Standardised performance tests, phase 1

Transcription

1 Marine Renewables Infrastructure Network Infrastructure Access Report Infrastructure: ECN Hydrodynamic and Ocean Engineering Tank User-Project: WRAM MkII Standardised performance tests, phase 1 Swirl Generators Limited Status: Version: Date: Draft 24-Aug-215 EC FP7 Capacities Specific Programme Research Infrastructure Action

2 ABOUT MARINET MARINET (Marine Renewables Infrastructure Network for emerging Energy Technologies) is an EC-funded network of research centres and organisations that are working together to accelerate the development of marine renewable energy - wave, tidal & offshore-wind. The initiative is funded through the EC's Seventh Framework Programme (FP7) and runs for four years until 215. The network of 29 partners with 42 specialist marine research facilities is spread across 11 EU countries and 1 International Cooperation Partner Country (Brazil). MARINET offers periods of free-of-charge access to test facilities at a range of world-class research centres. Companies and research groups can avail of this Transnational Access (TA) to test devices at any scale in areas such as wave energy, tidal energy, offshore-wind energy and environmental data or to conduct tests on cross-cutting areas such as power take-off systems, grid integration, materials or moorings. In total, over 7 weeks of access is available to an estimated 3 projects and 8 external users, with at least four calls for access applications over the 4-year initiative. MARINET partners are also working to implement common standards for testing in order to streamline the development process, conducting research to improve testing capabilities across the network, providing training at various facilities in the network in order to enhance personnel expertise and organising industry networking events in order to facilitate partnerships and knowledge exchange. The aim of the initiative is to streamline the capabilities of test infrastructures in order to enhance their impact and accelerate the commercialisation of marine renewable energy. See for more details. Partners Ireland University College Cork, HMRC (UCC_HMRC) Coordinator Sustainable Energy Authority of Ireland (SEAI_OEDU) Denmark Aalborg Universitet (AAU) Danmarks Tekniske Universitet (RISOE) France Ecole Centrale de Nantes (ECN) Institut Français de Recherche Pour l'exploitation de la Mer (IFREMER) United Kingdom National Renewable Energy Centre Ltd. (NAREC) The University of Exeter (UNEXE) European Marine Energy Centre Ltd. (EMEC) University of Strathclyde (UNI_STRATH) The University of Edinburgh (UEDIN) Queen s University Belfast (QUB) Plymouth University(PU) Spain Ente Vasco de la Energía (EVE) Tecnalia Research & Innovation Foundation (TECNALIA) Belgium 1-Tech (1_TECH) Netherlands Stichting Tidal Testing Centre (TTC) Stichting Energieonderzoek Centrum Nederland (ECNeth) Germany Fraunhofer-Gesellschaft Zur Foerderung Der Angewandten Forschung E.V (Fh_IWES) Gottfried Wilhelm Leibniz Universität Hannover (LUH) Universitaet Stuttgart (USTUTT) Portugal Wave Energy Centre Centro de Energia das Ondas (WavEC) Italy Università degli Studi di Firenze (UNIFI-CRIACIV) Università degli Studi di Firenze (UNIFI-PIN) Università degli Studi della Tuscia (UNI_TUS) Consiglio Nazionale delle Ricerche (CNR-INSEAN) Brazil Instituto de Pesquisas Tecnológicas do Estado de São Paulo S.A. (IPT) Norway Sintef Energi AS (SINTEF) Norges Teknisk-Naturvitenskapelige Universitet (NTNU) Page 2 of 17

3 DOCUMENT INFORMATION Title Standardised performance tests, phase 1 Distribution Public Document Reference MARINET-TA1-WRAM MkII User-Group Leader, Lead William Dick Swirl Generators Limited Author [Optional: Insert address and contact details] User-Group Members, Chris Signorelli Swirl Generators Limited Contributing Authors Sean O Callaghan Trinity College Dublin Infrastructure Accessed: ECN Hydrodynamic and Ocean Engineering Tank Infrastructure Manager Dr Sylvain Bourdier (or Main Contact) REVISION HISTORY Rev. Date Description Prepared by (Name) Approved By Infrastructure Manager Status (Draft/Final) 1 28/8/15 William Dick Draft 2 2/9/15 William Dick Dr Sylvain Bourdier Final Page 3 of 17

4 ABOUT THIS REPORT One of the requirements of the EC in enabling a user group to benefit from free-of-charge access to an infrastructure is that the user group must be entitled to disseminate the foreground (information and results) that they have generated under the project in order to progress the state-of-the-art of the sector. Notwithstanding this, the EC also state that dissemination activities shall be compatible with the protection of intellectual property rights, confidentiality obligations and the legitimate interests of the owner(s) of the foreground. The aim of this report is therefore to meet the first requirement of publicly disseminating the knowledge generated through this MARINET infrastructure access project in an accessible format in order to: progress the state-of-the-art publicise resulting progress made for the technology/industry provide evidence of progress made along the Structured Development Plan provide due diligence material for potential future investment and financing share lessons learned avoid potential future replication by others provide opportunities for future collaboration etc. In some cases, the user group may wish to protect some of this information which they deem commercially sensitive, and so may choose to present results in a normalised (non-dimensional) format or withhold certain design data this is acceptable and allowed for in the second requirement outlined above. ACKNOWLEDGEMENT The work described in this publication has received support from MARINET, a European Community - Research Infrastructure Action under the FP7 Capacities Specific Programme. LEGAL DISCLAIMER The views expressed, and responsibility for the content of this publication, lie solely with the authors. The European Commission is not liable for any use that may be made of the information contained herein. This work may rely on data from sources external to the MARINET project Consortium. Members of the Consortium do not accept liability for loss or damage suffered by any third party as a result of errors or inaccuracies in such data. The information in this document is provided as is and no guarantee or warranty is given that the information is fit for any particular purpose. The user thereof uses the information at its sole risk and neither the European Commission nor any member of the MARINET Consortium is liable for any use that may be made of the information. Page 4 of 17

5 EXECUTIVE SUMMARY The WRAM is a single-bodied heaving buoy point absorber in which mass, spring and damping may be controlled. The numerical solution of the hydrodynamic model had been validated by tank tests at 1/5 th scale when a proposal made under the MARINET 6 call was accepted. A model at nominally 1/2 th scale was built and instrumented. The pneumatic PTO damping was replicated using a series orifice plates, characterised using a purpose-built rig. The test programme was discussed and agreed with ECN and with DNV-GL who were invited to provide external validation. Arrangements for data-logging and post-processing were prepared in advance. Ballasting, trimming, and instrumentation checks were completed in Ireland before shipping the model to Nantes for five days access to the ECN wave tank. Setting up and wave-maker and instrument calibration (6 tests) occupied most of Day 1, followed by decay tests in still water and then a comprehensive series of 89 tests in monochromatic waves and long and short-crested polychromatic (Bretschneider) spectra. These yielded sufficient data for a comprehensive assessment of performance and of stability across a wide range of wave periods and heights. The tests confirmed that the WRAM has the potential to perform powerfully in an energetic North Atlantic site. The device is stable and any risk of parametric roll may be avoided. There is clear scope for improvement, notably by optimising the choice of eigen values for heave, and by reducing viscous drag. This short series of tests has contributed significantly to the development of the WRAM. Page 5 of 17

6 CONTENTS 1 INTRODUCTION & BACKGROUND INTRODUCTION DEVELOPMENT SO FAR Stage Gate Progress Plan for this access OUTLINE OF WORK CARRIED OUT PREPARATORY WORK Characterising PTO damping True-scale model Still water tests Instrumentation Calibration of the Freescale differential pressure sensors Calibration of the OSS Wavestaff internal water level sensors Calibration of the ECN wave probes Moorings Wave calibration tests Reflection tests TESTS Decay tests Pan-chromatic tests Mono-chromatic tests RESULTS Pan-chromatic power RAO ANALYSIS & CONCLUSIONS MAIN LEARNING OUTCOMES PROGRESS MADE Progress made: for this technology Progress Made: For Marine Renewable Energy Industry KEY LESSONS LEARNED REFERENCE Page 6 of 17

7 1 INTRODUCTION & BACKGROUND 1.1 INTRODUCTION The WRAM is a novel heaving buoy point absorber, developed with two essential objectives in mind, long-term survival on an offshore site and a significantly reduced LCOE. Survival is best assured by it being a single axisymmetric body without any articulations, hinges or end-stops, and secured on compliant moorings. A low LCOE is helped by the structural materials being a small fraction of the overall participating mass, a simple power train, and high power, the latter assured by an ability to control mass, spring and damping. The WRAM is in the form of a spar buoy, with a draught much greater than its cross-section. The tank testing of models much larger than 1/5 th scale requires access to correspondingly deep wave tanks, not available in Ireland. This short series of tests at a nominal 1/2 th scale in the ECN basin in Nantes, supported by MARINET, has been timely and a logical next step in the necessary R&D. 1.2 DEVELOPMENT SO FAR The hydrodynamic analysis, numerical solutions and boundary element computations were completed in late 214 and validated by proof-of-concept tests at 1/5 th scale in early 215. This made it possible to estimate probable average electrical power from a North Atlantic site. A tentative general arrangement for a utility-scale device was modelled in SolidWorks, from which a bill of materials was generated, leading to approximate capital costs. A first pass at probable LCOE was made from these estimates of performance and costs, augmented by previous detailed knowledge of heaving buoy point absorbers. It is now essential to establish an accurate power matrix for the WRAM based on standardised performance testing. This document is a summary record of the first phase in that process Stage Gate Progress Previously completed: Planned for this project: STAGE GATE CRITERIA Stage 1 Concept Validation Linear monochromatic waves to validate or calibrate numerical models of the system (25 1 waves) Finite monochromatic waves to include higher order effects (25 1 waves) Hull(s) sea worthiness in real seas (scaled duration at 3 hours) Restricted degrees of freedom (DofF) if required by the early mathematical models Provide the empirical hydrodynamic co-efficient associated with the device (for mathematical modelling tuning) Investigate physical process governing device response. May not be well defined theoretically or numerically solvable Real seaway productivity (scaled duration at 2-3 minutes) Initially 2-D (flume) test programme Short crested seas need only be run at this early stage if the devices anticipated performance would be significantly affected by them Evidence of the device seaworthiness Initial indication of the full system load regimes Stage 2 Design Validation Accurately simulated PTO characteristics Performance in real seaways (long and short crested) Status Page 7 of 17

8 STAGE GATE CRITERIA Survival loading and extreme motion behaviour. Active damping control (may be deferred to Stage 3) Device design changes and modifications Mooring arrangements and effects on motion Data for proposed PTO design and bench testing (Stage 3) Engineering Design (Prototype), feasibility and costing Site Review for Stage 3 and Stage 4 deployments Over topping rates Status Stage 3 Sub-Systems Validation To investigate physical properties not well scaled & validate performance figures To employ a realistic/actual PTO and generating system & develop control strategies To qualify environmental factors (i.e. the device on the environment and vice versa) e.g. marine growth, corrosion, windage and current drag To validate electrical supply quality and power electronic requirements. To quantify survival conditions, mooring behaviour and hull seaworthiness Manufacturing, deployment, recovery and O&M (component reliability) Project planning and management, including licensing, certification, insurance etc. Stage 4 Solo Device Validation Hull seaworthiness and survival strategies Mooring and cable connection issues, including failure modes PTO performance and reliability Component and assembly longevity Electricity supply quality (absorbed/pneumatic power-converted/electrical power) Application in local wave climate conditions Project management, manufacturing, deployment, recovery, etc Service, maintenance and operational experience [O&M] Accepted EIA Stage 5 Multi-Device Demonstration Economic Feasibility/Profitability Multiple units performance Device array interactions Power supply interaction & quality Environmental impact issues Full technical and economic due diligence Compliance of all operations with existing legal requirements Plan for this access These were the first standardised performance tests of the WRAM at this scale. The intention was to assess performance, following best practice guidelines and experimental procedures. The tests were expected to help define and to quantify design features and uncertainties that would benefit from further R&D and modification. Specifically, we wished to assess the accuracy of PTO simulation using characterised orifice plates, and to verify that any risk of parametric roll, a classic problem with deep draught spar buoys, could be readily avoided. Page 8 of 17

9 2 OUTLINE OF WORK CARRIED OUT 2.1 PREPARATORY WORK Characterising PTO damping The WRAM PTO is pneumatic, similar to that for an OWC device. This was replicated in the scale model using a range of interchangeable orifice plates fixed over a circular hole in the top deck, and sealed with a neoprene gasket. Each plate, formed from rigid plastic, has four identical and accurately dimensioned orifices, ranging in size from 8.2mm to 52.4mm diameter. Twenty plates were characterised using a purpose-built rig. Using theoretical relationship (Sheng & et.al., 213) between pressure, flow and damping coefficient for a given orifice plate, the characterisation curve was derived (damping coefficient vs. orifice diameter). Five damping coefficients were chosen for the tests: four of the above plates, plus no plate which provided light damping True-scale model The nominally 1:2 scale model was based on the sub-optimal geometry resulting from the initial hydrodynamic analysis and BEM (WAMIT) assessment, but incorporating lessons learnt from the proof-of-concept 1:5-scale tank tests, notably a lowered ballast and thus a better metacentric height. The drawings and estimations of masses and inertia were prepared in SolidWorks. The main frame was built in aluminium with polypropylene tanks and lead ballast. The as-built dimensions were comfortably within the specified tolerances of +/- 5mm or less Still water tests Preliminary still water tests were carried out at the 14 th Lock of the Grand Canal, with the permission of Inland Waterways, Ireland. These tests included ballasting, checking the trim (which required no adjustment) and measurement of the natural frequencies of heave and pitch for various settings of the inertial mass. This also provided an opportunity to check the integrity of the instrumentation and data logging Instrumentation The model carries two differential pressure sensors and three wave gauges on the deck plate for all tests. The differential pressure sensors are Freescale MPXV72 units. These are pre-amplified, calibrated and temperature compensated devices. One terminal of each sensor is connected, via tubing, to the air chamber; the other terminal is open to atmosphere. This arrangement permits monitoring of the pressure differential across the damping plate of the model under test. The wave gauges are Ocean Sensor Systems Wavestaff units with 5mm probes installed. They are spaced at 12 intervals around the edge of the air chamber and permit monitoring of the behaviour of the free surface within the chamber. The data collected may be used to identify any sloshing behaviour or clipping behaviour, as well as to confirm the data recorded by the pressure sensors. All instrumentation on the model is cabled 8 metres to a signal conditioning box before being converted from -5V and -1V output respectively, to ±1V output. This data conversion is carried out using Radionics Linecard Analogto-Analog converters. From there the data is transmitted the remaining 3 metres to the onsite data acquisition hardware over BNC terminated, shielded, twisted pair cabling. Page 9 of 17

10 volts Pressure (Pa) volts For certain tests the model also carried an X-IMU Inertial Measurement Unit. This device permits recording of the accelerative and gyroscopic forces at the deck plate of the model and has local logging capability which is correlated with the ECN test data by way of log file timestamps Calibration of the Freescale differential pressure sensors The pressure sensors were calibrated using ECN s high-spec pressure calibration equipment, where the known pressures (from the digital display of the pressure control device) were tabulated by hand and the sensor signals recorded on the analog DAQ system. The calibration data were fitted with a linear equation. The pressure sensors are internally temperature compensated between +1 to +6ᶛC and so no extra compensation was deemed necessary ve pressure w.r.t. atmosphere Sensor 1 Sensor ve pressure w.r.t. atmosphere Sensor 1 Sensor time (s) Figure 2-1: Pressure sensor measurements Pressure Sensor Calibration Curves Curve Fitting p1 = * v p2 = * v volts Figure 2-2: Pressure sensor calibration curve Sensor 1 Sensor Calibration of the OSS Wavestaff internal water level sensors The three internal air-chamber water level sensors were calibrated in the ECN tank. The crane was used to lift the model a series of short distances vertically. In still water the internal water level equals the amount of the model s heave displacement. The Qualisys motion detection system was used measure the heave displacement (in fractions of a metre) while the sensor data were logged on the analog DAQ system in volts, and calibration curves produced Calibration of the ECN wave probes ECN calibrated the wave probes out of the tank using a deep water container. The measured data were scaled according to the gain supplied by ECN. On their advice the signal offset bias was removed based on the measured values during the zero period of data. An initial zero-ing period of 5 sec was logged in the analog files for every test before the main data set Moorings The ECN tank is 5m deep, with a central pit 1m deep. A 4-point mooring system was adopted to hold the model over the centre of the tank. Each mooring comprised a 2.5m line from a drop-off point on the float to a 3 litre surface buoy and from each of these an 8m sinking line led to an anchor point at the corner of a 5.5 x 5.5 m square. Two of the anchor points were on a rail fixed to the tank floor, two to the base of massive concrete blocks. The measurement of mooring loads was not part of these tests. Page 1 of 17

11 Amplitude Amplitude Amplitude Wave calibration tests Before deploying the model in the wave tank, a short series of wave calibration tests were completed. A series of wave gauges were set in line of the direction of the waves, five at the position where the model was to be deployed and three to the side of the model. Calibration was based on a Bretschneider spectrum for a single significant wave height and six different peak periods H s (m) T (s) Reflection tests After the main testing period, a short series of beach reflection tests were completed. The five wave probes at the centre were used to measure various monochromatic wave sets. The data was analysed using code provided by ECN which uses the Mansard and Funke method for wave reflection calculation. The plots in Figure 2-3 show to the analysis for sea state H=.15m, T=2.57s where different combinations of probes have been used. It can be seen that the reflected waves are very small with respect to the incident waves. The wave probe positions are shown below: 1.5 Reflection Analysis Incident Waves: Probes Reflected Waves: Probes Incident Waves: Probes Reflected Waves: Probes Incident Waves: Probes Reflected Waves: Probes Wave probe number WG4 (closest to wave maker paddles) Distance w.r.t. front probe WG5.235 WG6.335 WG7.51 WG8.56 Table 2.1: Wave probes for reflection analysis Frequency (rad/s) Figure 2-3 Reflection test results Page 11 of 17

12 2.2 TESTS Decay tests ECN s Qualisys system was used to obtain both heave and pitch Eigen periods of the model, each for the two tank configurations of fully open and fully closed. Table 2.2 shows these Eigen periods identified from the frequency domain data, confirmed by the time domain data. Heave Pitch Fully Open Tank Fully Closed Tank Table 2.2: Identified Eigen periods (model scale) Figure 2-4: Heave, Fully Open Tank Figure 2-5: Pitch, Fully Open Tank Figure 2-6: Heave, Fully Closed Tank Figure 2-7: Pitch, Fully Closed Tank Page 12 of 17

13 2.2.2 Pan-chromatic tests Long crested Long crested pan-chromatic sea states were tested as outlined in Table 2.3 and Table 2.4. To maximise the range of damping values discrete damping values were selected for each row (wave-height) in each test scatter. Hs (m).4 D1 D1 D1 D1 D1 D1 D1.28 D2 D2 D2 D2 D2 D2 D2.18 D3 D3 D3 D3 D Tp (sec) Table 2.3: Long crested pan-chromatic tests - Fully closed tank Hs (m).4.28 D1 D1 D1 D1 D1 D1 D1.18 D2 D2 D2 D2 D2 D2 D2.1 D3 D3 D3 D3 D3 D3 D Tp (sec) Table 2.4: Long crested pan-chromatic tests 2/3 closed tank Short crested Short crested pan-chromatic sea states were tested as outlined in Table 2.5. Here, two sea states were used for all tank configurations. Hs (m).28 D2 D2.28 D2 D2.28 D2 D Tp (sec) Table 2.5: Short crested pan-chromatic tests Fully open, Fully closed and 2/3 closed tank Mono-chromatic tests Table 2.6 shows the mono-chromatic sea states tested. The purpose here was to obtain RAO information. H (m) D2 D2 D2 D2 D2 D2 D2.5 D3 D3 D3 D3 D3 D3 D Tp (sec) Table 2.6: Mono-chromatic tests - Fully closed tank Page 13 of 17

14 Power (normalised) Parametric roll Demonstration of parametric roll was achieved with a mono-chromatic wave at half the heave Eigen period of the model, also very close to the pitch Eigen period. Table 2.7 shows the heave, pitch and roll response periods, obtained from the frequency domain of measured data. Note the halving of the roll period between tank change (fully open to fully closed), indicating the removal of parametric roll. Waveheight (m) Wave Period (s) Heave Motion Period (s) Pitch Motion Period (s) Roll Motion Period (s) RESULTS Pan-chromatic power Table 2.7: Mono-chromatic tests - Fully closed tank Tank Damping Notes Fully Open Fully Closed 'D' 'D' Parametric Roll No Parametric Roll Mean Pan-Power, Full Tank, ECN Model Scale Hs(m) Tp(s) Figure 2-8 Normalised mean power for fully closed tanks Page 14 of 17

15 Power (normalised) Power (normalised) Mean Pan-Power, 2/3 Tank, ECN Model Scale Hs(m) Tp(s) Figure 2-9 Normalised mean power for 2/3 closed tanks Mean Pan-Power, All Tank Configs, ECN Model Scale, Hs:.28m Tank Config Tp(s) Figure 2-1 Normalised mean power for all tank configurations in spreaded sea states Page 15 of 17

16 RAO RAO RAO, Full Tank 4 3 X: 3 Y:.1 Z: X: 3 Y:.5 Z: X: 2.1 Y:.1 Z: H(m).6.4 X: 2.1 Y:.5 Z: T(s) Table 2.8: 3D RAO plot 2.4 ANALYSIS & CONCLUSIONS This was an appropriate scale and ECN was an excellent facility for the required tests. instrumentation and the team delivered all that could be expected in the short time available. The model, its The tests confirmed that the WRAM has the potential to be a powerful point absorber, and one that is capable of adapting to a wide range of wave conditions, confirming that average power of several hundred kilowatts from a North Atlantic site is achievable. The tests showed that there is significant scope for improvements that can readily be incorporated in a model of this scale. It is therefore intended to complete a further and more extended series of standardised tests at this or a similar scale. 3 MAIN LEARNING OUTCOMES 3.1 PROGRESS MADE This short series of tests (less than 4 days actual tank testing) has resulted in satisfactory progress in the further development of the WRAM. The following were achieved: Performance, - power RAO s recorded and adequate data for a power matrix Tuning: RAO s for different resonant heave frequencies, as anticipated Stability: the device was more stable than expected Parametric roll could always be avoided by altering settings of the reference mass and, to a lesser but important extent, by adjusting damping Uncertainties and modifications: o Scale: Although nominally 1:2, the results suggest that 1:3 may be more correct when considering a North Atlantic site as the Irish AWETS. This is most important when assessing LCOE. o Viscous drag: there is scope for better performance, a modified model will be more streamlined o Optimisation: there is scope for improved performance (ie over a full scatter diagram) by adjusting the size of the reference mass Page 16 of 17

17 o o PTO simulation: there is some uncertainty about the damping values attributed to the orifice plates; this will be addressed Instrumentation: ideally calibration should be carried out before and after tests, if time permits. Some minor and easily corrected faults were observed Progress made: for this technology Point absorbers Oscillating point absorbers tend to be regarded by many one of the less powerful classes of WECs. It is true that an axi-symmetric heaving buoy has, in theory, a maximum capture width of, a half or even a third of some alternative oscillating systems. However when the criteria of paramount importance, ie survival and reduced LCOE, are considered, a simple axi-symmetric heaving buoy has distinct advantages. This becomes more convincing if the device is fully tuneable, and well suited for autonomous control and deployment in utility-scale arrays. These are the priorities that drive the WRAM R&D Next Steps The intention is to return to the same or a similar facility for an extended series of standardised performance tests, plus and the measurement of loads and forces on the main elements of the device and its moorings. Before doing so, there is a need to modify and, as far as is reasonable at this scale, to optimise the basic dimensions, including the settings for the reference mass. The existing and now validated numerical model will be used to explore alternatives and (with coefficient from BEM computations) to predict probable performance metrics. A finite element analysis will be part of this preparatory work. CFD will be used to assess the extent of improved streamlining relative to the findings of the heave decay tests. The existing model will be re-built and the instrumentation extended to include the measurement of loadings Progress Made: For Marine Renewable Energy Industry This short series of tests is of minimal significance to the OE sector as a whole. It is hoped that this will change as the WRAM R&D progresses through to the higher TRL s. 3.2 KEY LESSONS LEARNED Working with an independent expert consultancy is highly beneficial, in this case DNV GL The MARINET project has been an excellent support mechanism for this and for many other developers A rigorous, and now standardised, approach to tank testing is important As much as possible by way of commissioning and still-water trials should be completed beforehand One week is rather too short for tests at this scale, a significant fraction necessarily being taken up by on-site calibration; it would be useful to have had the option of additional time Simulating a pneumatic PTO using characterised orifice plates has inherent uncertainties; access to a depressurised facility such as at MARIN has to be a better option. REFERENCE Sheng, W., & et.al. (213). Investigation to Air Compressibility of Oscillating Water Column Wave Energy Converters. ASME nd International Conference on Ocean, Offshore and Arctic Engineering. Page 17 of 17