors. The location along
the mooring line to place the sensor requires consideration. The main criteria should be
to follow the requests of the mooring design engineers but also to ensure all aspects of
the anchorage are revealed. Both bow and stern mooring should be monitored but for
symmetric configurations only port or starboard need to be connected for laboratory trials.
A recommended location for a transducer is at the chain stopper on the device. For
phase 1 models this can prove difficult since the site will still be in a splash zone and
waterproof strain gauges are heavy compared to the weight of the model. It might,
however, be possible to extend the mooring past a fairlead to a dry location. For medium
and large scale phase 2 and 3 models this is rarely a problem. At the small scale care
should be taken that the sensor’s own weight does not become a factor in the mooring
recordings. At large scale this is not relevant. An alternative approach in small scale
trials is to attach buoyant material to a waterproof sensor and attach it in line to either
the device or to the weight bearing float, if there is one in the system. The inertia of the
system will be influenced.
Although mooring measurements can be conducted at small scale this should only be
a guide for mooring designers and should not replace measurements from the medium
scale tests. The latter results can more confidently be applied to the mathematical safety
codes used by the design companies. Since snatch loading can be present in mooring
lines a fast acquisition rate might be required.
NOTE The calibration of the force transducers should be performed before and after the test
programme and regular zero load checks made during testing.
56
Tank Testing of Wave Energy Conversion Systems
5.4 Calibration
5.4.1 General
5.4.2.1 General
This should be performed prior to testing in enclosed tanks for three reasons:
Mechanical wave generation systems have frequency related transfer functions that
should be established for all installations. Providing no changes are made within the tank,
such as to the downstream energy absorption beach, this should not change over time.
A set of calibration curves for a range of monochromatic waves typical of this procedure
can be drawn up by programming the control computer in steps of increasing gain and
measuring the response in terms of wave height generated. If the generation system does
not have adjustable gain each individual wave height required should be run separately.
The wave period should similarly be checked during the same runs.
Water levels can influence the transfer function so tank levels should be checked
regularly. This procedure should be followed for single frequency, regular waves and
irregular seaways. Before beginning a set of important device trials, it is prudent to
confirm that the transfer function is still valid. These checks should be made prior to the
deployment of the model and preferably at the proposed test location within the tank.
A similar approach is taken with real seas when the full time history is run and the summary
statistics obtained for each seaway. A graph can then be produced of the theoretical and
practical significant wave height such that intermediate seaways can be generated. The
spectral shape should also be verified as described previously.
57
Marine Renewable Energy Guides
Once the standard wave climates are established for a facility they need only be statistically
verified at the model deployment station. If additional specific spectra are required these
should be calibrated following the above procedure. Again this should be completed prior
to the launch of the model.
The reasons for the occurrence of and methods of reducing reflected waves was
described in 5.2. However, they are ever present and should be quantified since they do
add a complication to testing in enclosed wave bodies. Even when the model is absent
some reflection from downstream beaches contaminates the wave field and should be
taken account of.
For buoyant point absorbers which surge over extreme distances an allowance during
analysis for height changes might be required. Variations are easily identified in regular
waves by inspecting the response time trace for differences in magnitude of the oscillations.
Transducers and other types of sensors may be obtained with or without manufacturers’
pre-calibration certificates. In either case it is good practice to pre-calibrate both types,
especially if this can be done in situ on the model and connected to the data acquisition
equipment and logger. The minimum should be a bench calibration of the complete
acquisition system for each sensor.
Wave recorders should also undergo calibration. In the laboratory the common use of
resistance probes can be checked reasonably easily by raising and lowering them in
still water, thus producing a calibration graph similar to Figure 25. When the step size
is known the voltage change can be related to the specified excursions such that the
calibration units become distance per unit digital change, i.e. mm/digit.
At sea things are more complicated. Most proprietary wave gauges are factory calibrated
but should occasionally be checked. This could involve returning the unit to the manu
facturer for service or the technique of deploying a second unit for a short period of time
to verify the primary recorder. In general the manufacturer’s service recommendations
should be followed verbatim, which can be expensive so should be included in the
budget planning.
58
Tank Testing of Wave Energy Conversion Systems
Probe calibration
6
2
Voltage (V)
-2
-4
-6
0 20 40 60 80 100 120 140 160 180 200
Data point
During phase 1, conducted in a laboratory, global scale results for comparison between
different configurations are paramount so that the accuracy can be relaxed somewhat. This
is because results are relative rather than absolute. However, repeatability is important.
By phase 2 the absolute value of the parameter is required so measuring tolerances should
be tightened but the resolution is not so critical.
Of more concern, particularly at small scale phase 1 trials, are the ranges of the parameters
measured. Sensors should be carefully selected to be appropriate to these limited scales,
which can restrict the choice. The accuracy of most gauges and transducers is a function
of their full scale deflection.
If an active control system is under investigation during the phase 2 part of the programme
then the transducer specifications should become even tighter. This requirement certainly
advances into phase 3 when a PTO control of some description will definitely be required.
The sensor selection also becomes a multidisciplinary task so the final decision of the actual
units should be left to the appropriate engineers. Providing the underpinned principles of
calibration are followed a system should function satisfactorily.
59
Marine Renewable Energy Guides
One aspect that should be agreed for phase 3 is that of sensor redundancy. For critical
components and parameters, dual or back-up sensors should be incorporated when
possible. One rationale for this is if the sensor will not be accessible at sea due to either
its location or integration in the control circuitry.
Table 6 offers a guide to some common parameters and the overall accuracy from the
sensor to the logger.
waves ≤ 5% 5% 10%
force/pressure ≤ 5% 2% 10%
generator ≤ 2%
6 Calculation procedures
6.1 General
Analysis of test results requires as much skill, experience and knowledge as conducting
the physical test programme. Although standard mathematical techniques can be
employed to validate data and produce the first stage of the analysis for each test run
and device configuration investigated, progressing this information into the summary
stage will be a bespoke activity requiring imagination, discipline and ingenuity.
6.2.1 General
The two basic approaches are time domain and frequency domain analysis. Both
methods have germane usage.
60
Tank Testing of Wave Energy Conversion Systems
6.2.2.1 General
Although spectral methods can be used for monochromatic results, and indeed should
be employed on certain occasions, it is usual to produce the various physical parameters
from time series. Important information can be obtained directly from raw data records
or derived from the combination of two or more parameter signals.
Figure 26 shows a typical time trace of four monitored signals from which considerable
relevant information regarding the characteristics of a device can be read directly.
Examples of the observations are:
• phase relationship;
• quality of signal;
• resonance proximity;
• signal statistics (root mean squared (RMS), maximum, minimum, mean, standard
deviation etc.).
Prototype scale
monochromatic Extension Piston(z)
F05 (H = 2.5 m, T = 9.4 s) Body Wave
4
2
Displacement (m)
-1
-2
-3
-4
500 516 532 548 564
Data point
61
Marine Renewable Energy Guides
Further insight into device behaviour, characteristic, performance etc. can be revealed
when signals are combined, such as shown in Figure 27. In this graph, the differentiated
motion signal (i.e. velocity), and the direct force signal product produces the instantaneous
power from which the required statistical values can be derived (e.g. mean power).
250
Abs. power output (kW)
200
150
100
50
0
500 516 532 548 564
Data points
A useful continuation of the time history analysis is to compare the measured device
performance with the incident wave creating that response. Two primary areas of
interest are:
• phase relationship.
The method of obtaining these values in regular waves is illustrated in Figure 28.
One wave excitation frequency produces a single measure and plot point of the device
response. Multiple frequency runs can then be combined to produce a complete picture
of the device behaviour.
62
Tank Testing of Wave Energy Conversion Systems
parameter under investigation can be generated. These summary plots provide the
basic design requirements so the WEC can be optimized. An example of this is shown in
Figure 29 depicting three different body motions. Design engineers can evaluate these
graphs to ensure the required characteristics are being exhibited.
1.5
RAO = —
Hi
1
H i
Hj
0.5
H j
Elevation
0
p i p j
-0.5
-1
-1.5
Phase, θ = Pj – Pi
-2
0 2 4 6 8 10 12 14 16 18 20
Time
1.4 0.007
Heave and surge RAO (mm/mm)
1.2 0.006
Pitch RAO (rad/mm)
1 0.005
0.8 0.004
0.6 0.003
0.4 0.002
0.2 0.001
0 0
0.5 0.7 0.9 1.1 1.3 1.5 1.7 1.9 2.1
Monochromatic wave period (s)
(Model scale)
[email protected] Confidential
63
Marine Renewable Energy Guides
This process should then progress a stage further by comparing different device
configurations, or settings, to highlight which combination produces the results
specified by the design team. An example of this is shown in Figure 30 (and Figure 8).
The unit under investigation was a heave mode extraction WEC for which pitch was to
be minimized. The governing mass properties were adjusted according to the equations
specified in Clause 5 to produce the three uniquely different responses shown in the
graph. From such comparisons the design engineer can select the combinations best
suited to a particular device.
OE BUOY PROJECT MARINE INSTITUTE IND/03/13
0.006
RAO (rads/mm)
0.004
0.002
0
0.5 0.7 0.9 1.1 1.3 1.5 1.7 1.9 2.1
Monochromatic wave period (s)
(Model scale)
[email protected] 19/4/09 SB Pitch RAO
Response to multi-frequency wave fields can also be analysed by time domain methods.
Figure 31 shows the real sea time series equivalent to the monochromatic traces in
Figure 26. Less detailed understanding of the behaviour can be extracted from these
results but important criteria such as extremes, averages, variance etc. (which are required
by designers, especially regarding seakeeping characteristics), can be extracted.
Also critical is the instantaneous power absorption under real wave conditions. Figure 32
shows a plot of the product of the simultaneous PTO velocity and force from Figure 31.
Such information is essential for the design of a power take-off system and the power
electronics that will convert the supply to an acceptable quality for connection to a
national grid, particularly that of dealing with the spikes.
64
Tank Testing of Wave Energy Conversion Systems
To provide a full set of design criteria these plots can be further analysed to provide
information regarding:
• the time period for energy above a set threshold, such as the average;
Panchromatic
wave heave pressure
wave probes
80 100
60 75
40 50
Wave elevation (mm)
Pressure (Pa)
20 25
Heave (mm)
0 0
-20 -25
-40 -50
-60 -75
-80 -100
140 141 142 143 144 145 146 147 148 149 150
Time (s)
Panchromatic Power,p&Q
instantaneous pneumatic power (kW) Average Power, p&Q
prototype scale Average Power, Q
Average Power, p
4500
4000
3500
3000
Inst. power (kW)
2500
2000
1500
1000
500
0
1400 1410 1420 1430 1440 1450 1460 1470 1480 1490 1500
Time (s)
65
Marine Renewable Energy Guides
Time series analysis is also required to aid mooring design. Figure 33 shows details
from two different mooring line configurations. Both are exposed to the same, or
similar, seaways. The first is a well designed system with pretension and compensation
buoys. The peak to average tension ratio is approximately 3 : 1 and no snatch loads are
present. In the second plot, representative of a slack moored catenary type mooring the
maximum to average tension ratio is approaching 10 : 1 and in this record snatch forces
become present. This information can be combined with the regular wave second order
work to verify and calibrate mooring design computer packages.
7
Mooring force (N)
0
100 120 140 160 180 200 220 240 260 280 300
Time (s)
Mooring force
panchromatic trial
40
35
30
Mooring force (N)
25
20
15
10
0
0 500 1000 1500 2000 2500 3000 3500 4000
Data points
66
Tank Testing of Wave Energy Conversion Systems
6.2.3.1 General
Analysis of irregular signals can also be undertaken by a spectral approach, since the
excitation waves can be represented by a frequency spectrum.
The principle behind this technique is that an irregular signal is the super-position of a
series of regular waves which can be decomposed into frequency components. Figure 34
illustrates this technique graphically.
S(f) Magnitude
Time series
Monochromatic Waves
(Randam Phase)
f Short Period
Oscillations
Time Domain Axis
(Irregular Surface Elevation)
Theoretically each wave frequency can vary randomly in amplitude and phase but in
ocean sea states it is more usual that amplitudes vary according to governing physical
laws of energy transfer whilst phase is regarded as random. This means that when
superimposed the regular waves produce an irregular oscillation as shown in Figure 34
on the time series axis. Along the frequency axis the spectral energy density that these
monochromatic harmonics are derived from is depicted.
Any of the measured physical parameter signals, as shown in Figure 31, can be processed
into a series of frequency ordinates by the reversal of the procedure described above. That is,
an irregular time series is measured and the spectral profile obtained by harmonic analysis.
The recommended algorithm used for computer based analysis is the Fast Fourier Transform
(FFT), which is an efficient method for producing the discrete Fourier series. The spectral
ordinates can be viewed as amplitude values (m) or energy density values (m2/Hz) or the
appropriate units for the parameter being analysed (e.g. pitch (rad2/Hz), energy (W2/Hz)).
67
Marine Renewable Energy Guides
There are two particular characteristics of spectral analysis that should be considered
when applying it to a data set.
• The analysis technique relies on decoupling of the signal into harmonic components,
which might not occur in nature. This methodology introduces the factor of periodicity
into the analysis when evidence suggests it might not have existed in the original signal.
However, although the individual amplitudes of the harmonic components can differ the
overall variance produced by the spectral method should not vary significantly from the
variance calculated from the time series.
This last statement has several caveats that should be accommodated when conducting this
work since ancillary application can influence the result accuracy. Examples of these are:
• filtering;
• smoothing;
• windowing;
Harmonic analysis is less universally used as an analysis tool than time series analysis but
does have several important contributions to add to the evaluation and understanding of
the process governing device performance.
Since monochromatic waves can’t be expected in the ocean it offers a useful tool to
investigating the excitation and response of a device once the laboratory is left, i.e.
phase 3 and beyond. Its use to produce the RAOs of a device was described in Clause 5.
Direct spectrum comparisons can also reveal important information. In Figure 35 the
relationship between the wave spectra and the power conversion of a typical buoyant
WEC in three different seaways is shown.
The conditions have been selected to illustrate seas with a peak period below the device
resonance (Tp < TD), coincidental with the device natural period (Tp ≈ TD) and the peak
energy period above the device eigen frequency (Tp > TD). As can clearly be seen the
peak in the power production does not migrate with the wave energy peak by as much
as might be intuitively expected. Indeed the period of maximum power absorption
varies little from seaway to seaway. It is also shown that the magnitude of the peak in the
resonant case is an order of magnitude greater than both cases either side of resonance.
68
Tank Testing of Wave Energy Conversion Systems
400 4000
Power spec (W^2/Hz)
300 3000
Pre
200 2000
100 1000
0 0
0 1 2 3 4 5 6 7
Period(s)
3500 7000
3000 6000
Power spec (W^2/Hz)
Resonant
2500 5000
2000 4000
1500 3000
1000 2000
500 1000
0 0
0 1 2 3 4 5 6 7
OE BUOY PROJECT Period (s) S.E.I. OE/03/006
500 0.05
Power spec (W2/Hz)
400 0.04
Post
300 0.03
200 0.02
100 0.01
0 0
0 1 2 3 4 5 6 7
Period (s)
[email protected] Figure 35c edited.xls 19/4/09
69
Marine Renewable Energy Guides
This detailed information can support the summary device specification presented from
the time series analysis and from which the power matrix performance scatter diagram
has evolved, see Figure 36.
2.0 57 88 115 136 148 153 152 147 138 127 116 104 93 83 74 66 59
2.5 89 138 180 212 231 238 238 230 216 199 181 163 146 130 116 103 92
3.0 129 198 260 305 332 340 332 315 292 266 240 219 210 188 167 149 132
3.5 – 270 354 415 438 440 424 404 377 362 326 292 260 230 215 202 180
4.0 – – 462 502 540 546 530 499 475 429 384 366 339 301 267 237 213
4.5 – – 544 635 642 648 628 590 562 528 473 432 382 356 338 300 266
5.0 – – – 739 726 731 707 687 670 607 557 521 472 417 369 348 328
5.5 – – – 750 750 750 750 750 737 667 658 586 530 496 446 395 355
6.0 – – – – 750 750 750 750 750 750 711 633 619 558 512 470 415
6.5 – – – – 750 750 750 750 750 750 750 743 658 621 579 512 481
7.0 – – – – – 750 750 750 750 750 750 750 750 676 613 584 525
7.5 – – – – – – 750 750 750 750 750 750 750 750 686 622 593
8.0 – – – – – – – 750 750 750 750 750 750 750 750 690 625
This diagram is at present being proposed as an industry standard for describing the
power conversion characteristics of wave energy converters such that they can be
equitably compared.
To be of value the method of derivation of the power value in each element of the power
map should be displayed in combination with this information.
At present most power maps are theoretically generated but if, as will eventually be the
case, the matrix is produced from measured, empirical data a further layer of evaluation
will be required for the following reason.
70
Tank Testing of Wave Energy Conversion Systems
the power absorption for that time slot. The variation in levels is obvious. Below each
highlighted black histogram is the actual frequency component mix that produced those
same summary statistics. For the left hand example, the real sea resembles a classical
Bretschneider and the WEC performance is similar to the scaled laboratory results. For the
middle example a twin peaked spectra was present with zero input frequency ordinates at
the peak period. Very little power was produced in these wave conditions even though the
summary statistics produced the same Hm0 and T02 combination. The example to the right
has a JONSWAP type narrow spectrum with a higher peak close to the peak period for the
Bretschneider. These wave conditions produced a slightly higher power production.
5_12a
75
50
25
0
1.4 1.4 1.4
Measured Measured Measured
1.2 1.2 1.2
Bretschneider Bretschneider Bretschneider
1 1 1
S(f) (m^2/Hz)
S(f) (m^2/Hz)
S(f) (m^2/Hz)
0 0 0
Page 1
0 0.1 0.2 0.3 0.4 0.5 0 0.1 0.2 0.3 0.4 0.5 0 0.1 0.2 0.3 0.4 0.5
Frequency (Hz) Frequency (Hz) Frequency (Hz)
Without the application of the spectral approach to the data analysis this type of
information would not be revealed to the device designers.
6.2.4.1 General
71
Marine Renewable Energy Guides
If typical characteristics of the physical parameter are known error traps can be imple-
mented on an individual signal. Often this is improved if an analysed, rather than raw,
trace is used since the error can be enhanced. A further modification can be to apply the
check to the product of more than one signal, though this method does not isolate the
problematic channel.
Figure 38a (displacement) shows a raw signal with small errors that would be difficult to
detect. Figure 38b (velocity) depicts the differential of that signal in which the problem
becomes more obvious and Figure 38c (power) is the product of the error signal and
a second, clean, trace (not shown) in which an error has now become evident and can
easily be identified digitally.
OE BUOY PROJECT
500
(mm)
400
300
200
0.01
(b)
Volume flow
0.005
0
-0.005
-0.01
30
(c)
Inst power (W)
25
20
15
10
5
0
2000 2050 2100 2150 2200 2250 2300 2350 2400 2450 2500
Time series
[email protected]
If such monitoring errors are not detected significant and misleading differences can be
incorporated in the summary records, and in particular the power statistics.
Several mathematical techniques exist to improve signal quality but how and which to
apply is a judgement decision. Problems can present themselves in various forms and
each requires a different approach for the best solution. Basically the solutions involved
are frequency filtering or time series smoothing.
There can also be gaps in the time series, which should be compensated for. This can
necessitate a data set being completed by zeroes or perhaps spline filling but care should
be taken or the signal can become distorted.
72
Tank Testing of Wave Energy Conversion Systems
a) Smoothing
High frequency noise or jitter on a signal can be smoothed in the time domain, taking
care not to introduce phase shift. Figure 39 shows an example of a raw and smoothed
signal and the improvement in the processed trace. A consequence of smoothing can be
a reduction in the amplitude, which can also be seen in Figure 39.
0.04
0.02
Volume flow (m^3/s)
-0.02
-0.04
-0.06
160 165 170 175 180 185 190
Time (s)
b) Filtering
Although similar to smoothing this method separates frequencies by allowing some
to pass and attenuating those that are unwanted. Initially applied in the frequency
domain, filtering can be particularly useful to remove low frequency oscillations that
can be masked in the time history. Electrical analogue filters can be incorporated in
the hardware or digital numerical filters in the software, but each should produce the
same effect. Unfortunately no electronic filters have ideal characteristics as they tend to
reduce a problem rather than remove it completely. There can also be difficulties when
the frequency of the unwanted part and the required data information occupy the same
section of the spectrum.
This is often the case with low frequency noise introduced during integration of a signal
such that the left hand side of the spectrum is masked by the noise. If the signal can be
cleaned up prior to analysis, errors can be minimized.
Band pass filters which are applied to the high and low frequency conditioning are
particularly useful.
Figure 40 shows a raw data signal and the resultants when the two fundamental types
of filter are applied. The left hand trace has allowed the high frequency components to
pass but removed the low frequencies. The right hand time trace has removed the high
frequency elements. The difference between the resultants is clear and the dangers of
such powerful tools being applied incorrectly is obvious.
73
Marine Renewable Energy Guides
c) Windowing
The efficiency of digital filters can be improved by the application of windows to the time
series filter kernel. Not only can the Sinc function become manageable for computing
but the pass band and stop band characteristics are enhanced. The characteristics are
shown in Table 7.
Even with improved quality filters there will be the inevitable alteration to the original
signal in phase and amplitude that might require correction prior to continued application.
d) Manual
It is not uncommon in laboratory studies that jumps occur in the time trace, as seen
in Figure 41. This problem is usually detected in the processed signal as described
above. These errors are more difficult to correct and often lead to either repeated trials
74
Tank Testing of Wave Energy Conversion Systems
or manual adjustment to the data set. Ideally this problem should be avoided as the
solution is not trivial and can be time-consuming.
Motions in y axis
Bow y Stbd y Stern y Plenum y
0
-50
-100
-150
y axis (mm)
-200
-250
-300
-350
-400
-450
-500
2100 2110 2120 2130 2140 2150 2160 2170 2180 2190 2200
Data series
6.3.1 General
One of the primary reasons that scale model testing is conducted is that there are rules
that enable it to be done. Physical laws of similarity and similitude can be applied such
that the fundamental behaviour of a process can be observed, even when the theory
governing the situation is not well defined or even poorly understood. The physical
model behaves as an analogue computer that introduces all the parameters and
interactions that would occur at full size. Most wave energy converters seem like simple
devices but there are usually quite complex hydrodynamic interactions.
However, as with all engineering solutions, there are conditions that should be followed.
In this instance the principle consideration is that not all physical processes scale to the
same laws. The consequence of this fact is that the dominant relationships are compared
and the appropriate similitude law adopted.
For wave energy physical modelling there are two main criteria to consider, Froude and
Reynolds. Froude equates inertial forces whilst Reynolds relates viscosity. Because of their
derivation both of these two criteria cannot be matched in the same model. Since inertia
dominates the majority of the forces experienced in wave energy device operation Froude
scaling is selected as the primary law to match between the model and the prototype.
It is, however, important to know that there will be some forces present that are not
to scale and their influence on the results should be qualified if not quantified. This is
75
Marine Renewable Energy Guides
particularly the case in small scale wave energy models, circa λ < 1 : 50, that have large
wave frequency, first order, excursions, especially heave. Such physical combinations
produce the worst deviations from full scale results and are the primary rationale for the
progressive increase in scale through the phases. This approach not only applies checks
to the scale effect but, since different facilities support different scales, the laboratory
effects also.
v = velocity
inertia force v
Froude: Fr = = g = acc. due to gravity
gravity force gL L = length scale
ρ = density
v = velocity
inertia forces ρvL vL
Reynolds: Re = = = L = length
viscosity forces µ γ
μ = coeff. dynamic viscosity
γ = kinematic viscosity
76
Tank Testing of Wave Energy Conversion Systems
The factors by which a physical parameter should be multiplied to convert from model to
prototype scale can be read from Table 8. There are two important considerations when
applying this:
The figures above indicate why certain groups are concerned with prototype projections
from small scale results. The governing similitude laws dictate that these calculations are
valid but small differences can become significant in the prediction of absolute value.
The percentage error will remain constant across the scale ranges. A particular cause
of discrepancy can be non-linear, or higher order, effects that are not in kinematic or
dynamic similarity.
The multiplicands also show why even the large scale models used at benign site sea
trials do not produce significant amounts of electricity. Even a quarter scale 1 MW device
would produce only 10 kW at the design rating. This means connection to a grid has little
value but has the advantage that manageable resistive load banks can be used to burn
off the generated power.
• it validates the application of the similitude law to wave energy device development;
To date there is little practical evidence to support the accuracy of predictions since
only a small number of devices have achieved full/large scale sea trials and of those not
all followed a progressive modelling path or designs went through significant changes
between phases.
77
Marine Renewable Energy Guides
Some data does exist for validation purposes. It should be remembered that not all
physical properties will conform to the same scrutiny even when extracted from the
same set of tests. A sensitivity matrix of all parameters should be referenced when
reviewing the data.
Figure 42 shows three phases of results at small, medium and large scale. These are
normalized power production of a WEC in several different seaways, represented by the
horizontal base of the pivot plot. Each sea state is coded with a different shading by the
peak period and the benign (λ = 1 : 4) results were selected when the scatter diagram
element resembled a Bretschneider spectra. This classical shape had been used during
the small and medium scale tests.
Average power output
Ï = 1 : 50 Tp Ï = 1 : 15 Tp Ï=1:4 Tp
These results are from a pneumatic type device so the power represents an exacting
parameter which contains several factors that have unknown scale effects, e.g. air
compressibility, viscosity. Reference to the three plots reveals close similarity at the
different scales, which does support small scale trials.
The conclusion of these comparisons, once more data has been made available, could be
that the intermediate scales can be skipped. However, endorsements from developers
who have been, or are engaged in, large scale sea trials, suggest phase 3 is valid for the
experience gained in bringing together a multidisciplinary team and several differing
technologies in a fully functional device. This practical skill is achieved with the added
benefit that the rehearsal is not prohibitively expensive.
7 Reporting
7.1 General
The basic requirement and rationale for model testing wave energy devices has been
outlined in this document. The importance of a phased technology readiness level
approach has been described and examples of how test schedules for each phase can be
constructed are included. Equally as important as completing the device development
programme is the reporting of the activities and results within each phase.
78
Tank Testing of Wave Energy Conversion Systems
It is to these data sets that stage gate criteria will be applied during the most important
evaluation stage.
The question of commercial confidentiality will obviously become a concern in the content
of these reports but that should influence the distribution rather than the production.
The actual test schedule flow diagram for each specific trial set will dictate the main part
of the reporting procedures and provide the data for the downstream activities such
as analysis and presentation. However, the other elements required to configure a test
programme should also be described following Clauses 3, 4, 5, and 6.
Each report will be formatted for the device under test and the phase of the trials.
However, there should be some underlying principles and minimum inclusions to all
the technical reports. A liberal use of photography, diagrams and graphs in a report is
encouraged together with the following information.
• Objectives of trials.
• Type of model.
• Type of tests.
7.3 Facilities
• Type (flume, basin, benign site).
• Specification of tank.
• Pre-calibration of waves.
7.4 Model
• Technical specifications and material selection.
• Special features.
79
Marine Renewable Energy Guides
• Instrumentation:
• Directions selected.
• Special investigations.
• Failure modes.
80
Tank Testing of Wave Energy Conversion Systems
81
Marine Renewable Energy Guides
Bibliography
Det Norske Veritas, Guidelines on Design and Operation of Wave Energy Converters,
Carbon Trust, 2005.
Development of Recommended Practices for Testing and Evaluating Ocean Energy Systems,
IEA-OES, 2003.
Frazer-Nash Consultancy, Marine Energy: More Than Just A Drop In The Ocean, Institute of
Mechanical Engineers, 2008.
Performance Assessment for Wave Energy Conversion Systems in Open Sea Test Facilities,
EMEC, 2005.
UKERC Marine (Wave and Tidal Current) Renewable Energy Technology Roadmap,
UKERC, 2007.
82