You are on page 1of 9

See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/224188114

Analysis of power extraction from irregular


waves by all-electric power take off
Conference Paper October 2010
DOI: 10.1109/ECCE.2010.5617893 Source: IEEE Xplore

CITATIONS

READS

18

59

4 authors:
E. Tedeschi

Marta Molinas

Norwegian University of Science and Techno

Norwegian University of Technology and Sci

54 PUBLICATIONS 534 CITATIONS

222 PUBLICATIONS 1,862 CITATIONS

SEE PROFILE

SEE PROFILE

Matteo Carraro

P. Mattavelli

16 PUBLICATIONS 96 CITATIONS

University of Padova

SEE PROFILE

232 PUBLICATIONS 6,225 CITATIONS


SEE PROFILE

All in-text references underlined in blue are linked to publications on ResearchGate,


letting you access and read them immediately.

Available from: E. Tedeschi


Retrieved on: 18 May 2016

Analysis of Power Extraction from Irregular Waves


by All-Electric Power Take Off
Elisabetta Tedeschi, Marta Molinas
Dept. of Electric Power Engineering
Norwegian University of Science and Technology
[elisabetta.tedeschi, marta.molinas]@elkraft.ntnu.no

Matteo Carraro
Paolo Mattavelli
DTG
Center of Power Electronics Systems
University of Padova Virginia Polytechnic Institute and State Univ.
mttcarraro@gmail.com
mattavelli@ieee.org

Abstract-The choice of the most suitable control strategy for


Wave Energy Converters (WECs) is often evaluated with
reference to the sinusoidal assumption for incident waves. Under
this hypothesis, linear techniques for the control of the extracted
power, as passive loading and optimum control are well-known
and widely analyzed. It can be shown, however, how their
performances are fundamentally different when irregular waves
are considered and the theoretical superiority of optimum
control is questionable under real wave conditions. Moreover,
the global optimization of WECs requires a rational design of
the power electronics equipment. This requires the analysis of
the instantaneous extracted power in addition to the average
one. In this paper the impact of irregular waves on the power
extraction when using different control techniques is analyzed in
the case of a point absorber in heave. It is also shown how a
convenient trade-off between high average power extraction and
limited power electronics overrating can be obtained by
applying simple power saturation techniques. Moreover, the
impact of power conversion efficiency on the control strategy is
analyzed.

I.

INTRODUCTION

Among renewable energies the exploitation of ocean waves


is certainly one of the most recent. In spite of its consistent
potential [1], due to high availability and stability of the
source and low environmental impact, a single leading
technology has not yet been established and different
concepts are being studied and tested worldwide. Among
them, a promising solution is represented by point absorbers,
i.e. floating buoys having small dimensions compared to the
wavelength of incident sea waves. The principle of the power
extraction lies in creating some kind of destructive
interference between the incident wave and the oscillation of
the buoy. This requires an active control of the device that
needs to be tuned according to the sea state. Thus, the control
of the motion is one of the crucial aspects to improve the
power performance of Wave Energy Converters and many
scientific contributions have been devoted to this topic [2-4].
The most of them focus on linear control techniques that are
quite easy to implement and have been systematically studied
in order to maximize the power extraction. It is worth noting
that a large number of these studies are developed under the
ideal assumption of sinusoidal incident waves [3-4]. It will be
shown in the following, however, that this can lead to mislead

978-1-4244-5287-3/10/$26.00 2010 IEEE

the conclusions when real waves are considered. On the other


hand, some studies on practical techniques to improve power
performance have also been proposed, but being limited to
few very specific applications, without necessarily claim to
generality [5-7].
Moreover, when dealing with WEC performance analysis,
the attention is usually focused mainly on mechanical and
hydrodynamic aspects, while the issues related to the sizing
and efficiency of electric and electronic equipment are often
neglected or oversimplified. The goal of this paper is at first
to underline the impact of irregular waves on the performance
of linear control techniques. Following, some considerations
on the relationship between the choice of the control strategy
and the rating of power electronic equipment are presented.
Consequently the usefulness of simple power saturation
techniques in order to maximize the average extracted power,
while avoiding excessive over ratings of the converter, is
proposed and exemplified through computer simulations. The
attention is then focused on the effects of a non ideal power
conversion efficiency and its influence on the choice of the
best control strategy is also shown. Finally, a complete model
of the system, including the electrical machine and power
electronic interface has been developed in Matlab/Simulink;
hence the proposed control solutions have been tested,
confirming the validity of the previous analyses.
II.

MODEL OF THE WAVE ENERGY CONVERTER

A. Hydrodynamic model
The following analysis is focused on a spherical point
absorber in heave [8], i.e. a single degree of freedom device
as the one schematically depicted in fig. 1, which is directly
connected to the electric Power Take Off (PTO) without any
intermediate hydraulic or pneumatic stage.
Under the assumption of plane progressive waves
propagating in an infinite water depth and if small motion is
assumed, linear theory can be applied and the hydrodynamic
diffraction model can be used to represent the interaction
between the buoy and the waves [9].

2370

Wave Energy Device


I

C
+

RL

ZL

uL
-

XL
Power
Take Off

Fig. 2. Electric analogue of a point absorber Wave Energy Converter,

Fig. 1 Simplified model of the Wave Energy Converter

A.1 Frequency-domain model


If sinusoidal incident waves (regular waves) are considered.
the system behavior can be described in the frequency
domain as [9]:


 

2 ( M + a ( )) X + j B ( ) X + K X = FE + FL (1.a)
where is the angular frequency of the incident wave, X is
the buoy position, and . denotes complex quantities. M is
the mass of the device (including also the contribution of the
generator inertia Jm, scaled by the gear ratio, n) and a is the
added mass at the considered frequency. Added mass takes
into account the water mass involved in the device movement
and depends on the radiation force caused by device
oscillation. B() is the mechanical damping (also including
the radiation resistance b, which is frequency dependent, too);
K is the hydrodynamic stiffness. Moreover FE is the
excitation wave force and FL is the force applied by the
Power Take Off. Hydrodynamic parameters for the reference
test case are listed in Tab. I
A.2 Time- domain model
In order to analyze the WEC behavior in irregular waves and
to cope with any non-linearity, a time domain model is
required, as the following one based on the Cummins
equation [10]:

(Fig. 2), which is valid as long as regular waves are assumed.


The excitation force, FE, corresponds to the sinusoidal
voltage, e, and the buoy velocity x to the current, i. Moreover,
the device mass M is represented by the inductance L, the
spring stiffness, K, by the inverse of the capacitance C and,
finally, the mechanical damping, B, by the resistance R. Thus,
the buoy, as a whole, is represented by the impedance Z.
Correspondingly, the force applied by the Power Take Off
corresponds to the load voltage uL, whose value is related to
the control parameters represented by load resistance RL and
load reactance XL, as will be better explained in Section III.
C. Model of the WEC for irregular waves tests
The time domain model of the WEC has been built from (1.b)
and is reported in Fig.3. The excitation force, FE, coming
from the waves is considered as a system input and its
derivation is described in detail in Section 3.C. The control
force applied by the PTO is a system input as well.
From the availability of the hydrodynamic coefficients of the
point absorber described in [4], the corresponding radiation
impulse response was obtained and the frequency response
derived. An identification method (Matlab command: N4sid)
was then applied to such frequency response in order to
finally model the radiation force (convolution integral) as a
4th order state space system.

( M + a ) x(t ) +

rad (t

) x (t )d + Kx (t ) = FE (t ) + FL (t )

(1.b)

In (1.b) a represents the value of added mass for t and


. Is time derivation operation. Moreover Krad is the
Radiation Impulse Response Function (RIRF) It is worth
noting that the radiation force acting on the buoy is noncausal.
B. Electric analogue of the WEC under sinusoidal
conditions
The system described by (1.a) corresponds to a mass-springdamper system and, in order to gain a better understanding of
its behavior, it can be useful to introduce its electric analogue

2371

FL(t)
FE(t)

x(t ) 1
1
s
( M + a )

+
+
- -

x (t ) 1 x(t )
s

RADIATION FORCE
STATE SPACE
MODEL

K
Fig. 3. Time domain model of the Wave Energy Converter , used in
irregular waves analyses

B. Sinusoidal incident waves


The power performance of point absorbers is well-known if
sinusoidal incident waves are considered [9] and it is here
recalled with reference to the most consolidated control
strategies for comparison purposes.
Passive loading application corresponds to have a zero
control reactive component, i.e. XL= 0.
To obtain the maximum possible extraction of average power
under this condition, the resistive component must be selected
so that:

TABLE I
DATA OF THE SELECTED SIMULATION TEST CASE
Quantity

Symbol

Unit of
measure

Value

Sinusoidal wave amplitude


Sinusoidal wave period
Significant wave height
(irreg. waves)
Energy period (irreg.waves)
Buoy radius
Buoy mass
Added mass at sinusoidal
wave frequency
Spring stiffness
Total buoy damping at
sinusoidal wave frequency

A
T

[m]
[s]

0.5
6

Hs

[m]

1.41

Te
r
M

[s]
[m]
[Kg]

6
5
267040

[Kg]

156940

[N/m]

785890

[Kg/s]

91520

RL =

R 2 + ( L

)2

(2)

On the other hand, application of optimum control implies:


X L = ( L

D. Model of the Power Take Off


The PTO is composed by an electric machine, controlled by
a bidirectional switching converter. Field Oriented Control
(FOC) is used to control the torque (i.e. the force) applied by
the PTO. One of the goals of the following power
performance analysis is to provide information about the
sizing of the generator and the power converter.
III.

CONTROL OF THE WAVE ENERGY CONVERTER

A. Linear control strategies


The simplest and most widespread control strategy for
heaving buoys is passive loading, where the force exerted by
the Power Take Off is proportional to the buoy velocity.
Acting this way, only the amplitude of the buoy motion can
be controlled. On the other hand, both the amplitude and the
phase of the motion need to be controlled in order to absorb
the absolute maximum average power. This is realized by
implementing the optimum (reactive) control, where the
applied force has a component which is proportional to the
buoy acceleration (reactive component) in addition to the one
proportional to the velocity (damping component). In this
case the theoretical condition of resonance between the
motion of the device and incident waves is achieved; however
a bidirectional power flow between the buoy and the PTO
must be allowed.

(3.a)

RL = R

(3.b)

leading to the resonance condition and, thus, to the absolute


maximum of the average power extraction.
The sinusoidal reference wave used in the following analysis
has amplitude A=0.5 m and period T = 6 s (Fig. 4.a).
C. Irregular incident waves
In order to evaluate the real performance of a WEC the
analysis of its behavior when subject to irregular waves is
mandatory. It is especially meaningful to develop some timedomain studies related to the power extraction. The stochastic
character of real waves can be well understood if the
superposition of infinite sinusoidal waves of infinitesimal
height, random phase and different frequencies is considered.
Several different models have been developed to provide a
suitable description of a real sea. Most of them use few
fundamental parameters that are considered representative of
a specific sea state to build up the energy spectrum of the sea
itself. In the following, the Bretschneider model [11] will be
adopted, where the energy spectrum (Fig. 5.a) can be
obtained as a function of the selected wave period and
significant wave height of the sea state. From the energy
spectrum the time domain expression of incident waves is
derived and, if the geometry of the buoy is known, also the
[m2s] 0.2

[m] 0.5

0.15

0
0.1

-0.5
60

61

[kW] 40

62

63

64

65

(a)

0.05

66

Pavg=23.73 kW

20
0
60

[kN]
61

[kW]

62

63

64

65

(b)

[rad/s]

(a)

400
200

66

Pavg=56.22 kW

200

-200

0
-400

-200
60

61

62

63

[s]

64

65

(c)

66

Fig. 4. Incident wave (a), instantaneous and average extracted power


applying passive loading (b) and optimum control (c)

10

20

30

40

50

[s]

60

70

80

90

100

(b)

Fig. 5. Energy spectrum according to Bretschneider model (a) and


example of an excitation force profile in irregular waves (b)

2372

100

[%]

[pu]

Pavg=5%

0.8

90

0.6
80
0.4
70
0.2
60

0
60

61

62

50

[pu]

40

[s]
63

64

65

(b)

64

65

(c)

66

Pavg=12%

0.8
0.6

30

0.4
20

0.2

10

-0.2

-0.4
0

20

Load factor (a)


40

60

80

-0.6
60

100

61

62

[s]
63

66

[%]
Fig. 6. Efficiency of electric and electronic devices (a) and its effects on
power extraction with passive loading (b) and optimum control (c)

excitation force it receives from the sea [12]; an example is


shown in Fig. 5.b.
It is worth noting that, for the purpose of control strategies
comparison, the energy period (also known as true average
period) Te = 6 s and the significant wave height Hs = 1.41 m
that are used to build the energy spectrum for the following
analyses, are selected according to a criterion of equal
energy period, Te (4.a) and equal energy transport, E (4.b)
between the cases of sinusoidal and irregular waves [11]:
(4.a)
Te = T
g 2
E sin =
A
2
E sin = E irr H s = 2 2 A
(4.b)
g
E irr =
Hs2
16
IV.

ANALYSIS OF THE POWER PERFORMANCE WITH


DIFFERENT CONTROL STRATEGIES

A. Impact of Control Strategies on the Converter Rating


In order to rationally choose the WEC control strategy, the
rating of the electric and power electronics equipment must
be taken into account. For such purpose, one interesting
parameter is the instantaneous extracted power or

equivalently the ratio, k, between peak and average power.


By simple analytical considerations it can be shown that
under sinusoidal conditions (T = 6 s, A = 0.5 m) if passive
loading is applied, the peak power always doubles the
average one. In the presented case the average power that can
be extracted is of 23.73 kW (Fig. 4.b). On the other hand, if
optimum control is adopted under sinusoidal conditions, k is
largely varying depending on the specific device and incoming waves: in the considered example the adoption of
optimum control leads to an average power of 56.22 kW and
peak to average power ratio is equal to 4.54 (Fig. 4.c). Thus,
under ideal conditions, optimum control enables the
extraction of a higher average power compared to passive
loading, at the expense of significantly higher power
electronics rating.
It is then interesting to verify if the same trend applies in
the case of irregular incident waves.
B. Power saturation effect
The analysis of the instantaneous trend of the power
extracted in irregular waves shows that it is extremely
fluctuating with sporadic very high peaks. If the conflicting
need to maximize the average extracted power, while limiting
excessive overrating of electronic equipment is also taken
into account, it is useful to consider the effect of saturation on
the instantaneous power, on the average power extraction.
Such saturation is obtained by conveniently reducing the PTO
applied force, so that the stated power limit is never
exceeded. This effect is represented in a very simplified way,
since more practically a torque limitation (possibly including
some time-dependant overloading conditions) and a power
converter current limitation are usually applied. All these
provisions are here aggregated in the PTO power saturation.
C. Efficiency of the Power Conversion
When trying to quantify the useful extractable power from
a WEC, another important consideration regards the
efficiency of the whole power transfer. It must be considered

1200

180

[kW]

[kW]

Pavg=18.38 kW

160

Pavg=28.38 kW

1000
800

140
600

120
400

100
200

80
0

60

-200

40

-400

20
0
200

-600

300

400

500

600

700

[s]

800

900

1000

1100

-800
200

1200

300

400

500

600

700

[s]

800

900

1000

1100

1200

Fig. 8. Instantaneous and average power using optimum control

Fig. 7. Instantaneous and average power using passive loading

2373

TABLE II - EFFECTS OF INSTANTANEOUS POWER SATURATION


PASSIVE LOADING
Irregular waves
Psat (kW)

OPTIMUM CONTROL
Sinusoidal waves

Pavg irr (kW)

kirr

Pavg sin (kW)

ksin

18,38

11,23

1,64

15,01

1,22

36,76

15,77

2,33

23,66

55,14

17,89

3,08

23,73

73,52

18,57

3,96

91,9

18,62

4,94

110,28

18,5

5,96

128,66

18,42

6,98

23,73

147,04

18,39

23,73

Irregular waves
Psat (kW)

Pavg irr (kW)

kirr

Pavg sin (kW)

ksin

73,52

5,43

13,5

10,26

7,17

1,55

147,04

10,62

13,85

20,84

7,06

(2)

255,2

18,19

14

56,22

4.54

23,73

(2)

367,6

22,84

16,09

56,22

(4.54)

23,73

(2)

551,4

26,59

20,74

56,22

(4.54)

23,73

(2)

735,2

27,83

26,42

56,22

(4.54)

(2)

919

28,32

32,45

56,22

(4.54)

(2)

1102,8

28,38

38,86

56,22

(4.54)

that electric machines and power converters have an


efficiency that is affected by several factors, such as the
loading conditions, the rotational speed, etc.; although a
precise evaluation of the effect of efficiency is rather complex
the simplified efficiency curve of Fig. 6a will be used as a
function of the load factor, i.e. the ratio between the extracted
power and the nominal (rated) one. As expected, the
efficiency is strongly decreased when the extracted power is a
small fraction of the nominal one. This has a different effect
depending on the chosen control strategy.
Two different aspects must be taken into account, that are
explained in the following, mainly with reference to the ideal
sinusoidal case. The first and most important one deals with
the fact that, if passive loading is applied, the power flow is
unidirectional, while if reactive control is selected the power
flow is bidirectional. This implies that, in the case of
reversible power flow, a reduction of the extracted power
needs to be taken into account in both directions (Fig. 6.c),
thus being a disadvantage for reactive control both in regular
and in irregular waves.
Moreover, under certain conditions, the application of
reactive control can lead the system to work in a low load
region for a longer part of the period, thus contributing to
disadvantage reactive control over passive loading. Its better
to point out, however, that this second effect strongly depends
on the specific test case (i.e. peak over average extracted
power) and it may become negligible in irregular waves. In
the case exemplified by Figs 6.b-c the average power loss due
to non-ideal efficiency results of 5% for passive loading and
12% for optimum control.
V.

Sinusoidal waves

SIMULATION RESULTS

The previous analyses have been verified by


Matlab/Simulink simulations, according to the test
conditions reported in TABLE I for both regular and irregular
waves. Although targeted to a specific test case, the analyses
can be extended to different system ratings and operating
conditions.

A. Linear controls performance in regular and irregular waves


At first, sinusoidal ideal conditions were tested to be used as a
reference case and to clearly emphasize the difference in the
power performance when applying the different control
strategies. As already mentioned, if the sinusoidal reference
wave is assumed, passive loading leads to extract 23.73 kW,
corresponding to only the 42% of the power that can be
theoretically extracted by applying optimum control.
Thus, for the sinusoidal case, a clear advantage in using
optimum control would be apparent. Then, in order to
compare the power performance of passive loading and
optimum control in irregular waves to the corresponding ones
under ideal conditions, waves force profiles were created, in
the form of profiles of 1000 seconds each (being the spectrum
discretization frequency resolution f = 0.001Hz).
While the data about average power extraction are the same
irrespective of the specific force profile that is applied (when
all of them are derived from the same spectrum), the
information about instantaneous power are related to the
specific test case since phases at different frequencies are
randomly generated. Thus the evaluation of the peak power
(and, consequently, of power ratio) must be performed on a
statistical base [10], considering the results coming, in our
case, from the simulation of 80 different force profiles.
The first set of simulations was aimed at the quantification of
the power performance when passive loading is used. In this
case the value of the damping component applied by the PTO
is fixed and equal to the ideal damping (2) for a sinusoidal
wave whose period equals the modal period [11] of the
energy spectrum. The average power can be quantified into
18.38 kW (Fig. 7). When compared to the corresponding
sinusoidal case, it shows how, due to irregular waves, a
reduction of almost 1/4 of the average power is to be
expected.
The same kind of test was developed while applying
optimum control. Also in this case the reactive and damping
coefficient are selected as the ones maximizing power
absorption under sinusoidal conditions when a sinusoidal
incident wave having period corresponding to the modal

2374

22

[kW]

30

[kW]

21

passive
loading

25

20

20

19

15

18

10

17

optimum control
0

100

200

300

400

500

600

700

800

900

16

1000 1100

[kW]

0.1

0.2

0.3

0.4

0.5

0.6

XL [%]

0.7

0.8

0.9

Fig. 10. Average power extracted with non-unity efficiency as a function


of the reactive component (evaluated as a percentage of the value used in
optimum control). Resistive component is the one used in passive loading

Fig. 9. Average power extraction as a function of the power saturation


level in presence of unity (dashed line) and non unity (continuous line)
efficiency of the power conversion, for passive and optimum control.

period of the selected energy spectrum is considered (3.a-b).


The corresponding power absorption is reported in Fig. 8.
The total average extracted power is now of 28.38 kW, which
is even more significantly reduced with respect to the
corresponding sinusoidal case. Then the statistical analysis on
irregular waves shows that, when passive loading is applied,
the peak to average power value is in the range 7.7-17.1 and a
power peak of 314 kW can be reached. In the case of reactive
control k is between 25.2 and 58.3 and the peak power
absorption (over the 80 tested cases) exceeds 1.6 MW.
Thus the irregular wave study clearly confirms the need for a
consistent overrating of the electric and electronic devices.
B. Test on power saturation effect
The second part of the analysis is focused on the effects of
power saturation, representing the maximum power that can
be handled by the electronic equipment and it is performed
with reference to a specific force profile.
Eight different saturation levels are chosen for both passive
loading and optimum control and analyzed with respect to
both irregular waves and corresponding sinusoidal waves.
Results are reported in Tab. II. It can be observed that in
irregular waves a saturation of the maximum instantaneous
power of 62% (from Psat= 147.04 kW to Psat = 55.14 kW)
leads to a loss of average power only of 3% for passive
loading, while in the case of optimum control a corresponding
reduction of 67% in the maximum power (from Psat= 1102.8
kW to Psat = 367.6 kW) results in a 19.5% loss of average
power, thus meaning that a significant reduction in the power
electronics rating can be achieved with limited average power
drop. With reference to the optimum control it is also worth
noting that, if an even smaller saturation of 43% (from Psat=
255.2 kW to Psat = 147.04 kW) is applied under sinusoidal
conditions, it leads to an average power loss of 63%, thus
meaning that the beneficial effect of a power saturation is
specifically related to irregular waves and could be not
appreciated if only sinusoidal waves are considered. From the
data of passive loading applied in irregular waves it can be

also noted how the adoption of a non-linear control (i.e.


saturation) can lead to the absorption of a higher average
power than the correspondent (unconstrained) linear
technique, under certain conditions, as already reported in
literature [4]. Finally it is interesting to note the fact that, if
the same limit for peak power is assumed for both passive
loading and optimum control (e.g. Psat = 73.52 kW), the
implementation of passive loading results much more
convenient than reactive control, if irregular waves are
considered.
C. Tests on non-ideal efficiency effect
The goal of the following set of simulations is to evaluate
the effect of a non ideal efficiency of electric and electronic
devices (fig, 6.a) on the power conversion and its specific
impact on the different control strategies.
In Fig. 9, the average power extraction as a function of the
power saturation level is reported for both ideal and non-ideal
conversion efficiency, in the case of passive loading and
optimum control.
The first important thing to underline is that, irrespective of
the control technique that is applied, the effect of non-ideal
conversion efficiency is of determining a point of maximum
in the average power extraction, corresponding to a specific
power saturation level. This means that even if a higher rate
of the power electronics equipment was allowed, no gain
would follow in the final power extraction. The reason is that
when the Power Take Off has a consistent overrating due to
irregular waves, it works mainly in low load conditions and
consequently the whole power conversion becomes more and
more inefficient.
It was already mentioned with reference to Tab. II that, in
case of very low saturation level, passive loading leads to
higher average power extraction than reactive control. When
considering also the effect of non-unity efficiency the
conclusion is even more decisive. When comparing the nonideal curves of Fig. 9 for passive loading and reactive control,
it can be clearly seen that the maximum power extraction
obtained from passive loading (Pavg = 16.7 kW) is more than

2375

the maximum power extraction that can be reached by


optimum control even in its most favorable condition (Pavg =
15 kW). Moreover, the highest power extraction in the
passive loading case can be reached at expense of a
consistently reduced power electronic equipment (Psat = 75
kW), compared to the optimum control option (having its
maximum at Psat= 500 kW). Thus, in the considered test case,
optimization of the average power extraction can be obtained
by designing electric and power electronics equipment whose
power rating is around 100 kW and by choosing passive
loading as the most convenient control strategy.
The proven superiority of passive loading over optimum
control when non-ideal efficiency is taken into account, leads
to wonder if an intermediate control solution can ensure a
higher average power extraction than passive loading under
the same conditions. A specific test was carried out by
adopting a saturation level of 110 kW. The resistive
component of the control was fixed as in the case of passive
loading while a reactive component being an increasing
fraction of the one prescribed by optimum control was
added. Average extracted power was then evaluated as the
mean value of several sets of simulations. As can be seen
from Fig. 10, when a reactive component being 50% of the
one adopted in optimum control is applied, an average power
of 21.8 kW can be obtained, corresponding to a 23% increase
with respect to pure passive loading. Consequently, an
intermediate control solution including a suitable reactive
component is advisable even under non-unity efficiency
conditions and must be carefully optimized to improve the

whole power performance.


D. Detailed system simulation
Based on the proposed analysis, detailed simulation model
including the electrical drive and the grid-interface converter
has been developed. For this specific example, the electric
machine is an asynchronous generator with a nominal power
of 110 kW and a rated speed of 1487 rpm, while other
electrical parameters are reported in Table III. The machine
model is implemented in the dq reference frame and the Field
Oriented Control has been adopted, as shown in Fig. 11 [13].
Two-level Voltage Source Inverter (VSI) has been adopted
both for the electrical drive and for the grid-connected
converter with fsw=8 kHz. The second one behaves as an
Active Front End (AFE), where an inner dq current control
imposes the line currents to be in phase with the line voltages
and an outer control regulates the dc-bus voltage. The control
algorithm has been implemented in the dq rotating reference
frame and the angle , needed for the transformation, has
been generated by a Phase-Looked-Loop (PLL). The control
used both for the grid-connected VSI and for the electrical
drive is based on well-established knowledge available in the
literature. Main converter parameters are also listed in Table
III. The PI regulators of the current loops of both the
electrical drive and the AFE are tuned to achieve a bandwidth
of 500 Hz with a phase margin of 60. The PI regulating the
DC link voltage is designed for a bandwidth of 20 Hz and a
phase margin of 70 and a feedforward action is added in the
dc bus voltage loop (i.e. a term proportional to the generated
power is added at the output of the PI_Vdc) to reduce dc-link
voltage fluctuation.

Fig. 11 Block diagram of the detailed system simulated model , including main control components

2376

control and design of wave energy converters. At first the


effect of irregular waves when two different linear control
techniques are applied has been shown. Consequently, from
the consideration of the highly varying power profiles, the
convenience of adopting power saturation provisions to
reduce the rating of power electronics equipment without
excessive drop in the average power extraction has been
proved by computer simulations. Finally the non-unity
efficiency of the real electric power conversion has been
considered in order analyze its different impact on the
selected control techniques.

TABLE III PARAMETERS OF THE FULL-SYSTEM SIMULATED


Parameter

Value

Unit of
measure

Description

Pn
Vn
Lm
Ls
Lr
Rs
Rr
p
Rl
Ll
Udc
Cdc
n
rw
Jm

110
400
10,4
10,6
10,6
21,6
12,3
2
28,0
900
800
3.3
40
0,1
2,3

[kW]
[V]
[mH]
[mH]
[mH]
[m]
[m]
[m]
[H]
[V]
[mF]
[m]
[kgm2]

Generator nominal power


Generator nominal voltage
Generator mutual inductance
Generator stator inductance
Generator rotor inductance
Generator stator resistance
Generator rotor resistance
Generator pole pairs
EMI filter resistance
EMI filter inductance
DC link voltage
DC link capacitor
Gear box ratio
Pignon radius
Generator mechanical inertia

REFERENCES
[1]
[2]
[3]
[4]

One example of the results obtained with the detailed model


is reported in Fig. 12. Fig. 12a shows the irregular wave
profile and Fig. 12b the torque applied by the electrical
machine when passive loading is used. It can be observed that
the DC voltage (Fig. 12.c) reflects the low frequency
oscillations of sea waves and such fluctuation is present also
in the grid current waveform (Fig. 12.d). In this specific
example, the dc-link capacitor is not providing any energy
smoothing and the power fluctuation generated by the PTO is
directly injected into the grid. From fig. 13 it can be noted
that the current is injected into the grid with a unity power
factor, as required.
The detailed model confirms the analysis previously reported
in Fig. 9. As an example, using a power saturation equal to
110 kW, the average extracted power using the detailed
model is 16.4 kW, as marked by the square shown in Fig. 9.
CONCLUSIONS

VI.

[5]

[6]

[7]

[8]
[9]
[10]

[11]

This paper has analyzed the effect of irregular waves in the

[12]

0
-0.5

Torque [Nm]

50

DC voltage [V]

60

65

70

75
time [s]

80

85

90

95

100

300

200
0
-200
50

Grid current [A]

55

ea

200

55

60

65

70

75
time [s]

80

85

90

95

iag

100

100

[V],[A]

Wave
amplitude [m]

[13]
0.5

T.W. Thorpe, An Overview of Wave Energy Technologies: Status,


Performance and Costs, Wave Power Moving Towards Commercial
Viability, IMECHE Seminar, London, UK, 1999
U.A. Korde, "Control System Applications in Wave Energy Conversion",
Procs of the MTS/IEEE Oceans 2000, Providence, Sept. 2000, pp. 1817-24
J.K.H.Shek, D.E.Macpherson, M.A.Mueller,J.Xiang,"Reaction force control
of a linear electrical generator for direct drive wave energy conversion", IET
Renew. Power Generation, Vol.1, N.1, March 2007, pp. 17 -24
J. Hals, T. Bjarte-Larsson, J.Falnes, Optimum reactive control and control
by latching of a wave-absorbing semisubmerged heaving sphere, Proc. of
the 21th International Conf. on Offshore Mechanics and Artic Engineering,
June 23-28, 2002, Oslo, Norway, pp.1-9
D.Valerio, P.Beirao, M. J.G.C. Mendes, J.S da Costa "Comparison of
control strategies performance for a Wave Energy Converter", Proc. of the
16th Mediterranean Conference on Control and Automation, 25-27 June
2008, pp. 773-778
E.A.Amon, A.A.Schacher, T.K.A. Brekken, "A Novel Maximum Power
Point Tracking Algorithm for Ocean Wave Energy Devices", Proceedings of
the IEEE Energy Conversion Congress and Exposition (ECCE09), San Jos,
California, 20-24 September 2009, pp.2635-2641
M. F. P. Lopes, J. Hals, R.P.F Gomes, T. Moan, L. M. C. Gato, A. F. de O.
Falco, Experimental and numerical investigation of non-predictive phasecontrol strategies for a point-absorbing wave energy converter. Ocean
Engineering, 2009, Vol. 36, No. 5, pp. 386-402.
K. Budal, J. Falnes, A resonant point absorber of ocean-wave power
Nature, Vol. 256, 1975, pp. 478-479.
J. Falnes, Ocean Wave Energy and Oscillating Systems: Linear Interaction
Including Wave-Energy Extraction. Cambridge Univ. Press; 1 Ed. 2002.
P. Ricci, J.B. Saulnier, A.F. de O. Falcao, M.T. Pontes "Time-Domain
models and Wave Energy Converters Performance Assessment", Proc. of
the 27th International Conference on Offshore Mechanics and Artic
Engineering (OMAE08), Estoril, Portugal, 15-20 June 2008, pp.1-10
W. H. Michel, "Sea Spectra Revisited", Marine Technology, Vol. 36 N.4
winter 1999, pp.211-227
K.A. Ansari, N.U. Khan, Computation of Environmental Forces for Use in
offshore vessel Dynamic Analysis, International Journal of Energy
R. Krishnan, Electric Motor Drives: Modeling, Analysis, and Control,
Prentice Hall, 2001

810
800

-100

790
55

60

65

70

75
time [s]

80

85

90

95

100

-200
100
0

-300

-100
50

55

60

65

70

75
time [s]

80

85

90

95

70.605

100

70.61

70.615

70.62

70.625

70.63

70.635

70.64

[s ]

Fig. 12.Sea waves profile (a), generator torque (b), DC voltage reference
and DC actual voltage (c), current injected into the grid in phase a(d)

2377

Fig. 13. Detail of the voltage and current of phase a at the grid
section