Hardware in the loop simulation of a dielectric
elastomer generator for oscillating water column
wave energy converters
G. Moretti, G.P. Papini Rosati, M. Fontana
PERCRO-SEES, TeCIP Institute Scuola SuperiorSant’Anna
Pisa, Italy
g.moretti@sssup.it, g.rosatipapini@sssup.it m.fontana@sssup.it
R. Vertechy
Department of Industrial Engineering University of Bologna
Bologna, Italy rocco.vertechy@unibo.it
Abstract— This paper presents a small-scale test-bench that
can be employed for the study of a new class of Oscillating Water Column (OWC) Wave Energy Converters (WECs) with Power Take Off (PTO) unit based on Dielectric Elastomer Generators (DEGs).
Such a test-bench is designed to perform Hardware-In-the-Loop (HIL) simulations of a small-scale DEG PTO prototype, that includes dedicated power and control electronics, with a real-time model that emulates OWC plant hydrodynamics.
The paper illustrates the theoretical model that is assumed to describe the dynamics of the OWC plant, the hardware employed to replicate the response of the DEG PTO and the algorithms for the implementation of the closed-loop controller that performs the HIL simulations. Some preliminary tests are reported considering a floating OWC collector and a simple energy-harvesting control law for the DEG PTO that is based on the measurement of the water level into the OWC structure.
Keywords—OWC; Dielectric Elastomer; Wave Energy; PTO; I. INTRODUCTION
Dielectric Elastomer Generators (DEGs) are deformable variable capacitors, made of rubber-like dielectric material and compliant electrodes, that can be used to directly convert mechanical energy into direct current electricity. DEGs present several features that make them suitable for the conversion of wave energy into electricity such as: large energy densities, good energy conversion efficiency that is rather independent of cycle frequency, easiness in manufacturing and assembling, high shock resistance, silent operation and low cost.
In the last years, Wave Energy Converters (WECs) based on DEGs have attracted a lot of attention due to their potential
performance.
The first prototype of a buoy equipped with a DEG has been conceived by Chiba [1]. More recently, the company SBM offshore developed a submerged tubular WEC, made by several ring-shaped DEGs, that has been tested in a wave-tank [2]. In addition, in the context of the EpoSil project [3], researchers from Bosch GmbH are studying new DEG solutions for ocean energy harvesting. In our previous works, we have proposed the concept of the Poly-OWC, that is a WEC based on an Oscillating Water Column (OWC) system equipped with a Circular Diaphragm DEG (CD-DEG) as Power-Take-Off (PTO) system [4], [5].
In this paper, we describe a small-scale laboratory test-bench that has been designed to investigate the PTO system of Poly-OWC devices at a subscale prototype level [6]. Such an experimental setup is able to impose programmed time-varying pressures on the CD-DEG, which replicate the operating conditions of the Poly-OWC air chamber, thereby making it possible to perform dry-run tests of the PTO unit. Specifically, the test-bench is employed to run Hardware-In-the-Loop (HIL) simulations of the small-scale CD-DEG PTO prototype, which are implemented by coupling it with an actuator, sensors and a model-based real-time controller that makes it possible to emulate the hydrodynamics of the OWC as excited by incoming sea waves.
Results from preliminary experiments are reported to show the test-bench functionality and measurement capabilities. Upon testing with the set-up, the obtained results can be scaled up to real system dimensions, thereby making it possible to estimate potential energy production and conversion efficiency of Poly-OWC systems.
The paper is organized as follows. Section II is dedicated to describe the working principle of the full-scale Poly-OWC system and the general architecture of the down-scaled HIL simulation system. Section III provides a description of the hydrodynamic model that is employed in the HIL simulation, while Section IV illustrates the details of the employed hardware. Section V shows an example of a preliminary test. The work presented in this paper is developed in the context of the project
PolyWEC (www.polywec.org ), a FP7 FET Energy project. The research leading to these results has received funding from the European Union Seventh Framework Programme [FP7/2007-2013] under grant agreement n° 309139.
The work has been partly supported by the Project “Studi e valutazioni sulla produzione di energia elettrica dalle correnti marine e dal moto ondoso” funded by the Italian Ministry of Economic Development.
II. SYSTEM ARCHITECTURE
A. Full-scale Poly-OWC system
A Poly-OWC is a partially submerged (usually referred to as “collector”) featuring opened to the sea action, and an upper part DEG membrane and forming an air chamb partially encloses a column of water that incident wave field at the bottom and to pressure at the top. As the waves break o structure, wave-induced pressure oscillations interface cause the reciprocating motion of with a concomitant compression-expansion o in the upper chamber, and the resulting inf the CD-DEG membrane. To generate e charges are put on the CD-DEG electrodes w is maximally inflated (namely, when it is exp the membrane deflates, CD-DEG capacitance makes the charges increase their electric converting the work done by the air-chamb DEG membrane into direct current electricity The potentialities of Poly-OWCs have investigated with reference to different t suitable for onshore [5] and nearshore/offsho respectively (see Figure 1).
Figure 1. Schemes of possible implementations onshore (left) and nearshore/offshor B. Architecture and control of the HIL simu
With the aim of testing the potentialit systems and, in particular, for the assessment PTO and the related energy harvesting c simulator has been developed. As depicted simulator comprises a cylindrical air cham volume that is closed at the top with a ful DEG (with power and control electronics in bottom with a movable piston. This latter is and controlled so as to emulate the reciproca water column inside the OWC structure tha bottom by the incident ocean waves and variable air-chamber pressure. Emulation motion is performed via a real-time hydrod the specific OWC structure in exam, which r a sensor that measures the relative pressure chamber and outputs the reference position ( water column free surface that needs to b actuated piston.
A schematic representation of the conside provided in Figure 2. Details on the employ E
d hollow structure an immersed part t closed by a CD-ber. The structure
is exposed to the o the chamber air on the Poly-OWC s at the underwater the water column, of the air entrapped flation-deflation of
lectricity, electric when the membrane
panded in area). As e decreases, which potential, thereby er pressure on the y. been theoretically types of collector ore [8] applications s of the Poly-OWC: re (right). lator ies of Poly-OWC t of their CD-DEG controller, a HIL d in Figure 3, the mber with variable lly-functional CD-ncluded) and at the
s properly actuated ating motion of the at is excited at the at the top by the
of water column dynamic model of eceives input from
(p) within the air (z) of the emulated be tracked by the ered HIL system is yed hydrodynamic
model and of the developed tes and IV.
Figure 2. Scheme of the Hardw
III. HYDRODYN In this paragraph, a simplif presented, which relies on line flow theory. The so-called “ used; that is, the free surface of assumed to remain flat during behaves as a rigid piston, the alternate compression and dec volume [9], [10].
Indicating with z the vertica water with respect to its equilib the water column is described b
(
) ( )
0
t
M z∞+
³
Kτ ξ
− zξ ξ ρ
d +where
• M is the added mass o frequency;
• the convolution integral a generated by the water col • ȡ is the water density, S i water column, and the pr stiffness associated with th • p=pc-pa is the difference
the chamber (pc) and the a
• Fe(IJ) is the time-series o
force, which is the sum o contributions.
The convolution kernel K(IJ induced force in response to an water column and its expres domain is
( )
( )
( )
ˆ ˆ ˆ
r add
K
ω
=Bω
+ ¬iω
ªMω
where Ȧ is the angular frequenc radiation damping and added frequency.
st-bench are given in Sections III
ware-In-the-Loop (HIL) system.
NAMIC MODEL
fied model of OWC dynamics is ear hydrodynamics and potential rigid piston” approximation is f the water within the chamber is
the operation. Such free surface e motion of which induces the ompression of the overhead air al position of the free surface of brium position, the dynamics of by the following motion law:
( )
e
gS z pS F
ρ
⋅ = − +τ
(1)of the water column at infinite accounts for the radiated waves lumn motion;
is the cross-sectional area of the roduct kb=ȡgS is the hydrostatic
he free surface displacement. between the air pressure inside atmospheric pressure (pa).
of the wave-induced excitation of Froude-Krylov and diffraction
IJ) expresses the radiated wave-n impulsive displacemewave-nt of the ssion in the Fourier frequency
M∞º
− ¼ (2)
cy, and Br and Madd represent the
Irregular waves can be described as a monochromatic regular waves, whose represented by a wave spectrum [10]. As a general form of the wave induced force, F waves, can be approximately expressed as sinusoidal addenda; namely
( ) (
)
, 1 cos n e w i i i i i F A ω ω τ ϕ = =¦
Γ + In equation (3), ī(Ȧ) is a frequency-depende coefficient, iji are random phases, and Aw,i arethe different monochromatic waves that ar function of the wave spectrum, SȦ(Ȧ), as it fo
( )
, 2
w i i i
A = ΔωSω ω
where ǻȦi is the step between the i-th and (
frequency value.
The wave spectrum, SȦ(Ȧ), can be expr
wave parameters; for instance, the significan and the energy period, Te. In this work, a P
spectrum is used, according to which SȦ(Ȧ) re
( )
262.9 2 4 5exp 1054(
4 4s e e
Sω ω = H T−ω− − T−ω−
If, instead, regular waves are considere takes the following simple form:
( ) ( )
2 cos , with = 2 e
F =H Γ
ω
ωτ
ω π
Twhere H (wave height) and T (wave characteristic parameters of regular waves, been set to zero (as this does not prov generality).
Notice that equation (1) is linear except which, depending on the specific considered linear. Such a non-linearity makes it neces OWC dynamics in the time domain (by differential equations solvers) rather than domain.
The hydrodynamic parameters involved model (namely, M, Madd(Ȧ), Br(Ȧ) and
calculated using a frequency-domain B Method (BEM) code (for instance, WAM AQWA). The convolution integral in computational expensive and can be repla state-space approximation, the parameters found with a frequency-domain identification
IV. HARDWARE SETUP The test-bench that has been developed fo simulations of Poly-OWC systems is schema Such test-bench comprises: a pneumatic pis
a superposition of distribution is a consequence, the Fe, in irregular sea s a finite sum of (3)
ent wave excitation e the amplitudes of re expressed as a ollows: (4) i+1)-th considered ressed in terms of nt wave height, Hs, Pierson-Moskowitz eads as [10]:
)
(5)d, the wave force
(6) period) are the and the phase has voke any loss of t for the term in p
PTO, may be non-ssary to solve the
y using ordinary in the frequency d in the described d ī(Ȧ)) can be
oundary Element MIT and ANSYS
equation (1) is aced with a linear
of which can be n procedure [11].
or performing HIL atized in Figure 3. ston; a small-scale
CD-DEG prototype; a high-vo DEG charge and discharge; a r control and data acquisition. system are provided below.
Figure 3. Schematic of t A. Electromechanical sub-syst To simulate the Poly-OW
custom made pneumatic cylind cylinder tube (with internal d made of Delrin®; a clamping that is able to host interchang can be made of different m dielectric and for the complian minimized by using a plana screwed fixture between C appropriate pneumatic circul cylinder. The piston is actuate (P01-37x120F/200x280-HP b encoder that is used to measur sensor (MPX12 by Freescale S cylinder tube is used to meas between cylinder chamber and accuracy CCD laser displac Keyence), which is mounted o used in order to measure CD-pneumatic cylinder is mounted is introduced in order to com piston and motor slider.
B. CD-DEG.
The CD-DEG is a circular different layers of dielectric materials (see the exploded sch
The capacitance of the CD difference is applied betwe
oltage (HV) electronics for CD-real-time unit for overall system
Specific details of each
sub-the experimental test-rig. tem.
WC air chamber we employ a der that features: a polycarbonate diameter e = 130mm); a piston
system on the top of the piston geable CD-DEG specimens that materials (both for the elastic
nt electrodes). Air leakages are r annular gasket with a tight CD-DEG and cylinder, and
ar seals between piston and ed via a brushless linear motor by LinMot), with embedded
re piston position, z. A pressure Semiconductor) installed on the ure the differential pressure, p, ambient air. A speed high-cement sensor (LK-G152 by
on top of the cylinder head, is -DEG tip displacement, h. The d vertically and an elastic spring mpensate for the weight of both
membrane obtained by stacking c and conductive deformable
eme reported in Figure 4). D-DEG varies when a pressure een its compliant electrodes.
Specifically, the capacitance increase as the m since the surface of the electrodes increases of the dielectric layers decreases (indeed, the the dielectric layers is incompressible). The for the CD-DEG capacitance is reached when fully inflated and the minimum value o configuration.
Figure 4. Schematic of the CD-DEG a C. HV Electronics.
A custom made HV driving electronics h to control CD-DEG charge and discharge Figure 5, the considered driving electronics power supply (10C24-P125 by UltraVolt) relays (HM12-1A69-150 by MEDER ele supplementary HV capacitors.
In Figure 5, C2 is a capacitor that is perm
in parallel with the CD-DEG (whose varia indicated in the figure as Cd). The capacitor C
absorb and store part of the charge, i.e. lim voltage, during the generation phase of the An appropriate choice for the value of C2 m
increase the energy harvested in each cycl CD-DEG to work as close as possible breakdown limit [12]. The capacitor C1 is lar
employed to deliver the charge to the CD-Specifically, the charging of the CD-DEG place when the relay S2 is closed (with S1 be same time).
In the circuit, resistor R2 (R2 = 20MΩ) is resistive component to the load of the powe resistor R1 (R1 = 1MΩ) is used to limit the occurs during CD-DEG discharge as the relay
To measure the electric potential differ CD-DEG electrodes, a custom made HV prob Figure 5) has been implemented that feature resistance (nearly 50GΩ), which drastically charge from the CD-DEG electrodes, and which is obtained thanks to a capacitor comp In order to minimize the current leakag electronics, all the HV wirings and comp encapsulated via thick layers of silicone g Raytech) and acrylic tape (VHB 4905 by identical HV probe is employed to measu across the capacitor C1. This voltage re charging phase of the CD-DEG (and of C2) m
precisely evaluate the initial value of Cd.
membrane inflates and the thickness e material used for e maximum value n the membrane is ccurs for its flat
assembly.
has been developed e. As depicted in s comprises a HV ), three HV reed ectronic) and two manently connected
able capacitance is
C2 is introduced to
miting the rise of harvesting cycles. makes it possible to le by pushing the to its dielectric rge in value and is
DEG (and to C2).
(and of C2) takes
eing opened at the s used to provide a er supply, whereas e peak current that
y S3 is closed. ence, Vd, between
be (not depicted in es very high input limits the drain of large bandwidth, pensation network. ge of the power
onents have been gel (Magic gel by 3M®). A second ure the voltage V1
ading during the makes it possible to
D. Controller implementation
A MatLab® xPC Target® rea real-time target machine by Sp the real-time hydrodynamic m governs the motion of the pist status (charging voltage and st
Figure 5. Simplified schema
The developed test-bench makes it possible to implem controlling the harvesting cy simple regulation strategy co electrical activation to the CD-the actual value of its capacitan measured or can be estimate following sensor readings: air the piston, displacement of th advanced versions of the C measures can also be employed can further improve the energy generator.
A basic controller that can experiments is described in the V. PRELIM This section reports prel conducted considering the hyd Poly-OWC based on the OW company Sendekia. Details of t the dimensioning of this WEC match the diameter of the O hardware cylinder tube, a sma geometric scale factor of 1:77 computation of the hydrodynam The employed CD-DEG s dielectric elastomer membrane together two layers of a com sensitive acrylic tape (VHB 4 onto a polycarbonate ring an compliant carbon conductive gr 846). The dimension of th polycarbonate ring is e = 130 equi-biaxial with value λp = undeformed virgin condition) o membrane is t0 = 1.0mm, whic
al-time machine (Performance peedGoat®) is employed to run model of the OWC plant, which ton, and to control the charging tatus of relays) of the CD-DEG.
atic of the HV Driving circuit.
h, with its sensing equipment, ment different algorithms for ycles of the CD-DEG. A first onsists in imposing a state of -DEG that is only a function of nce. Such a value can be directly ed through at least one of the
r-chamber pressure, position of e tip of the CD-DEG. In more CD-DEG control, the acquired
d to build suitable estimators that y harvesting performance of the be used to conduct preliminary following section.
MINARY TESTS
liminary tests that have been drodynamic model of a floating WC collector developed by the the hydrodynamic model and on
are reported in [8]. In order to WC collector with that of the all-scale OWC prototype (with 7) has been considered for the mic parameters.
specimen is made by an active e, which is obtained by gluing mmercial double-sided
pressure-4905 by 3M®), that is stretched nd coated on both sides with rease electrodes (MG-Chemicals he internal diameter of the
mm; acrylic film pre-stretch is 3; the initial thickness (in the of the active dielectric elastomer ch excludes the thickness of the
compliant electrodes. A picture of the CD-D has been employed for this study is reported i Preliminary results have been collected control law that triggers the activation statu on the basis of the knowledge of the wate inside the chamber. Specifically, the CD-DEG
z reaches a maximum (or minimum) and it i z crosses the zero.
Figure 7 reports the experimental result three tests of the same CD-DEG subje activation voltages and excited by the sam polychromatic incident waves, generated usin equation (5) with the following parameters:
Te = 0.91s, that, at full scale, correspond
Te = 8.0s according to the rules of Froude
plots reported in figure are for the different measured with the set-up: the voltage (V1) ac the voltage (Vd) across the CD-DEG (and C pressure (p) with respect to atmosphere inside emulates the pressure inside the OWC position (z) of the piston (that emulates the column of water inside the OWC). In each p the three tests performed with different activ represented by red, blue and black solid lines
The reported results highlight the followin
•
In all wave periods and for all the thre electrical activation, voltage Vd increase deflates. That is, irrespective of wave and of the value chosen for the activ considered control law based on the m water level inside the OWC structure makes it possible to convert pneum electricity via the CD-DEG.•
Activation of the CD-DEG determines inside the cylinder chamber, with the dr the activation voltage is higher. Th indicates that the mechanical work abso to deflate the CD-DEG is lower than th the inflation phase, with the difference b energy that is extracted by the CD-DE that are due to material viscoe consequence, the energy that can be har DEG is larger as the activation voltage i•
The motion amplitude of the piston activation voltage of the CD-DEG i indicates that the energy extraction perf DEG can significantly affect the dyna inside the OWC structure. This demons DEG can extract a significant portion of available in the incident water waves.DEG prototype that in Figure 6. d implementing a us of the CD-DEG er column level z G is charged when s discharged when ts acquired during ected to different me emulated set of ng the spectrum of : Hs = 4.55cm and to Hs = 3.5m and scaling [13]. The t variables that are cross capacitor C1; C2); the differential e the cylinder (that air-chamber); the free surface of the plot, the results of vation voltages are
. ng:
ee test cases, upon es as the CD-DEG height and period vation voltage, the measurement of the e is effective and matic energy into a drop in pressure rop being larger as his pressure drop orbed by the piston hat provided during being the electrical EG (plus the losses elasticity). As a rvested by the CD-is higher. diminishes as the is increased. This formed by the CD-amics of the water strates that the
CD-f the energy that is
Figure 6. Picture of the inflat
VI. CO This paper presented an ex used to study the control of Os Wave Energy Converters (WE Diaphragm Dielectric Elastome The test-rig makes it pos operating conditions of this k hardware-in-the-loop simulation After a presentation of the set-up and global controller arc study, which is based on the converter, is provided to sho energy harvesting performan specimen made with acrylic d and carbon conductive grease e
In the future, hardware-in performed using the developed CD-DEG power take-off sy controllers for WECs that are b
ted CD-DEG during operation. ONCLUSIONS
xperimental test-rig that can be scillating Water Column (OWC) ECs) based on inflated Circular er Generators (CD-DEGs). ssible to simulate the practical
kind of WECs by performing ns.
e purposely developed hardware chitecture, an experimental case emulation of a floating OWC ow the dynamic response and nce of a particular CD-DEG
dielectric elastomer membranes electrodes.
n-the-loop simulations will be d test-rig to optimize and assess ystems and energy harvesting
Figure 7. Example of the measured variables time-series during experiments with simulated polychromatic waves (Hs = 4.5cm,
ACKNOWLEDGMENT
The authors would like to acknowledge the support of: • SELMAR company (Italy) for the development of the
hardware setup;
• SENDEKIA company (Spain) for the concession of the geometry of their offshore OWC concept;
• Marco Alves form WavEC (Portugal) for providing the hydrodynamic parameters of the OWC collector.
REFERENCES
[1] Chiba, S., Waki, M., Kornbluh, R., & Pelrine, R., “Innovative power generators for energy harvesting using electroactive polymer artificial muscles,” in Proc. SPIE 6927(1), 692715–692719 (2008).
[2] Jean, P., Wattez, A., Ardoise, G., Melis, C., Van Kessel, R., Fourmon, A., Barrabino, E., Heemskerk, J., and Queau, J.P., “Standing wave tube electro active polymer wave energy converter,” in Proc. of SPIE Vol 8340, 83400C–1 (2012).
[3] Scherber, B., Grauer, M., & Köllnberger, A. “Electroactive polymers for gaining sea power”. In SPIE Smart Structures and Materials+ Nondestructive Evaluation and Health Monitoring (pp. 86870K-86870K). International Society for Optics and Photonics (April, 2013). [4] Vertechy, R., Fontana, M., Rosati Papini, G. P., & Bergamasco, M.;
“Oscillating-water-column wave-energy-converter based on dielectric elastomer generator”. Proc. SPIE 8687, Electroactive Polymer Actuators and Devices (EAPAD) 2013, 86870I (April, 2013).
[5] Vertechy, R., Rosati, G. P. P., & Fontana, M. (2015). “Reduced Model and Application of Inflating Circular Diaphragm Dielectric Elastomer Generators for Wave Energy Harvesting”. Journal of Vibration and Acoustics, 137(1), 011016.
[6] Vertechy R., Fontana M., Rosati Papini G. P., & Forehand D. “In-tank tests of a dielectric elastomer generator for wave energy harvesting”. Proc. SPIE 9056, Electroactive Polymer Actuators and Devices (EAPAD) 2014, 90561G (March 8, 2014).
[7] Righi, M., Vertechy, & R. Fontana M., “Experimental Characterization 0f a Circular Diaphragm Dielectric Elastomer Generator”. In Proceedings of the ASME 2014 Conference on Smart Materials, Adaptive Structures and Intelligent Systems (SMASIS 2014), Newport, Rhode Island, USA September 8–10, 2014.
[8] Moretti G., Papini Rosati G.P., Alves M., Grases M., Vertechy, R., & Fontana M., “Analysis And Design of an Oscillating Water Column Wave Energy Converter with Dielectric Elastomer Power Take-Off”. To appear in ASME 2015 34rd International Conference on Ocean, Offshore and Arctic Engineering (OMAE2015).
[9] Gomes, R. P. F., Henriques, J. C. C., Gato, L. M. C., & Falcao, A. F. O. (2012). “Hydrodynamic optimization of an axisymmetric floating oscillating water column for wave energy conversion”. Renewable Energy, 44, 328-339.
[10] Henriques, J. C., Chong, J. C., Falcão, A. F., & Gomes, R. P. “Latching Control of a Floating Oscillating Water Column Wave Energy Converter in Irregular Waves”. In ASME 2014 33rd International Conference on Ocean, Offshore and Arctic Engineering (OMAE 2014).
[11] Yu, Z., & Falnes, J. (1995). “State-space Modelling of a Vertical Cylinder in Heave”. Applied Ocean Research, 17(5), 265-275.
[12] Huang, J., Shian, S., Suo, Z., & Clarke, D. R. (2013). “Maximizing the Energy Density of Dielectric Elastomer Generators Using Equi-Biaxial Loading”. Advanced Functional Materials, 23(40), 5056-5061. [13] Holmes, B. (2009). “Tank testing of wave energy conversion systems”.
