» ASK THE EXPERTS

From Blue to Green

if control can be used to get energy energy typically increases with in-
I n this issue of IEEE Control Systems
from ocean waves. Does anyone at creasing latitude from the equator,
Magazine we ask Ted Brekken, Belinda
IEEE Control Systems Magazine both in the northern and southern
Batten, and Ean Amon to respond to a
know anything about this? Thanks hemispheres. This resource varies
query on the uses of systems and control
for your help. between summer and winter, with
technology in wave-energy production.
Ted, Belinda, and Ean: Thanks for peak available energy in the winter
Ted, Belinda, and Ean are all involved
asking. Control can be used for a va- months in the northern hemisphere.
in wave energy projects at Oregon State
riety of reasons in the context of wave Data gathered from the National
University.
energy, such as maximizing the energy Data Buoy Center indicates that the
extraction from waves, minimizing yearly average power extractable on
excessive wear or damage to devices the West Coast of the United States
Q. I’m really interested in “green en- from unusually large waves or storm typically ranges from 30 kW/m to
ergy,” but I also thought my controls events, and coordinating arrays of 40 kW/m, where kW/m is units of
course was neat. I was wondering wave energy converters (WECs). kilowatts of power per meter of wave
It’s helpful to keep in mind that front [2]. A schematic of annual aver-
Digital Object Identifier 10.1109/MCS.2011.941960
waves are most energetic on the age wave power across the world is
Date of publication: 16 September 2011 west coasts of continents, and wave shown in Figure 1.

Average Annual Wave Power (kW/m) <5

5–10
10–15
15–20
20–30
30–40
40–60
> 60

FIGURE 1 Average annual wave power worldwide. Average annual wave power in kW/m (kilowatts per meter of wave front) is shown at
locations around the world. The highest average power tends to be at the west coasts of continents. Average power tends to increase
with increasing latitude in both hemispheres. (Image courtesy of OCEANOR and ECMWF [1].)

18 IEEE CONTROL SYSTEMS MAGAZINE » OCTOBER 2011 1066-033X/11/$26.00©2011IEEE

 Several states are instituting re- Oscillating Water Column As shown in Figure 3(a), water from
newable portfolio standards (RPSs) An OWC uses an inverted, enclosed the basin is drained out through a
specifying the amount of electricity chamber to enable the heaving of the turbine, and the energy is extracted
to be provided by renewable energy water surface, which is effectively the in the same manner as a low-head hy-
sources. Oregon, for example, has a floor of the chamber, to alternately dropower system. Wave Dragon Ltd.
state RPS of 25% of the total electricity compress and decompress air trapped has deployed an overtopping device,
consumed in the state generated from in the chamber. The air rushes in and shown in Figure 3(b).
renewables by 2025. With large-scale out of a small opening in the cham-
commercial development, wave en- ber, and a unirotational wind turbine, Oscillating Body
ergy could provide approximately half such as a Wells turbine, is located at The last category, the oscillating body,
of this requirement. In fact, the entire the opening to convert the energy to uses an arrangement of physical struc-
energy needs of some island commu- electricity. This process is depicted in tures that rock, heave, or pivot in the
nities found across the world could be Figure 2(a). An example of an OWC waves. Examples of this category are
provided by wave energy. by Oceanlinx that can be used in deep shown in Figure 4: the Pelamis by
water is shown in Figure 2(b). OWCs Pelamis Power in (a), the Oyster by
TECHNOLOGY OVERVIEW can also be used at locations where Aquamarine Power in (b), the Sea-
The earliest recorded interest in uti- ocean energy must be dissipated to Ray by Columbia Power Technologies,
lizing ocean waves for on-shore use reduce erosion and damage, for ex- shown in (c), and the PowerBuoy by
is a patent for a wave-energy-con- ample, at a jetty or at the shore. Voith Ocean Power Technologies in (d). In
version device filed by a French in- Hydro Wavegen, Ltd. has deployed particular, the OPT PowerBuoy is one
ventor in 1799. Since then, there has such a device, the Land Installed Ma- type of point absorber, which we use
been a plethora of ideas on how best rine Powered Energy Transformer, as a WEC example here.
to convert wave energy. In contrast to or Limpet 500, on the island of Islay, Figure 5(a) shows a point absorber
the modern wind industry, which is Scotland, since 2000; this device is WEC that was developed and tested by
dominated by three-bladed horizon- shown in Figure 2(c). Columbia Power Technologies, in col-
tal-axis wind turbine designs, there is laboration with Oregon State Universi-
yet no single dominant technological Overtopping ty and with funding support from the
paradigm for ocean-wave-energy con- Overtopping devices use walls or U.S. Navy and Columbia Power Tech-
version. However, past and current arms to gather and focus waves to a nologies. This point absorber is called
designs can be classified within three single point. At this point, the focused the L10, for “linear 10 kW.” The L10 has
main categories, namely, oscillating waves spill up and over a ramp into two main bodies, namely, a float and a
water column (OWC), overtopping, a basin that has a mean water level spar. The float, which is connected to a
and oscillating body. higher than the surrounding ocean. magnet array, is designed to be excited

Air Is Turbine
Punched
Through
Turbine
Mean
Sea Level

Air Is Turbine
Punched
Through
Turbine
Mean
Sea Level

(a) (b) (c)

FIGURE 2 Oscillating water columns. (a) is a concept sketch of an oscillating water column showing air movement through the turbine
generating electricity with the rising and falling of the wave. (b) shows the MK2 oscillating water column deployed in deep water at Port
Kembla by Oceanlinx. (c) shows the 500-kW Limpet deployed on Islay, Scotland. This device has been deployed since 2000 and is an
example of an OWC used on land. [(a) and (b) are courtesy of Oceanlinx. (c) is courtesy of Voith Hydro Wavegen, Ltd.]

OCTOBER 2011 « IEEE CONTROL SYSTEMS MAGAZINE 19

 by the waves in the vertical, that is, capture theory for an ocean wave
Overtopping heave direction. The spar, which is the energy converter was formalized in
Reservoir
smaller-diameter central cylinder, is the late 1970s by several researchers,
designed to be relatively unexcited by for example see [3]–[5]; a histori-
Turbine Outlet the waves. Twelve armature coils are cal overview is provided in [6]. The
(a) embedded within the spar. The spar mathematics for wave energy extrac-
is designed to provide a stationary tion are developed in [6], and the
reaction reference for the float and its theory presented here is based on
attached magnet section. This is an ex- that development.
ample of a direct drive system in that A few key assumptions are needed
the force and velocity imparted by the to simplify the illustrative example.
waves are converted directly to electri- For the L10 point absorber, it is as-
cal power without intermediate stages sumed that the central tall and thin
involving hydraulics or pneumatics. cylindrical spar that houses the gen-
Figure 5(b) shows the L10 deployed at erator armature is moored in such a
(b) Yaquina Head, Oregon. manner that it remains stationary in
FIGURE 3 Overtopping devices. (a) shows the waves. In other words, the spar
the mechanics of an overtopping device, POINT ABSORBER PLANT MODEL is assumed to be a stable reference
where water collected by the device re- To focus our discussion on control and reaction point for the generator.
enters the ocean through a turbine outlet of WECs, we now turn to a discus- In this case, the generator relative ve-
thereby generating electricity. (b) shows
sion of a model of a point absorber. locity, that is, velocity of the magnets
the Wave Dragon, an example of an over-
topping device. [(a) and (b) are used with In particular, we first develop the past the armature, is equal to the ver-
permission of Wave Dragon Ltd.] linear WEC model. Optimal energy tical velocity with which the float is

(a) (b)

(c) (d)

FIGURE 4 Oscillating body wave energy converters (WECs). (a) – (d) show examples of WECs categorized as oscillating bodies. In
(a), the Pelamis prototype is shown during sea trials. In (b), the Oyster 1 wave power device is shown before installation at sea. Part
(c) shows the SeaRay deployed in Puget Sound, Washington, in spring 2011. (d) shows PowerBuoy 150 deployed in Hawaii. [(a) is
courtesy of Pelamis Wave Power; (b) is courtesy of Aquamarine Power; (c) is courtesy of Columbia Power Technologies; and (d) is
courtesy of Ocean Power Technologies, Inc.]

20 IEEE CONTROL SYSTEMS MAGAZINE » OCTOBER 2011

 Buoy

Tether
Power
Line
Float
Ocean
Floor
Stationary
Generator Coils
Siding
Magnet Assembly

(a) (b)

FIGURE 5 The L10 Point Absorber wave energy converter (WEC). (a) shows the operation of the L10 point absorber WEC. This
device was developed and tested by Columbia Power Technologies in collaboration with Oregon State University and with funding
support from the U.S. Navy and Columbia Power Technologies. The heaving motion of the buoy generates electricity as the magnet
assembly moves past stationary generator coils. (b) shows the L10 deployed at Yaquina Head, Oregon, in September 2008. [(a) is
courtesy of Smithsonian Magazine [20].]

moving, that is, the float vertical ve- where Fe is the excitation force im- hydrodynamic concept that water is
locity. It is also assumed that the WEC parted on the float by the wave, Fr accelerated with and around a mov-
is constrained to move in only the is the radiation force, Fb is the buoy- ing body, thus causing the body to
heave direction. ancy force, and Fpto is the force upon behave as if it has more mass than it
Note that there are six possible the float from the power take-off. The actually does. For many simple body
degrees of freedom (DOFs) in a wave float mass is m, the float vertical veloc- geometries, the frequency-dependent
energy system, namely, heave; surge, ity is V, and s is the first derivative in added mass A 1 s 2 can be approximat-
defined as motion back and forth, in the frequency domain, hence sV 1 s 2 is ed well as a constant. The damping
the direction of wave propagation; the vertical acceleration. An addition- B(s) is the velocity-dependent force
sway, which is side-to-side motion al friction force that is proportional to term resulting from the creation of
orthogonal to the direction of wave velocity is sometimes included to ac- waves by the moving WEC.
propagation; roll; pitch; and yaw. The count for fluid friction and friction in For a cylindrical float geometry
full formulation of the equations of the power take-off; this term is omit- that is neither fully submerged nor
motion for all 6DOFs can be cast as six ted here for simplicity. fully removed from the water, the
scalar second-order equations or one It has been said that a good wave buoyancy force Fb is proportional to
matrix equation [6]. For both simplic- energy absorber is also a good wave the amount of displaced water and
ity and also since most point absorb- maker. This concept can be under- thus appears as the spring term
ers are designed to have a dominant stood if one recognizes that the pro-
response in heave, only the heave di- cess of absorbing waves can also be Fb 1 s 2 5 2k V 1 s 2 / s, (3)
rection is illustrated here. viewed as producing waves that de-
While a generator was used in the structively interfere with incoming where V 1 s 2 / s is the vertical position.
L10, the idea of a more general power waves. Thus, devices that are good at The power take-off force Fpto is
take-off will be used in the follow- making waves will be good at destroy- reacted against the spar, which as
ing discussion. In general, the power ing incoming waves, hence absorbing. explained above is assumed to be sta-
take-off mechanism could be hydrau- The creation of waves by the WEC is tionary relative to the waves; the pow-
lic or, depending on the application, represented by the radiation force Fr. er-take force Fpto is the control input
may never require electricity produc- This force can be expressed as variable.
tion. Linear wave theory is also as- In this model the excitation force Fe
sumed. The heave equation of motion Fr 1 s 2 5 2 1 sA 1 s 2 1 B 1 s 22 V 1 s 2 , (2) is regarded as an independent distur-
for the float in the frequency domain bance, but it can be estimated with a
is then where A 1 s 2 is the frequency-depen- Kalman filter, for example. For simu-
dent added mass and B 1 s 2 is the lation purposes, it is often useful to
Fe 1 s 2 1Fr 1 s 2 1Fb 1 s 2 1Fpto 1 s 2 5 sV 1 s 2 m, frequency-dependent damping. The have a time series for Fe. It is possible
(1) added mass is the embodiment of the to create an approximation of Fe from

OCTOBER 2011 « IEEE CONTROL SYSTEMS MAGAZINE 21

a water-surface-elevation time series Zpto, opt 1 s 2 5 2Z*i 1 s 2 . (8) relation is expressed in the time do-
using a method derived from the lin- main as
earized Morison equations for force In fact, by making the analogy of
`
on a submerged body in a moving force to voltage, velocity to current,
vopt 1 t 2 5 3 hFV 1 t2t 2 fe 1 t 2 dt. (13)
fluid [7]. This approximation assumes and mechanical impedance to electri- 2`
the frequency-dependent added mass cal impedance, an analogous equiva-
and frequency-dependent damping lent circuit for (7) can be drawn,
can be approximated by constants A^ consisting of the series connection of IMPLEMENTATION ISSUES
and B^ . This procedure yields the excitation force, intrinsic imped-
ance, and apparent power take-off- Causality
Fe 1 s 2 5 1 s2A^ 1 sB^ 1 k 2 h 1 s 2 , (4) produced mechanical impedance; the Equations (10) and (13) present equiv-
“voltage drop” across the series con- alent methods for specifying the op-
where h is the water-surface eleva- nection is 2Fpto. timal control of the WEC, assuming
tion with respect to the mean water- In common terms, the optimal pow- that “optimal” refers only to energy
surface elevation, that is, the vertical er capture condition requires the pow- capture. However, several practical
position of the water at the float. er take-off to appear as a virtual mass, implementation issues arise.
damper, and spring such that the virtu- First and foremost, implementing
OPTIMAL CONTROL al mass and virtual spring, that is, the (10) and (13) in practice requires future
It is useful to rewrite (1) by collecting imaginary components of Zpto, cancel knowledge of the vertical velocity v 1 t 2
the acceleration, velocity, and position the actual mass and spring terms. Con- and the excitation force fe 1 t 2 . Consider
terms as sequently, the power take-off damping (10) where the impulse response zi 1 t 2
is equal to the actual float radiation of the intrinsic mechanical impedance
Fe 1 s 2 1 Fpto 1 s 2 5 Zi 1 s 2 V 1 s 2 , (5) damping. The optimal control law for is the response of radiation, buoyancy,
the power take-off force then becomes and inertial forces—as well as friction
where Zi 1 s 2 is the mechanical intrinsic and other parasitic forces, if included
impedance. If we assume the power Fpto, opt 1 s 2 5 Zpto, opt 1 s 2 V 1 s 2 in the model—to a velocity impulse.
take-off force Fpto is controlled such 5 2Z*i 1 s 2 V 1 s 2 , (9) However, (10) requires convolution
that it has components proportional with the complex conjugate of the
to vertical acceleration, vertical ve- which in the time domain can be ex- impulse response in the frequency
locity, and vertical position, then the pressed as the convolution domain, which corresponds to time-
power take-off appears to the system reversal in the time domain. Because
`
as a mechanical impedance as well. In zi 1 t 2 is physical and therefore causal,
Fpto, opt 1 t 2 5 23 zi 1 2 1 t 2 t 22 v 1 t2 dt,
other words, the action of the power 2`
it follows that zi 1 2t 2 is anticausal.
take-off appears to an outside force (10) Therefore, determination of Fpto, opt 1 t 2
as augmented inertia, damping, and at time t requires future knowledge of
stiffness that add to or subtract from where v 1 t 2 is the vertical velocity. We v 1 t 2 as reflected in (10).
the actual inertia, damping, and stiff- discuss these equations further in the Similarly with (13), Re 5 Zi 1 s 2 6 is a
ness of the system. This term can thus next section on causality. real, even function in the frequency
be written as Alternatively, the optimal control domain. Therefore, 1/ 1 2 Re 5 Zi 1 s 2 6 2
law can be specified in terms of the 5 HFV 1 s 2 is a real, even function, and
Fpto 1 s 2 5 Zpto 1 s 2 V 1 s 2 , (6) desired velocity Vopt 1 s 2 . If (9) is satis- thus the impulse response hFV 1 t 2 is a
fied, then (5) becomes real, even function and therefore non-
Fe 1 s 2 5 1 Zi 1 s 2 2 Zpto 1 s 22 V 1 s 2 . (7) causal. Determining vopt 1 t 2 at time t
thus requires future knowledge of the
Fe 1 s 2 5 Zi 1 s 2 Vopt 1 s 2 2 Fpto, opt 1 s 2
Note that, in the convention used time-domain excitation force fe 1 t 2 .
here, negative power from the power 5 1 Zi 1 s 2 1 Z*i 1 s 2 2 Vopt 1 s 2 There are various ways to address
take-off denotes power take-off force 5 2 Re 5 Zi 1 s 2 6 Vopt 1 s 2 , (11) the causality problem. First, we could
acting in opposition to velocity and 1 determine causal transfer functions
therefore ocean energy converted to Vopt 1 s 2 5 Fe 1 s 2 that approximate Z*i 1 s 2 and HFV 1 s 2 .
2 Re 5 Zi 1 s 2 6
electrical energy, that is, generation of To obtain such transfer functions, we
power. 5 HFV 1 s 2 Fe 1 s 2 (12) assume that the radiation force (3)
Recognizing that power is force can be approximated as frequency
times velocity, optimal power capture where HFV denotes the transfer func- independent. This assumption im-
by the power take-off now becomes an tion from the excitation force Fe to the plies that the frequency-dependent
impedance-matching problem, that is optimal vertical velocity Vopt. This added mass A 1 s 2 and damping B 1 s 2

22 IEEE CONTROL SYSTEMS MAGAZINE » OCTOBER 2011

are approximated as constants A^ and near resonance—can be large. Thus the er available in the waves than the sys-
B^ over the frequencies of interest. In reactive power requirement can be tem can handle, a detuning strategy
this case, much larger than the actual average that allows the WEC to operate within
power delivered. In such cases, the its limits may be required. A widely
Fpto, opt 1 t 2 5 instantaneous power maximum is accepted method of power shedding
t
#
1 m 1 A^ 2 v 1 t 2 2 B^ v 1 t 2 1 k3 v 1 t2 dt much larger than the average power. for the different types of WECs has
2`
(14) This situation results in poor efficiency not yet been established.
and the need to specify the power take-
and off system for a much higher power Latching
than the device actually averages. Latching is a nonlinear control ap-
vopt 1 t 2 5 fe 1 t 2 / 1 2B^ 2 , (15) To avoid this situation, a nonopti- proach in which a mechanical,
mal condition must be accepted, such hydraulic, or electromechanical mech-
both of which are realizable in real as augmenting the control law (14) to anism holds the point absorber float
time. make Fpto 5 0 when Fpto would other- for approximately one quarter of the
A second approach is to predict the wise be acting in the same direction as incoming wave, allowing the force on
future excitation force fe 1 t 2 based on velocity. Another alternative would be the float to reach high levels. The float
current and past observations using to remove the derivative and integral is then released, and it moves quickly
a model, for example ARIMA, and to terms and choose a suboptimal damp- to the other end of the stroke, where it
then use the prediction with the non- ing value replacing B^ to maximize is held again to be released in the op-
causal control laws. energy capture in this context. This posite direction. This approach tends
damping value can be shown to be to phase-synchronize the float force
Reactive Power Requirement and float velocity while achieving
Another control issue is the necessity Bsubopt 5 " 1 v 1 m 1 A 2 2 k / v 2 2 1 B^ 2, high float velocities. However, in ap-
of “back driving,” which refers to mo- plications for large commercial-scale
mentary flow of energy from the WEC where v represents the dominant point absorbers, forces may approach
power take-off to the ocean, even frequency in the wave climate. This hundreds of thousands of newtons or
though the energy on average flows type of damping is called nonreactive more. A mechanical system that can
in the opposite direction. This power damping. latch and release forces of this magni-
is delivered to the ocean and then re- tude is difficult to construct. Therefore,
turned in a form of reactive power. AREAS FOR DEVELOPMENT there is interest in a “soft” latching ap-
The frequency of the reactive power is Many areas in wave energy control proach, in which the float is partially
twice the WEC motion frequency and are ripe for further research. A brief restrained at a force magnitude within
thus is altogether different from grid description of some of these areas is the ability of the generator and power
reactive power; typical periods for provided below. takeoff mechanism [8]–[10].
WEC motion are approximately 10 s.
The reactive power requirement can Life-Extending Control Array Control
be seen in (14), where a nonzero m, A^ , Maintenance costs are expected to WECs in an array may be able to ex-
or k requires the power take-off force be significant for WECs. Therefore, it ploit antenna concepts, such as beam
and velocity to occasionally have the may be worthwhile to develop control focusing, to shelter or intensify power
same sign, which in the convention strategies that balance energy conver- at certain WECs within the array. Ad-
used here signifies power delivered to sion, that is, revenue, with damage ditionally, WECs at the front of the ar-
the ocean through the power take-off. mitigation, that is, costs, to maximize ray may be able to pass information to
This scenario can be considered an overall revenue. Optimal control ap- other WECs about the incoming waves,
investment of energy in the mass and proaches with objective functions that thus providing enhanced estimates of
spring effect emulated by the power take these aspects into consideration future waves for control. Placement of
take-off that is later returned, minus likely have a role in this context [21]. WECs within an array to either lever-
losses, and is similar to the energy age or diminish device interaction is
stored and returned by reactive circuit Power Shedding also an active research topic [11]–[16].
elements such as inductance and Wind turbines typically have a rated-
capacitance. For WECs operating in power region of operation in which Nonlinear and Robust Control
frequency ranges far from their natural the power available in the resource Large waves and devices other than
resonance, the “matching” effort exceeds their rated power. In this case, simple point absorbers exhibit sig-
required by the power take-off—that the excess power is typically shed by nificant nonlinear behavior. Robust
is, the emulated mass and spring effect controlling blade pitch. Similarly, for and fast nonlinear control and model-
necessary to bring the system to or situations in which there is more pow- ing approaches are needed. Adaptive

OCTOBER 2011 « IEEE CONTROL SYSTEMS MAGAZINE 23

control approaches, such as maximum CAREER award and the IEEE Power [6] J. Falnes, Ocean Waves and Oscillating Systems.
Cambridge, U.K.: Cambridge Univ. Press, 2002.
power point tracking [17], [18], may & Energy Society 2011 Outstanding [7] M. Patel, Dynamics of Offshore Structures. Lon-
be a practical and effective approach Young Engineer award. don, U.K.: Butterworth, 1989.
to dealing with nonlinearities and Belinda A. Batten (bbatten@engr. [8] M. Eriksson, J. Isberg, and M. Leijon, “Theory
and experiment on an elastically moored cy-
changing operating conditions. As orst.edu) is director of the Northwest lindrical buoy,” IEEE J. Oceanic Eng., vol. 31, pp.
more sophisticated models of the flu- National Marine Renewable Energy 959–963, Oct. 2006.
id-structure interactions are coupled to Center at Oregon State University and [9] A. Babarit and A. Clément, “Optimal latch-
ing control of a wave energy device in regular
device models to more accurately pre- a professor of dynamics and control. and irregular waves,” Appl. Ocean Res., vol. 28,
dict forces, reduced-order model and She received the Ph.D. in 1991 from pp. 77–91, Apr. 2006.
control designs will be important [19]. Clemson University in mathematical [10] J. Falnes and P. M. Lillebekken, “Budal’s latch-
ing-controlled-buoy type wave-power plant,” in
sciences. While an assistant profes- Proc. 5th European Wave Energy Int. Conf., Univ. Col-
Model Predictive Control sor, she received an Alexander von lege Cork, Ireland, 17–20 Sept. 2003, (A. Lewis and
Model predictive control provides a Humboldt fellowship and visited Uni- G. Thomas, Eds. Cork, Ireland: Hydraulics and
Maritime Research Centre, 2005) pp. 233–244.
natural framework for handling sev- versitaet Trier, Germany. Her research [11] P. C. Vicente, A. F. de O. Falcão, L. M. Gato,
eral aspects of WEC control, such as has focused on modeling and control and P. A. Justino, “Dynamics of arrays of floating
survivability, by the ability to handle of distributed parameter systems, in- point-absorber wave energy converters with in-
ter-body and bottom slack-mooring connections,”
and consider velocity, force, and posi- cluding computational algorithms for Appl. Ocean Res., vol. 31, no. 4, pp. 267–281, 2009.
tion constraints and optimal energy reduced-order controllers. Her cur- [12] M. Folley and T. Whittaker, “The effect of
capture by providing prediction as rent research includes dynamics and sub-optimal control and the spectral wave cli-
mate on the performance of wave energy con-
part of the framework [22]. control of micro air vehicles, as well as verter arrays,” Appl. Ocean Res., vol. 31, no. 4, pp.
optimization and control of wave en- 260–266, 2009.
CONCLUSIONS ergy devices. She is an associate editor [13] P. Siddorn and R. E. Taylor, “Diffraction and
independent radiation by an array of floating
In summary, as the demand for elec- for the SIAM Advances in Design and cylinders,” Elsevier Ocean Eng., vol. 35, no. 13, pp.
tricity from renewable energy sources Control book series, a member of the 1289–1303, 2008.
increases, interest turns to new ways AIAA, ASME, and SIAM, as well as a [14] G. Duclos and A. Clément, “Wave propaga-
tion through arrays of unevenly spaced verti-
of meeting that demand. The solution Senior Member of IEEE. In 2002, she re- cal piles,” Elsevier Ocean Eng., vol. 31, no. 13, pp.
to energy independence will likely take ceived the Outstanding Young Alumni 1655–1668, 2004.
shape as a portfolio of technologies, Award from the College of Science and [15] A. O. Falcão, “Wave-power absorption by
a periodic linear array of oscillating water col-
and wave energy has promise to be one Engineering at Clemson University. umns,” Elsevier Ocean Eng., vol. 29, no. 10, pp.
of those technologies. Commercializ- Ean A. Amon is an assistant pro- 1163–1186, 2002.
ing this energy source requires systems fessor in the Electrical Engineering [16] A. Kiprakis, A. Nambiar, D. Forehand, and
A. Wallace, “Modelling arrays of wave energy
and control tools to understand the re- Department at Oregon State Univer- converters connected to weak rural electricity
source and how to efficiently maximize sity. He received the B.S., M.S., and networks,” in Proc. Int. Conf. Sustainable Power
the energy extracted from waves. Ph.D. from Oregon State University Generation and Supply (SUPERGEN’09), 2009,
pp. 1–7.
in 2004, 2007, and 2010, respectively. [17] E. Amon, A. Schacher, and T. K. A. Brekken,
AUTHOR INFORMATION His research interests include power “A maximum power point tracking algorithm
Ted K.A. Brekken is an assistant pro- electronics and control in high-per- for ocean wave energy devices,” IEEE Ind. Appli-
cat. Mag., to be published.
fessor in energy systems at Oregon formance motor drives and converter [18] E. Amon, A. Schacher, and T. K. A. Brekken,
State University. He received the B.S., topologies applied to electric vehicles “A novel maximum power point tracking algo-
M.S., and Ph.D. from the University of and renewable energy systems. rithm for ocean wave energy devices,” in Proc.
Energy Conversion Congr. and Exposition (ECCE),
Minnesota in 1999, 2002, and 2005, re- Sept. 2009, pp. 2635–2641.
spectively. He studied electric vehicle REFERENCES [19] K. A. Evans and B. A. Batten, “Reduced order
motor design at Postech in Pohang, [1] S. Barstow and G. Mork. (2008, Mar.). World- compensators via balancing and central control
waves wave energy map. Fugro OCEANOR AS, Tech. design for a structural control problem,” Int. J.
South Korea, in 1999. He also studied Rep. [Online]. Available: http://www.oceanor. Control, vol. 83, pp. 563–574, 2010.
wind turbine control at the Norwegian no/related/59149/wavemap_background.pdf [20] E. Rusch, “Catching a wave, powering an
University of Science and Technology [2] P. Lenee-Bluhm, R. Paasch, and T. Ozkan- electrical grid?” Smithsonian Mag., July 2009.
Haller, “Characterizing the wave energy re- [21] C. Stillinger, T. K. A. Brekken, A. von Jouanne,
(NTNU) in Trondheim, Norway in source of the US Pacific Northwest,” Renewable R. Paasch, D. Naviaux, K. Rhinefrank, J. Prudell,
2004–2005 on a Fulbright scholarship. Energy, vol. 36, pp. 2106–2119, 2011. A. Schacher, and E. Hammagren, “WEC proto-
His research interests include control, [3] K. Budal and J. Falnes, “Optimum operation type advancement with consideration of a real-
of improved wave-power converter,” Marine Sci. time damage accumulation algorithm,” in Proc.
power electronics, and electric drives; Commun., vol. 3, pp. 133–150, 1977. PowerTech, Trondheim, Norway, June 2011.
specifically, digital control techniques [4] S. H. Salter, D. C. Jeffery, and J. R. M. Taylor, [22] T. K. A. Brekken, “On model predictive
applied to renewable energy systems. “The architecture of nodding duck wave power control for a point absorber wave energy con-
generators,” Naval Architect, pp. 21–24, Jan. 1976. verter,” in Proc. PowerTech, Trondheim, Norway,
He is codirector of the Wallace En- [5] D. V. Evans, “A theory for wave-power ab- June 2001.
ergy Systems and Renewables Facility sorption by oscillating bodies,” J. Fluid Mech.,
(WESRF). He is a recipient of an NSF vol. 77, pp. 1–25, 1976.

24 IEEE CONTROL SYSTEMS MAGAZINE » OCTOBER 2011