Wave Dragon, prototype wave power production

Abstract

Wave Dragon is a floating wave energy converter working by extracting energy principally by means of overtopping of waves into a reservoir. A 1:4.5 scale prototype has been sea tested for 20 months. This paper presents results from testing, experiences gained and developments made during this extended period. The prototype is highly instrumented. The overtopping characteristic and the power produced are presented here. This has enabled comparison between the prototype and earlier results from both laboratory model and computer simulation. This gives the optimal operating point and the expected power of the device. The project development team has gained much soft experience from working in the harsh offshore environment. In particular the effect of marine growth in the draft tubes of the turbines has been investigated. The control of the device has been a focus for development as is operates automatically for most of the time. This has led to improvements in the power take off, trim control and stability of the device.

Full text

Wave Dragon, prototype wave power production
J. Tedda,b, J.P. Kofoeda, W. Knappc, E. Friis-Madsenb, H.C. Sørensenb
a Department of Civil Engineering, Aalborg University, Aalborg 9000, Denmark
b Wave Dragon ApS., Blegdamsvej 4, Copenhagen N 2200, Denmark
c Laboratorium für Hydraulische Maschinen, TU München, 80290 München, Germany
Wave Dragon is a floating wave energy converter working by extracting energy principally by
means of overtopping of waves into a reservoir. A 1:4.5 scale prototype has been sea tested for 20
months. This paper presents results from testing, experiences gained and developments made
during this extended period. The prototype is highly instrumented. The overtopping characteristic
and the power produced are presented here. This has enabled comparison between the prototype
and earlier results from both laboratory model and computer simulation. This gives the optimal
operating point and the expected power of the device. The project development team has gained
much soft experience from working in the harsh offshore environment. In particular the effect of
marine growth in the draft tubes of the turbines has been investigated. The control of the device has
been a focus for development as is operates automatically for most of the time. This has led to
improvements in the power take off, trim control and stability of the device.
Keywords: Wave Power, Overtopping, Production, Prototype
Introduction
The Wave Dragon is a floating offshore wave
energy converter of the overtopping type. A
full scale Wave Dragon designed for an
appropriate climate would have a installed
power of 4-11 MW. A prototype scaled at
1:4.5 of a North Sea model and rated at 20
kW has been tested in Nissum Bredning, a
large inland waterway in Denmark from May
2003 to January 2005.
The concept works by waves overtopping a
ramp, filling a floating reservoir with water at
a higher level than the mean sea level. This
head of water is used for power production
through the specially designed hydro turbines.
Figure 1: The Wave Dragon Nissum Bredning
Prototype.
The prototype has all the features of a
operational power plant including: slender
wave reflectors to focus the energy of the
waves towards the ramp, a pneumatic system
to adjust the floating level of the platform;
seven Propeller turbines mounted with
permanent magnet (PM) generators to convert
the potential energy of the water; and an
inverter system to control the variable speed
of the turbines. Furthermore, three calibrated
dummy turbines are used to process
overtopping flow rates that exceed the
capacity of the Propeller turbines. The power
generated is exported to the Danish national
grid via a three phase sub-sea power cable.
Availability
The Wave Dragon Nissum Bredning
Prototype has been tested in real sea for
approx. 2 years. During the period May 2003
to December 2004 the availability of the
system has been continuously increasing to up
to 80% at the last part of the period.
Monthly operating experience of the power
production systems from May 2003 until end

of 2004 is summarised in Figure 2. This
reflects time logged in the PLC system where
the Wave Dragon’s power production system
has been in active operation, either
• “Continuous operation” (green) mode
which covers longer periods where the
prototype has been in automatic
production mode. Not necessarily
aiming at optimum power output.
• When carrying out specific test runs
(labelled “Testing”, yellow). This
covers tests of control systems and
tests of hydraulic response, i.e. effect
on floating level and stability.
The additional time has been spent on:
• Re-construction and re-configuration
activities (labelled “Re-construction”,
grey)
• Waiting time (labelled “Out of
operation”, red). This covers as an
example holidays and simply evenings
and nights in the periods without a
working automatic fire extinguishing
systems (insurance question). In these
periods power production has been
stopped.
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
May 03 Aug. 03 Nov. 03 Feb. 04 May 04 Aug. 04 Nov. 04
Continuous production Testing
Re-construction Out of operation
Figure 2: Availability of WD-NB over the
period of real sea testing.
The Wave Dragon’s power production system
referred to covers turbines, generators,
inverters and rectifiers plus PLC system.
A close to 100% availability has been
achieved for other Wave Dragon systems, like
the automatic floating level and stability
system and the remote control and
communication systems.
From the 1st October 2004 to 9th January 2005
measurements were made almost
continuously on the prototype. Sample
periods of 30 minutes were chosen - 300 to
500 waves. This prevents too much scatter
within a result set, while preventing a loss of
definition due to fast wave build up in Nissum
Bredning which has a relatively short fetch.
Of the approximately 4800 such sample
periods during the time 3969 records were
made including the most important
measurements. Of these 1577 had high
quality enough measurements to allow full
time series analysis of all the flows, and of
these 247 had significant wave height great
enough to give some power production. This
relatively low proportion is due to the fact
that the platform was not fully ballasted – for
safety reasons – and thus it never floated at a
level low enough to permit overtopping at
very low wave heights. These 247 sample
periods are those shown in this paper.
Power capture
The water flow overtopping the ramp and the
hydraulic power which passes through the
turbines are the two main measures used to
compare power production performance at
this stage. The overtopping flow is compared
to predictions based on earlier laboratory
tests. The hydraulic power passing through
the turbines accounts for the lower head
across the turbine than the crest freeboard.
The electrical power generated by the real
turbines is recorded. A third measure of
power is the estimated electrical power, this is
the power which would have been produced
by the hydraulic power if the dummy turbine
flow had passed through a functioning turbine
and if the speed control of the PM generators
was working optimally.
The overtopping flow is defined as the flow
which passes through the turbines, ignoring

any spill from the reservoir back to the sea.
The individual turbine flow is calculated from
the turbine characteristic (Keller et al, 2001)
as a function of head and rotational speed.
Pressure transducers in the reservoir measure
the head.
Kofoed et al, 2006, presents the overtopping
relationship for the Wave Dragon based on
the results from tank testing of a scale model.
It is clear that there is a wave length
dependency on the overtopping. The form of
the non-dimensional units Q* and R* depend
on the wave steepness, based on the breaking
criterion. The slope of the ramp of the Wave
Dragon is rather steep; to avoid loss of energy
during breaking. A better model for the rate
of overtopping over a low crested, non-
breaking and floating structure is desirable.
Figure 3: Vertical distribution of energy in
water column
A different method to include the frequency
dependency of the waves into the non-
dimensional form is given by Kofoed, 2002.
Its physical basis has more relevance for this
case of a floating overtopping device, and is
shown above in Figure 3. The average
overtopping rate QN is non-dimensionalized
as normal and modified by the ratio of the
energy in the water column between the free
surface and the draft of the device to the total
energy in the water column.
3
1
S
d
NgHW
Q
Q
r
λ
= (1)
Where:
=Q overtopping rate [m3/s]
=
S
H Significant wave height [m]
=
W Ramp width 21.6 m [m]
=
r
d
λ
Ratio of energy between free surface
and device draft Ef,dr to incident wave
energy Ef,d. [m]
dkdk
dkdk
pp
d
d
p
d
d
p
d
rr
r2)2sinh(
)1(2))1(2sinh(
1+
−+−
−=
λ
=
p
k Wave number at peak period [m-1]
=
d Depth of water [m]
=
r
d Draft of device [m]
Figure 4 shows the overtopping flow
relationship, between the QN and the relative
crest freeboard (Rc/Hs). The predicted
overtopping from the old theory (Hald and
Frigaard, 2001) is presented to compare the
flow through the turbines measured on the
prototype. The scatter in these results is due
to the difference in the form of the
relationship, an exponential best fit line is
plotted for these. The measured flow through
the turbines is plotted, with the size of the
markers indicating the proportion of time the
reservoir was within 0.01 m of full. The
larger points show a full level between 50 %
and 75 % of the time.
0
0.02
0.04
0.06
0.08
0.1
0.00 0.50 1.00 1.50 2.00
Rc/Hs
Q
N
Old theory Measured Flow
Figure 4: Overtopping flow
When the crest freeboard was high (Rc/Hs >
1) the overtopping rate was generally higher

than predicted from the old formulation. In
lower crest freeboard the water flow through
the turbines was considerably less than the
predicted flow. In these cases the reservoir
was very close to the full level for over half
the period. It is probable that this loss in
flow is due to considerable spill from the
reservoir back to the sea. This is a greater
problem than expected as the flow capacity of
the prototype was less than designed. This
was caused by an incorrect setting of the
inverter speed control, causing the generators
to spin at a below optimal speed, and also
three of the generators were out of order for
this period. The flow capacity of the
prototype is thought to have been around 65%
of the designed capacity.
Figure 5 below shows the average power
produced in various sea states. The
‘Produced’ power is the electricity delivered
from the PM generators on the working
turbines. The ‘Estimated’ Power is the
electrical power which would have produced
if the dummy turbines had been producing at
the same efficiency as the actual turbines, and
if the inverter speed control had been
functioning correctly. The ‘Hydraulic’ power
is the power of the water passing through the
turbines.
0.0
2.0
4.0
6.0
8.0
10.0
12.0
0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00
Hs [m]
Average Power [kW]
Produced Estimated Hydraulic
Figure 5: Energy captured
There is a significant difference between the
hydraulic and the electrical energy. This is
due to the low overall efficiency (0.3-0.65
depending on head) of the scaled down
turbines and generators, mainly due to fixed
losses such as bearing friction. In the full
scale the overall efficiency of this stage will
be between 0.80 and 0.85. The total levels of
energy production are also low as the
platform was mostly operated at a too high
floating level and with insufficient turbine
capacity at the lower levels.
Figure 6 shows the ratio of energy captured
by the Wave Dragon. The ‘Hydraulic’
efficiency is defined as the ratio of the
average power of the water through the
turbines to the theoretical incoming wave
power across a width equal to the Wave
Dragon ramp width (see Falnes, 2002). The
‘Production’ energy is the ratio of the average
electric power generated by the operating
turbines to the same theoretical incoming
wave power.
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.70
0.00 0.50 1.00 1.50 2.00
Rc/Hs
Efficiency
Hydraulic Production
Figure 6: Efficiency

From this diagram it appears that the optimal
relative crest freeboard for energy production
is around Rc/Hs = 1.0. The hydraulic
efficiency was lower at the lower crest levels
due to the lack of turbine capacity.
Production efficiency is much lower at these
points as here much of the flow passed
Figure 7: Time series of sample record
through the dummy turbines and so did not
generate any electricity. Previous simulation
work has shown an optimal relative floating
level of Rc/Hs 0.7-0.8 with full capacity.
This is still believed to be accurate.
Currently work is being conducted by Knapp
in the Technical University of Munich on the
simulation program. This work is trying to
simulate the production of the prototype
Wave Dragon, operating as it did during this
period, in particular taking into account the
faulty turbine speed behaviour. This will
enable a good comparison of whether the
energy captured is realistic, and how large the
improvement in production would be if the
turbines had been operated at full capacity.
The time series shown in Figure 7 shows a 10
minute sample from the record of December
16th 2004 at just after 9 am. The
corresponding point lies roughly in the middle
of the records and was of a sample with Hs =
0.62 m and a floating level of 0.45 m. There
is a great deal of dummy turbine activity here,
which does not show up on the actual
electrical production. The power is quite
smooth, with a ratio between the peaks and
troughs of the estimated power of around 3.
0
5
10
15
1000 1100 1200 1300 1400 1500 1600
Time [s]
Power [kW]
Production Estimate Hydraulic
Soft issues
In operating and maintaining the device
during the testing period, the development
team has gained invaluable “soft”
experiences. This can be grouped into the
following categories:
• Maintenance: the experience has shown
that access to the device, transportation of
pieces of equipment and work on board
are only possible in relatively calm
weather conditions. This can be planned
on the basis of weather forecasts, but
major operations need to be planned
including a withdrawal scenario for the
case that the work needs to be stopped due
to bad weather.
• Corrosion: for many stationary steel parts,
a conventional epoxy paint system has
proven sufficient. For some of the moving
parts, however, more expensive corrosion
resistant materials needed to be used. This
proved to be of particular importance in
parts of the power train such as turbine
shafts and bearings. For strongly stressed

components, stress corrosion cracking
needs to be considered.
• Marine growth: On components
underneath the waterline, heavy marine
growth has accumulated within a short
time. In some components this is
acceptable, as it just means additional
weight. In others, such as the turbine draft
tubes, the layer of growth increases the
friction losses and reduces the
performance. This problem was solved by
using suitable non-toxic anti-fouling
coatings, which proved very efficient.
• Electrical equipment: A number of
components that were classified IP66
failed although they were just exposed to
rain and wind. The spray of sea water is
carried into places that seem to be
relatively well sheltered. The lesson learnt
is that sensitive equipment must
assembled with utmost care to make sure
the sealing properties are maintained, and
it needs additional protection against
splash and spray exposure. Also the effect
of corrosion attack onto the sealing
surfaces needs to be considered.
Overall, there were no problems that could
not be solved, but a lot of problems that were
not anticipated.
Conclusion
The real sea testing of the Wave Dragon
prototype has proven its seaworthiness,
floating stability and power production
potential. Operation of the device in the harsh
offshore environment has led to a number of
smaller component failures. All of these have
been investigated, and technical solutions
have been found.
An enormous quantity of data has been
collected during the testing period, which has
not yet been fully analysed. However, the
work done up to now has confirmed that the
performance predicted on the basis of wave
tank testing and turbine model tests will be
achieved in a full scale prototype.
Acknowledgements
The prototype project has only been possible
with generous support funding from the
European Commission (Contract No: ENK5-
CT-2002-00603), Danish Energy Authority,
The Obel Family Trust and the dedication of
the participating partners The first author is a
research fellow funded by the Marie Curie
WaveTrain training network.
Further information
More information can be found on the project
at the website www.wavedragon.net.
References
Falnes J. “Ocean Waves and Oscillating
Systems”, Cambridge University Press, 2002
Frigaard P. and Kofoed J.P. “Power
production experience from Wave Dragon
prototype testing in Nissum Bredning: 2003 to
2005” Aalborg University, 2005.
Hald, T. and Frigaard, P.: Forces and
Overtopping on 2. generation WD for Nissum
Bredning. Phase 3 project, Danish Energy
Agency. Project No. ENS-51191/00-0067.
Hydraulics & Coastal Engineering
Laboratory, Aalborg University, Denmark,
2001.
Kofoed J.P. “Wave overtopping of Marine
Structures – Utilization of Wave Energy”
Aalborg University, 2002.
Kofoed J.P., Frigaard P., Friis-Madsen E. and
Sørensen H.C. “Prototype testing of the wave
energy converter wave dragon” Renewable
Energy 31, 2006.
Keller J., Rohne W., Böhm C. and Knapp W.
“Wave Dragon, Development and Tests of a
Variable Speed Axial Turbine” TU München,
2001.