Experimental Results from the Lysekil Wave Power Research Site

Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology 957

Experimental results from the Lysekil Wave Power Research Site

OLLE SVENSSON

ACTA UNIVERSITATIS UPSALIENSIS ISSN 1651-6214 ISBN 978-91-554-8433-0 UPPSALA urn:nbn:se:uu:diva-179098 2012 Dissertation presented at Uppsala University to be publicly examined in Polhem Å 10134, Ångströmlaboratoriet, Lägerhyddsvägen 1, Uppsala, Friday, September 28, 2012 at 13:00 for the degree of Doctor of Philosophy. The examination will be conducted in English.

Abstract Svensson, O. 2012. Experimental results from the Lysekil Wave Power Research Site. Acta Universitatis Upsaliensis. Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology 957. 101 pp. Uppsala. ISBN 978-91-554-8433-0.

This thesis presents how experimental results, from wave power research performed offshore at the Lysekil research site, were obtained. The data were used to verify theoretical models as well as evaluate the feasibility of wave power as a future sustainable energy source. The first experiments carried out at the research site was the measurement of the force in a line where one end was connected to a buoy with a diameter of 3 m and the other end to a set of springs with limited stroke length. The system is exposed to high peak forces compared to average forces. The maximum measured force in the line, when the buoy motion is limited by a stiff stopper rope is ten times the average force in that particular sea state. The experiment performed on the first wave energy converter tested at the Lysekil Research Site is described. The infrastructure of the site is presented where the central connection point is the measuring station. The key finding is that it is possible to transform the motions of ocean waves into electrical energy and distribute it to land. Many wave energy converters must be interconnected if large amounts of energy are to be harvested from the waves. The first submerged substation intended for aggregation of energy from wave power converters is described, with focus on the measurement and control system placed inside the substation. During this experiment period the generators were equipped with many different sensors; these measurements are explained in the thesis. The system that aggregates power from the studied wave energy converter is regularly exposed to peak power of up to 20 times the maximum average output from the converter. Vertical and horizontal movement of the buoy has been measured in different ways. The result is that the vertical displacement of the buoy can be measured with a simple accelerometer circuit but it is much more complicated to measure the horizontal displacement. A special method for measuring the horizontal displacement has been implemented by measuring the strain in the enclosure and the force in the line.

Keywords: Wave power, Lysekil, Marine Substation, Offshore measurement, strain gauge, lateral force, Invlination and azimuth angles, Wave energy converter, Temperature measurements, Inverter, Energy, Control sustem, CompactRIO, Vågkraft, Mätteknik, Styrsystem, Lysekil

Olle Svensson, Uppsala University, Department of Engineering Sciences, Electricity, Box 534, SE-751 21 Uppsala, Sweden.

ISSN 1651-6214 ISBN 978-91-554-8433-0 urn:nbn:se:uu:diva-179098 (http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-179098)

To my late father who started the journey ”from cultivation to education”

List of Papers

This thesis is based on the following papers, which are referred to in the text by their Roman numerals.

Paper published 2005 I Gustafsson, S., Svensson, O., Sundberg, J., Bernhoff, H., Leijon, M., Danielsson, O., Ericsson, M., Thorburn, K., Strand, K., Henfridsson, U., Ericsson, E., Bergman, K., “Experiments at Islandsberg on the west coast of Sweden in preparation of the construction of a pilot wave power plant” Proceedings of the 6th EWTEC conference, Glasgow, UK, 2005.

Papers published 2007 II Waters, R., Stålberg, M., Danielsson, O., Svensson, O., Gustafsson, S., Strömstedt, E., Eriksson, M., Sundberg, J., and Leijon, M. “Experimen- tal results from sea trials of an offshore wave energy system”. Applied Physics Letters, 90:034105, 2007. doi: http://dx.doi.org/10.1063/1.2432168

III Eriksson, M., Waters, R., Svensson, O., Isberg, J., and Leijon, M. “Wave power absorption: Experiments in open sea and simulation”. Journal of Applied Physics, 102:084910, 2007. doi: http://dx.doi.org/10.1063/1.2801002

Papers published 2008 IV Leijon, M., Boström, C., Danielsson, O., Gustafsson, S., Haikonen, K., Langhamer, O., Strömstedt, E., Stålberg, M., Sundberg, J., Svensson, O., Tyrberg, S., and Waters, R. “Wave Energy from the North Sea: Experiences from the Lysekil Research Site”. Surveys in Geophys- ics, Springer, 2008. Doi: 10.1007/s10712-008-9047-x

Papers published 2009 V Leijon, M., Waters, R., Rahm, M., Svensson, O., Boström, C., Strömstedt, E., Engström, J., Tyrberg, S., Savin, A., Gravrakmo, H., Bernhoff, H., Sundberg, J., Isberg, J., Agree, O., Danielsson, O., Eriks- son, M., Lejerskog, E., Bolund, B., Gustafsson, S., Thorburn, K.,, “Catch the wave to electricity”, IEEE Power and Energy Magazine, Volume 7, Issue 1, January-February 2009 Page(s):50 - 54, doi: 10.1109/MPE.2008.930658

VI Svensson, O., Boström, C., Rahm, M., & Leijon, M., “Description of the control and measurement system used in the Low Voltage Marine Sub- station at the Lysekil research site.” Proceedings of the 8th European wave and tidal energy conference, EWTEC09, Uppsala, Sweden, pp. 44– 50.

VII Boström, C., Waters, R., Lejerskog, E., Svensson, O., Stålberg, M., Strömstedt, E., Leijon, M. “Study of a Wave Energy Converter Con- nected to a Nonlinear Load” IEEE Journal of Oceanic Engineering, Vol-

ume. 34, issue 2, pp. 123-127, 2009., DOI: 10.1109/JOE.2009.2015021

VIII Rahm, M., Boström, C., Svensson, O., Grabbe, M., Bülow, F. and Lei- jon, M. “Laboratory experimental verification of a marine substation” Proceedings of the 8th European wave and tidal energy conference, EWTEC09, Uppsala, Sweden, pp. 51–58.

IX Boström C., Svensson O., Rahm M., Lejerskog E., Savin A., Strömstedt E., Engström J., Gravråkmo H., Haikonen K., Waters R., Björklöf D., Johansson T., Sundberg J. and Leijon M., “Design proposal of electrical system for linear generator wave power plants”, Proc. of IECON 2009, 35th annual conference of the IEEE Industrial Electronics and Society, Porto, Portugal 3-5 November 2009, no. PD-027448, pp. 4429-4434. doi: 10.1109/IECON.2009.5414903

X Savin A., Svensson O., Strömstedt E., Boström C. and Leijon M., “De- termining the service life of a steel wire under a working load in the Wave Energy Converter (WEC) ”, Proceedings of the 28th International Conference on Ocean, Offshore and Arctic Engineering (OMAE 2009), May 31 to June 5, 2009, Honolulu, Hawaii. doi:

http://dx.doi.org/10.1115/OMAE2009-79164

Papers published 2010 XI Rahm M., Boström C., Svensson O., Grabbe M., Bülow F. and Leijon M., “Offshore underwater substation for wave energy converter arrays”, Renewable Power Generation, IET, Volume.4, no.6, pp.602-612, No- vember 2010 doi: 10.1049/iet-rpg.2009.0180

XII Boström, C., Lejerskog, E., Tyrberg, S., Svensson, O., Waters, R., Savin, A., Bolund, B., Eriksson, M., Leijon, M., “Experimental results from an offshore wave energy converter”, Journal of Offshore Mechan- ics and Arctic Engineering, Volume. 132, Issue 4, 041103, 5 pages, 2010. doi:10.1115/1.4001443

XIII Gravråkmo, H., Leijon, M., Strömstedt, E., Engström, J., Tyrberg, S., Savin, A., Svensson, O., Waters, R., “Description of a Torus-shaped Buoy for Wave Energy Point Absorber”, Renewable energy 2010, 27June-2July, 2010, pacifico Yokohama, Yokohama, Japan.

Papers published 2011 XIV Tyrberg, S., Svensson, O., Kurupath, V., Engström, J., Strömstedt, E., and Leijon, M. “Wave Buoy and Translator Motions — On-site Meas- urements and Simulations”. IEEE Journal of Oceanic Engineering, Vol- ume. 36, Issue 3, pp. 377–385, July 2011.

doi:10.1109/JOE.2011.2136970

XV Waters, R., Rahm, M., Eriksson, M., Svensson, O., Strömstedt, E., Boström, C., Sundberg, J., Leijon, M. “Ocean wave energy absorption in response to wave period and amplitude - offshore experiments on a wave energy converter”, Renewable Power Generation, IET Volume: 5, Issue: 6 pp. 465-469, November 2011. doi: 10.1049/iet-rpg.2010.0124 ,

XVI Lejerskog, E., Gravråkmo, H., Savin, A., Strömstedt, E., Tyrberg, S., Haikonen, K., Krishna R., Boström, C., Rahm, M., Ekström, R., Svensson, O., Engström, J., Ekergård, B., Baudoin, A., Kurupath, V., Hai, L., Li, W., Sundberg, J., Waters R., and M. Leijon. “Lysekil Re- search Site, Sweden: A Status Update” Proceedings of the Ninth Euro-

pean Wave and Tidal Energy. Southampton .UK 2011.

XVII Svensson, O., Strömstedt, E., Savin, A. Leijon, M. “Sensors and Meas- urements Inside the second and Third Wave Energy Converter at the Lysekil Research Site” Proceedings of the Ninth European Wave and Tidal Energy. Southampton .UK 2011.

Papers published 2012 XVIII Boström, C., Rahm, M., Svensson, O., Strömstedt, E., Savin, A., Wa- ters, R., and Leijon, M., “Temperature Measurements in a Linear Gen- erator and Marine Substation for Wave Power”, Journal of Offshore Me- chanics and Arctic Engineering, Volume 134, Issue 2, 21901, 6 pages,

May 2012. doi:10.1115/1.4004629

XIX Strömstedt, E., Svensson, O., Leijon, M. “A Set-Up of 7 Laser Triangu- lation Sensors and a Draw-Wire Sensor for Measuring Relative Dis- placement of a Piston Rod Mechanical Lead-Through Transmission in an Offshore Wave Energy Converter on the Ocean Floor ” ISRN Renew- able Energy, Volume. 2012, Article ID 746865, 32 pages, 2012.

doi:10.5402/2012/746865

XX Savin, A., Svensson, O and Leijon, M. “Estimation of Stress in the Inner Framework Structure of a Single Heaving Buoy Wave Energy Con- verter”, IEEE journal of Oceanic Engineering, Volume 37 number 2 p.p.

309-307 , Issue 99, /ISSN 0369-9059) doi: 10.1109/JOE.2012.2188614

XXI Savin, A., Svensson, O., Leijon, M. “Azimuth-inclination angles and snatch load on a tight mooring system”, Ocean Engineering, Volume 40, pp 40-49, Feb 2012, doi:

http://dx.doi.org/10.1016/j.oceaneng.2011.12.007

XXII Rahm, M., Svensson, O., Boström, C., Waters, R., Leijon, M., “Experi- mental results from the operation of aggregated wave energy converters” Renewable Power Generation, IET , Volume.6, no.3, pp.149-160, May 2012 doi: 10.1049/iet-rpg.2010.0234

Submitted Papers XXIII Svensson, O and Leijon, M. “Peak Force Measurements on a Cylindrical Buoy with Limited Elastic Mooring” Submitted to IEEE Oceanic Engi-

neering, accepted with major revision 2011-11-02.

XXIV Svensson, O., Boström, C and Leijon, M. “A Study of the Possible Power Extraction from a Point Absorbing Wave Energy Converter” Submitted to Energies, accepted with major revisions 2012-05-12.

Supplemental material XXV National instrument Graphical design achievement 2010.

Conference papers not included in the thesis, that later have been accepted as journal papers and patent

A. Tyrberg, S., Stålberg, M., Boström, C., Waters, R., Svensson, O., Strömstedt, E., Savin, A., Engström, J., Langhamer, O., Gravråkmo, H., Haikonen, K., Tedelid, J., Sundberg, J., and Leijon, M. “The Lysekil Wave Power Project: Status Update”. Proceedings of the 10th

World Renewable Energy Conference, WREC, Glasgow, UK, 2008.

B. Boström, C., Lejerskog, E., Tyrberg, S., Svensson, O., Waters, R., Savin, A., Bolund, B., Eriksson, M., and Leijon, M. “Experimental results from an offshore wave energy converter”. Proceedings of the 27th International Conference on Offshore Mechanics and Arctic

Engineering, OMAE, Estoril, Portugal, 2008.

C. Cecilia Boström, M. Rahm, O. Svensson, E. Strömstedt, A. Savin, R. Waters and M. Leijon, “Temperature measurements in a linear generator and marine substation for wave power”, Proceedings of the ASME 2010 29th International Conference on Ocean, Offshore and Arctic Engineering (OMAE 2010), June 6-11, 2010, Shanghai, China,

OMAE2010-20881.

D. A wave power unit International patent, publication number WO 2010/085188, published 2010-07-29.

Contents

1. Introduction ...... 17 1.1. Peak and average power ...... 17 1.2. The early history of water wave research ...... 17 1.3. Point absorber ...... 18 1.4. Linear generator...... 18 1.5. My work ...... 19 1.5.1. Classification of articles...... 19 2. The Uppsala University wave energy concept ...... 22 2.1. The wave energy converter...... 22 2.2. The substation...... 22 2.3. Previous thesis from the wave power group...... 22 3. The Lysekil project...... 26 3.1. Project purpose, challenges and project stages...... 26 3.1.1. Papers describing project purpose ...... 26 3.1.2. Purpose ...... 26 3.1.3. Challenges...... 26 3.1.4. Theoretical verification stage...... 26 3.1.5. Experimental verification stages...... 27 3.2. The research area ...... 28 4. Theory ...... 30 4.1. Wave theory...... 30 4.1.1. Water particle movement...... 30 4.1.2. Wave group theory...... 31 4.2. Generator theory ...... 32 4.3. Damping factor ...... 33 4.4. WEC theory ...... 33 4.4.1. Model of a point absorber for heave motion...... 33 5. Wave climate measurements at the research site...... 37 5.1. Articles with main focus on the wave climate measurement at the research area...... 37 5.2. Measurement set up ...... 37

6. Measurements with the measuring equipment placed inside the buoy...... 38 6.1. Background to buoy measurements...... 38 6.2. General measurement set-up...... 38 6.3. Measurement system ...... 40 6.3.1. The force sensor...... 40 6.3.2. Accuracy of force measurements...... 40 6.3.3. Electronics ...... 41 6.3.4. The data logger ...... 43 6.4. Calibration ...... 44 6.5. The launch of the first force measurement at the research site...... 44 6.6. Developments and changes of the measurement system in following experiments ...... 45 6.6.1. Buoy measurements with the first WEC L1 ...... 45 6.6.2. The torus-shaped buoy...... 46 6.6.3. Cylindrical buoys 2 and 3 ...... 48 7. The measuring station ...... 49 7.1. Introduction...... 49 7.2. The first substantial data...... 52 7.3. Measuring Station Control System...... 52 7.3.1. The first compactRIO system ...... 52 7.4. Non-linear load ...... 54 7.5. The measuring station during the park experiments 2009...... 55 8. The Marine Substation ...... 56 8.1. The enclosure...... 56 8.2. The electrical system ...... 56 8.2.1. The electrical main circuit ...... 56 8.2.2. Auxiliary Power Supply System...... 57 9. The control system of the marine substation...... 58 9.1. The safety and relay control system...... 59 9.2. The control of the inverter ...... 60 9.3. Dedicated data acquisition system...... 60 10. Sensors and measurements during the park test ...... 61 10.1. Position of the translator...... 61 10.1.1. Measurement set-up...... 61 10.1.2. Calibration ...... 62 10.2. Strain measurements...... 63 10.3. Search-Coil sensors ...... 64 10.4. Temperature measurements ...... 65 10.5. Water level and leakage detection ...... 65 10.6. Piston and seal housing displacement...... 65

11. Buoy and translator movements ...... 66 11.1. Time and “unnecessary measurements” ...... 66 11.2. Accelerometer measurements and influence of gravity and rotation...... 67 11.3. Measurement of the buoy position with strain gauges placed on the capsule ...... 68 11.4. Optical measurements...... 68 11.5. Comparison of the different measurements...... 68 12. Comments on presented power data from one single generator...... 70 12.1. Power absorption with resistive linear load...... 70 12.2. Power absorption with a unlinear load ...... 71 12.3. Generator power output ...... 72 12.4. Power calculations from L2, L3 ...... 74 13. Power smoothening effects from two or more WECs...... 75 13.1. Discussions about power smoothening effects and the outline of the collecting system ...... 75 14. Results from force measurements ...... 76 14.1. Peak and average power ...... 78 15. Summary of Papers ...... 80 16. Conclusions ...... 90 16.1. The control system...... 90 16.2. Measurements inside the buoy...... 91 16.2.1. Force measurements ...... 91 16.2.2. Measurements of buoy movement...... 91 16.3. Relation between peak energy and average energy ...... 92 17. Svensk sammanfattning...... 93 17.1. Varierande energikälla...... 93 17.2. Mitt arbete...... 93 17.3. Mätningar med sensorer och datalagring inuti bojen...... 94 17.4. Mätstugan ...... 94 17.5. Ställverket och mätningarna i generatorerna ...... 94 17.6. Resultat ...... 95 17.6.1. Bojens rörelse ...... 95 17.6.2. Kraften i linan ...... 96 17.6.3. Medelenergin i förhållande till den maximala energin från vågkraftverket...... 96 18. Acknowledgements ...... 97 19. Bibliography...... 99

Glossary

Buoy The float on the ocean surface. Linear Generator The generator that is a part of the WEC. Translator Is for the linear generator, the same as the rotor in a rotating generator. The Lysekil Research The area outside Lysekil close to Härmanö where the Site experiments have been performed. Wave energy Converter The linear generator in its enclosure connected to buoy. Line The connection between the generator and the buoy. Rope A line made of plastic fiber. Wire A line made of metal fiber. Substation The place where energy from multiple WEC are aggregated and adapted for transmission. The measuring station A cabin located on the island Härmanö. Wave crest The top of the wave Wave trough The bottom of the wave Piston rod, piston The steel cylinder connecting the translator and the line. Seal Housing The seal between the piston rod and the generator housing.

Abbreviations

WEC Wave Energy Converter LRS Lysekil Research Site L1, L2, … Names of the Lysekil project WECs, Lysekil 1, Lysekil 2 AC Alternating Current DC Direct Current RG Generator inner resistance LS Synchronous inductance Lg Generator inductance Rk Sea cable resistance Lk Sea cable inductance PAC Programmable Automation Controller SHDSL Symmetric High-Speed Digital Subscriber Line

Nomenclature

General Symbol Unit Quantity B T Magnetic flux density D / mC 2 Electric displacement field E V/m Electric field F Newton Force f Hz Frequency g 2 Gravity constant / sm H A/m Magnetizing field M Significant Wave Height H s P W Power S Energy period Te λ M Wave length ω Rad/s Angular frequency ρ 3 Charge density c / mC

Chapter 4.3 – Theory

Symbol Unit Quantity x t)( M Translator displacement from equilibrium y t)( M Buoy displacement from equilibrium ∞ kg Added mass in the infinite frequency limit mα kg Buoy mass mb kg Translator mass (Piston mass in III) mp N/m Impulse response functions for the excitation force fe t)( N Excitation force Fe t)( h t)( kg/s Radiation impedance ∗ Convolution product η t)( m Incident wave elevation in the center of the buoy for undisturbed surface as a function of time.

N Hydrostatic reaction force Fh ρ / mkg 3 Density of water 2 Water plane area of the buoy Sw m Line force Fw N Line spring constant kw N/m Energy storage spring force Fs N Static spring force in equilibrium Fstatic N Static spring force ks N Upper end stop spring force Fu N Lower end stop spring force Fl N γ Nm/s Damping constant

1. Introduction

This thesis presents how experimental results from wave power research performed offshore at the Lysekil research site were obtained. The data was used to verify theoretical models as well as evaluate the feasibility of wave power as a sustainable energy source for the future. The intention of this comprehensive summary is to put the academic results presented in the papers in context.

1.1. Peak and average power The energy in ocean waves has been known and respected by sailors for centuries. They have managed to control the energy in the wind but only to handle the energy in the waves. Why has no one managed to harvest the energy of the ocean waves and transform it to useful energy throughout the centuries? The problem with extracting energy from the ocean waves is that the average power (energy) is much smaller than the power peaks. Salter com- mented on this in 1974, stating that: “The essential problem is finding a method to convert dispersed, random, alternating forces into concentrated, direct force, using a mechanism which is efficient at low levels and yet ro- bust enough to withstand the worst conditions” [1].

1.2. The early history of water wave research The basic mathematical theory of fluid motion and water waves was described centuries ago by Bernoulli (1700-1782), Gerstner (1756-1832), Fara- day (1791-1867) and Stokes (1819-1903) [2-5]. Lamb (1849-1934) made an excellent compilation of the previous research and wrote a textbook that is still in use today [6]. The energy in water wave, governing equations and overview of the early research was summarized by Evans [7]. Many types of wave energy absorbers have been suggested and patented; there was a spate of them after World War I. But it was not until 1965 when Yoshio Masuda commercialized a navigation buoy powered with an air turbine, that energy in the water waves was converted to useful energy [8]. During the oil crises beginning in 1973 there was an increase in wave power research funding. The

17 research activity rapidly decreased in the eighties when the price of oil fell. A few first generation prototypes were tested at sea during that period [9].

1.3. Point absorber One type of absorber of the energy in water waves is the point absorber. Budal and Falnes described the point absorber in Nature 1975: “A point absorber of ocean-wave power is a system that has its horizontal extent much smaller than one wavelength” [10]. The Lysekil project is based on the concept of a point absorbing buoy floating on the water surface connected with a line to a linear generator placed on the ocean floor.

1.4. Linear generator The linear generator is not a new invention; Michael Faraday was the inven- tor of the electric generator when he in 1831 in a series of experiments first discovered the transformer by experimenting with his induction ring. Later the same year he discovered magneto electro induction and produced a small DC voltage with the Faraday disc [11]. The first commercial generation of electricity took place only in 1881 after Edison’s work in power stations. Wheatstone patented a pioneering early design of linear induction motor1.

1 http://www.kcl.ac.uk/aboutkings/history/famouspeople/charleswheatstone.aspx (retrieved 2012-07-05)

18 1.5. My work My work has been a diversified work, it spans from the initial discussions about the first wave energy converter (WEC) deployed in 2006, to the implementation of the data acquisition and control system in the submerged substation. Even though I have been involved in many different designs, force measurements have been a thread throughout my Ph.D. Not only because it was the first experiment performed at the research site but also because, the peak force and the repetitive medium forces, relative to the produced energy is what wave power system design is all about.

1.5.1. Classification of articles The author has been the first person involved in the measurements at the Lysekil research site from 2005 onward, therefore the author is a co-writer in almost all papers presenting experimental data from the test site. The papers the author has contributed merely with experimental data are included in the thesis because they are the fruit of a large and long-term underlying work at the department that the author has been a part of. Because of this, many papers are enclosed in the thesis. To assist the reader to sift between the articles, this subparagraph is written as an addition to paragraph 15.

The papers in which the author has contributed merely with experimental data The most important papers in this category are Paper II and Paper III. In Paper II the first energy data from the research site is presented. The power is transmitted to shore and measured in the measuring station. It shows that wave energy can be absorbed offshore and transported to the consumer. Paper III investigates the theory for the Wave Energy Converter (WEC) and presents a simulation model. The results from the simulations are compared with measured data from the research site.

The contribution to Paper XV is solely the measurement data

The author accounts for most of the content in the following papers The articles in which the Author alone accounts for most of the content are: Paper VI, Paper XVII, and Paper XXIII, Paper XXIV.

The control and measuring system inside the substation Paper VI describes the control measurement and safety system implemented in the first submerged substation. The implementation of these systems has solely been a work performed by the author.

19 Sensors inside the WECs Paper XVII describes the measurements with different sensors performed in the second and third WEC. The author had the main responsibility for these implementations.

Force measurements and peak to average power ratio Paper XXIII describes the force measurements performed before the first WEC were constructed; the measurements system used were constructed by the author. Paper XXIV investigates the peak and average relation one step further. In this article focus is on peak and average power delivered from the WEC instead of peak and average force in the line connecting the WEC and the buoy.

Articles that have been initiated in collaboration with the article's first author Paper XIV describes and compares different methods to measure the buoy motion. Paper XXII presents a new measurement method for measuring the inclination and azimuth angle between the generator and the floating buoy.

Important contribution and/or description of important work In the following papers, the author has made contributions to the written content of the paper or/and the paper describes important work performed by the author: Paper I, Paper IV, Paper VII, Paper VIII, Paper IX, Paper X, Pa- per XI, , Paper XXI and Paper XXII.

General papers regarding The Lysekil Research Site Paper I, is the first description of the outline and goal of the research site. Paper IV addresses the vase that has been achieved in all parts of the project until 2008.

The electrical system and nonlinear load Dr Boström has had the main responsibility for the electrical system but we have worked closely together with the measuring station and the substation and our main tasks are intertwined. In Paper IX the electrical system of a wave power park is described and examples of the implementation at the university are given. Paper VII and Paper XII describes the experimental set-up for unlinear loading and the results from the WEC loaded with a rectifier connected to a huge capacitor in parallel with difference resistors.

20 Descriptions of the substation and experimental data from the wave power park Paper XI, Paper VIII and Paper XXII, describes the substation. The author has constructed, assembled as well as tested the substation in cooperation with the first and second author of the papers.

Major contributions to the implementation of the specific measurement resulting in the paper Paper X, Paper XX, Paper XIII, XVIII, Paper XIX.

Application of force measurements Paper X describes the axial tension in a steel wire, the paper uses force data collected by the author as its main source of presented data. Paper XIII describes the torus buoy and the special three point force measuring system suggested by the author.

Applications of strain measurements Paper XX, Estimation of Stress in the Inner Framework Structure of a Single Heaving Buoy Wave Energy Converter

Applications of temperature measurements In Paper XVIII the temperatures inside the WEC and the substation are evaluated at different sea states.

Application of the measurements on the piston rod in the WEC Paper XIX describes the measurements on the piston rod in the WEC.

Other papers Comprehensive descriptions of the Lysekil research Site that does not fit in the other categories Paper V, Paper XVI.

21 2. The Uppsala University wave energy concept

2.1. The wave energy converter The Uppsala University wave energy concept is based on a point absorbing buoy, floating on the ocean surface. A line connects the buoy to the linear generator placed on the ocean seabed. Because of the connection between the buoy and the translator, the motion of the buoy on the water surface is reflected by the motion of the translator when the line is stretched. The water particles, at the surface working on the buoy is constantly varying there vertical speed [27]. This has a direct impact on the output from the generator that varies in frequency and magnitude.

2.2. The substation The probability that the power output from one linear generator following the ocean surface should be equal to the power from any other linear generator in a wave power park at a certain moment is very low. This is because waves observed in the ocean are extremely irregular [28]. If the converters were connected to individual resistive loads the voltage would also never be compatible with the other voltages. Therefore the voltage produced by each single generator must be converted to a value that is common for all generators every moment of time. The different powers produced will be managed by changes in the current instead of the voltage. The aggregation of the different power sources are done in the substation.

2.3. Previous thesis from the wave power group Irina Ivanova, Simulation of Linear Permanent Magnet Octagonal Genera- tor for Sea Wave Energy Conversion. In this licentiate thesis the first simulations of the linear generator and the outline of the concept with a linear generator placed on the ocean floor connected to a buoy on the ocean surface is presented. Further on the idea with the collector substation is presented as a method to suppress voltage pulsations. The simulated generator has eight stator sections.

22 Examples of challenges are given, as the: irregularity of the waves, overloads at stormy conditions, slow vertical wave motion, disturbances on ship traffic and fishing, installation, exploitation and maintenance of the submerged converters. [12].

Karin Thorburn, Electric Energy Conversion Systems: Wave Energy and Hydropower. The frequency output of the linear generator depends on the constantly varying speed of the generator hence the frequency of the output voltage also varies. An analytic model for identifying the harmonics in the voltage output of the linear generator is proposed in this Ph.D. Further on wave energy transmission concepts for linear generator arrays is presented within the work of Dr. Thorburn [13].

Oskar Danielsson, Wave Energy Conversion, Linear Synchronous Perma- nent Magnet Generator. In this Ph.D. thesis the design of the magnetic circuit in the first full scale offshore experiment and the experimental generator are described in detail. End effects in a linear generator are discussed [14].

Mikael Eriksson, Modelling and Experimental Verification of Direct Drive Wave Energy Conversion. Buoy-Generator Dynamics. This Ph.D. thesis has the focus on the ocean, buoy and generator interaction. Models and modeling of a wave energy converter in operation are presented. Through out the thesis linear potential wave theory has been used to describe the wave-buoy interaction [15].

Rafael Waters, Energy from Ocean Waves. Full Scale Experimental Verifi- cation of a Wave Energy Converter. In this Ph.D. thesis the first offshore wave energy converter is described and the first results are thoroughly investigated. A method for generator air gap measurements has been developed and wave climate studies are presented [16].

Olivia Langhamer, Wave energy conversion and the marine environment: Colonization patterns and habitat dynamics. The Ph.D. thesis main conclusion is that main biological impact of large scale wave power conversion of the type studied is because of hard substrates added to the area. These substrates have ecological impact on organisms for which these types of substrates are essential. The organisms within the wave power park are not harmed and the biodiversity is not decreased within the park during the studied period [17].

23 Magnus Rahm, Ocean Wave Energy: Underwater Substation System for Wave Energy Converters. The focus of the thesis is a system for operation of direct driven offshore wave energy converters. The system is experimentally tested by the design implementation and offshore testing of and underwater collector substation [18].

Remya Krishna, Multilevel Inverter for Wave Power Conversion. This licentiate thesis describes a Matlab/Simulink model of the generator that is the first step towards a simulation of the whole energy conversion process in a wave power park, inclusive the synchronization to the grid. Multi level inverters, current and voltage control are described [19]

Halvar Gravråkmo, Buoy for linear wave energy converter. In the licentiate thesis linear wave theory and hydrodynamics are explained with focus on a torus shaped buoy [20].

Cecilia Boström, Electrical systems for wave energy conversion. The work presented focus on the first step in the electric energy conversion, converting the voltage out from the generators into DC. The purpose has been to investigate how the generator will operate when it is subjected to different load cases. Guidelines on how future systems can be implemented are presented [21].

Jens Engström, Hydrodynamic Modelling for a Point Absorbing Wave En- ergy Converter. The author of the Ph.D. thesis investigates further the work that the author of [15] established. The work of the thesis focuses mainly on the energy transport of ocean waves and on increasing the transfer of energy from the waves to the WEC. [22]

Simon Lindroth, Buoy and Generator Interaction with Ocean Waves. Differ- ent aspects of the interaction between wave, buoy and generator are the focus of this Ph.D. thesis. The main work performed by the author at the research site is the implementation of optical measurements of buoy movement. This measurement has been compared to other measurements performed at the research site. Further on the performance of the WEC is evaluated with changing water levels, wave heights and spectral shapes [23].

Boel Ekergård, Electromagnetic Energy Converters - Rotating Motors and Linear Generators. This licentiate thesis deals with two particular areas applied to the linear generator, longitudinal ends effects in a linear permanent magnet generator and the WEC connected to a special resonance circuit [24].

24 Andrej Savin, Experimental measurement of lateral force in a submerged single heaving buoy wave energy converter. The results of offshore experiment where strain gauge sensors are instrumented on the capsule and the inner framework of the structure are presented in the thesis. Stress estimation analyses using strain gauges are carried out. A method for calculating the inclination angle between the WEC and the buoy based on the strain measurements on the capsule and the measured line force has been carried out [25].

Kalle Haikonen, Environmental Impact from Wave Energy Converters - Underwater Noise. The licentiate thesis presents a study of underwater noise from a full scale WEC [26].

25 3. The Lysekil project

3.1. Project purpose, challenges and project stages 3.1.1. Papers describing project purpose Paper V “Catch the wave to electricity” describes the outline of the project the purpose as well as the challenges in extracting energy from water waves. Paper XI “Offshore underwater substation for wave energy converter arrays” describes the outline of the electrical system for the suggested wave power park.

3.1.2. Purpose The purpose of the Lysekil project is to thoroughly study the Uppsala Uni- versity wave energy concept, in real conditions and over several years.

3.1.3. Challenges Some of the major challenges are: the investment costs associated with large structures, the survivability of parts exposed to the large powers and forces from the ocean, excessive over-dimensioning needed to handle mechanical overloads and long life mooring difficulties. Due to the nature of waves, the power carried by a wave group may vary by a factor of 50 over fractions of a minute. The peak powers handled by a wave power plant will thus be much higher than the average power. Since the power of ocean waves increases with the square of the wave height, huge forces are manifested during storms and must be handled by the wave power plants. Survivability is thus a key issue in wave power technology.

3.1.4. Theoretical verification stage FEM – based simulations of the linear generator were performed at the department before the experimental verification stages. The electrical design of the generator was simulated [30]-[36].

26 3.1.5. Experimental verification stages The first stage As a first step in verifying the theoretical simulations of the PM synchronous linear generator, an 8 kW machine was designed and constructed at the Ång- ström laboratory. The experimental setup was run in a laboratory test of a one-half wave period of a 2 m high sinusoidal wave. The experiment demonstrated very good agreement with the FEM-based simulated results [36], [37].

The second stage During the second stage of the experimental verification, a buoy with a diameter of 3 m was installed off the Swedish west coast in the proximity of Lysekil. The buoy was attached to a set of powerful springs located on a concrete foundation on the ocean floor. A force sensor was connected between the buoy and the springs to study the interaction between them. Ac- celeration of the buoy was measured. The study was a part of the design of the future wave power plant. The experiment was performed in 2005, between March 16 and May 3 when the springs were retrieved for analyzation. The experimental set up was fully operational until April 6, Paper I, Paper XXIII, [38].

The third stage A full-scale wave energy converter was constructed and deployed at the research site. Once it was placed on the seabed at 25 m depth, the power plant was connected, via a 3 km long sea cable, to a measuring station located on a nearby island where experiments with different loads were performed. Three experimental periods were performed with the first WEC during the third verification stage. Between March 13 and May 24 2006 the linear generator was connected to a plastic fiber line with braded cover. The line was connected to a cylindrical buoy with diameter 3 m and height 0.8m. The electrical output of the generator was connected to purely resistive loads. The voltage was measured and power for each half hour was calculated and presented in Paper II. A model of the WEC was developed and verified by experimental data Paper III. The second test period of the first linear generator was between March 3 and July 27 2007. The same 3 m buoy was used. The line was still made of plastic fiber but a new guiding system was constructed. The first guiding system had four wheels with small radii, only a few times the radii of the line. After the first test period when the guiding section had been unscrewed and raised to the water surface, it was discovered that the wheels were stuck. Water works as lubricant between the guiding system and the line, a development of the wheel guidance system that

27 ensured low resistive torque throughout the WEC lifetime where discharged. Instead a solution based on low friction and bigger bend radius was suggested. A funnel was constructed and put in place on the generator by divers. Force and acceleration was measured as well as voltage. Force was unfortunately only measured until March 15 when the signal cable between the buoy and the force transducer broke. During this period the first tests with a non-linear loading of the generator were made. The non-linear loading was achieved by rectifying the AC voltages from the generator before loading with capacitors paralleled with resistors. It is decried in the following papers: Paper VII, Paper IX and Paper XII. Between May 21 and October 26 2008 a torus-shaped buoy connected to a steel wire with a plastic cover was tested. Force and acceleration was measured during this period, Paper XIII.

The fourth stage A small wave power park was installed at the Lysekil wave energy research site during 2009. It consisted of three wave energy converters connected to one submerged substation. The purpose was to evaluate the system design, rectification of the individual generator voltages and collecting of the energy on a common busbar, Paper XI, Paper VI, Paper VIII.

3.2. The research area Paper I is the first paper that describes the outline of the research area. The paper describes why the area was chosen. It also gives an overview of the wave measuring system and the force measuring system. Paper IV thoroughly describes the research area. It describes: the initial study, the connection to shore, the measuring station, the deployment of the first WEC, examples of measured line force, WEC output voltage and power, vertical buoy speed calculated from acceleration measurements. The power absorption depending on the wave climate is also presented as well as the electrical system and a sample of the output in DC load. The origin to the right side of Figure 1 was presented in Paper IV and has been updated with the switchgear. The observation tower is presented. Examples from the environmental studies performed are displayed as well as the energy content of the wave climate. Paper XVI describes the status 2011 and what experiments have been performed up to November 2010. The paper also gives an overview of which buoy types and how many generators that have been tested at the research site.

Figure 1 shows the location and some of the devices that have been tested at the research site. Figure 2 shows the positions of the generators and the ob-

28 servation tower during the park test. Figure 2 (a) shows the measured positions of the generators relative to each other. Figure 2 (b) shows the distance between the marking buoy of L3 and the observation tower, seen from a satellite, before the buoys were attached to the generators.

Figure 1. The location of the research site to the left in the figure and different devices at the research area to the right.

Figure 2. The measured position of the WECs at the park test to the left and the positions of the marking buoys of the generators relative to the observation tower to the right. The generators were placed on the ocean floor before the real buoys were connected to them.

29 4. Theory

4.1. Wave theory 4.1.1. Water particle movement Many of the previous theses from the wave power group in Uppsala have described the average power in the waves and presented general formulas. To understand the nature of the waves as well as the qualities of the energy in them, basic Newtonian physics is very important. In detailed studies of water waves, the momentum of each single water particle is influenced by the surrounding ones. For pure harmonic waves the motion of each particle is elliptic-harmonic. When the depth exceeds half a wave-length, it can be assumed that each particle describes a circle, with constant angular velocity. The radii of these circles are given by an exponential function that depends on the distance to the surface and therefore diminish rapidly downwards [27]. The velocity of a particle lying in a surface, must be tangential (or zero) relative to the surface, because otherwise we should have a finite flow of fluid across it [39]. The energy in the wave can be divided in kinetic, gravity potential, and surface tension energies. Water surface waves are a combination of both transverse and longitudinal components. The transverse component accounts for the vertical motion of the water surface, while the longitudinal component refers to the horizontal motion due to waves. Due to the longitudinal wave motion, the wave crest takes a sharper shape, while the wave trough expands horizontally and becomes smoother; this is the characteristic of the Gerstner wave in Figure 3. Gerstner was the first to publish a complete solution to a wave theory in an infinite deep channel in 1802 [3], [6].

Figure 3: Drawing from Gerstners publication Theorie der Wellen. The horizontal lines are the lines of equal pressure.

4.1.2. Wave group theory One can be misled into thinking that the ocean waves propagate smoothly and consistently based on observations of very small waves in lakes. But when waves are created by the wind, a spectrum of waves with different frequencies, length and heights are created. These waves travel with different phase speed, in deep water longer waves travel faster than shorter waves. All the waves interact with each other and forms constantly changing wave- forms. Sometimes the velocity of all the particles in the wave has the same direction and a very big wave is created. Other times the velocity has different vertical directions and the motions of the particles interact in such a way that the velocity of the surface becomes zero. If the frequency spectrum in the wave is narrow and two frequencies’ dominate, the motion of the water surface results in a beat pattern. The interaction between the frequency’s in the ocean waves are the main reason to the varying measured forces and powers presented in this thesis. The interested reader is recommended to read the thesis Hydrodynamic Modelling for a point Absorbing Wave Energy Converter [22].

31 4.2. Generator theory Maxwell’s equations 1-4 are the basis for generator design and modeling. ∂B E −=×∇ , (1) ∂t

∂D jH +=×∇ , (2) ∂t

=•∇ ρ D c , (3)

B =•∇ ,0 (4) E is the electric field: B is the magnetic flux density, H is the magnetizing field, j represents the electric current density, D is the electric displacement ρ field and c is the free charge density. Maxwell’s equations are used in the simulation software when designing the linear generator. A number of previous dissertations with focus on generator design and/or simulation have been presented and are available for the interested reader [14], [40] - [45]. After the design of the generator when the no load voltage dependence due to translator speed is known, a simpler model of the generator with a no load voltage, a resistor and an inductor is often sufficient. The no load voltage ei in this model presented in Figure 4 can be described by a function based on the translator speed and the common area between the translator and the stator. This is used in section 4.4 to describe the damping factor. The damping factor is very important for the power absorption of the WEC.

Figure 4: Simple equivalent circuit model of a generator

32 4.3. Damping factor If the buoy exactly follows the wave no force is working on the generator, (the force in the line is equal to the force from the spring and the weight of the translator). For energy to be absorbed from the water the buoy must perform a work on the water particles in it. The motion of the buoy must lag the motion of the wave to absorb energy from the wave. When absorbing energy in from an ocean wave a wave with different vertical direction than the incoming wave has to be created, the result of the two waves after the absorber is a smaller wave. Falnes writes 2001 “Maximum power is absorbed from the wave when the destructive interference between the re-radiated wave and the incoming wave is largest.” [46]. The control of the buoy at the Lysekil research Site (LRS) is done by changing the loading of the WEC.

4.4. WEC theory 4.4.1. Model of a point absorber for heave motion In Paper III a model of a wave energy converter using potential wave theory is described by Ericsson. The author of this thesis has contributed with com- plementary experimental results that verify the model. The model explains the function of the WEC in a clear way; therefore this model is briefly summarized in the theory section. In Figure 5 the parts of the WEC is shown to the right and the corresponding mechanical model to the left.

Figure 5: The real WEC to the right and the model of the WEC to the left.

The buoy in the top of Figure 5 is accelerated by the ocean waves. The acceleration of the buoy depends on the mass that is accelerated and the forces working on it. The acceleration of the translator depends on the line force, the electromagnetic force from the stator the force from the energy storage spring and the endstop forces. The coupled integral differential equations of motion are given in equation 5.

∞ η −−∗−∗=+ α b  e  )()()()()( FFtythtftymm wh , (5) )( +−−= FFFFtxm p  semw endstop

The excitation force of the buoy is the convolution product between the impulse response function and the incident wave elevation (without the buoy present) in the center of the buoy as described by equation 6. ∗= η e e ttftF )()()( (6)

The impulse response function th )( is entirely described by the radiation resistance, equation 7

34 2 ∞ )( = dwwtwRth .)cos()( (7) π 0

The hydrostatic reaction force is zero when the buoy is in its equilibrium position. When the buoy is displaced from its equilibrium position the hydrostatic reaction force is the weight of the water volume displaced given by −= ρ h w tygStF ).()( (8)

The connection between buoy and translator is modeled as a very stiff spring with a lineforce Fw ),( >− xyxyk =  w Fw  (9)  ,0 else, where x is the displacement of the translator from its equilibrium position, in a calm sea state it is the centre of the stator. y is the displacement of the buoy from its equilibrium position in a calm sea state.

The force of the spring connected to the lower end of the translator pulls the translator downwards in wave troughs. The force is described by equation 10. The spring functions as energy storage so that electric energy can be delivered also in the downward motion of the wave.

+= s static s xkFF , (10)

Equation 11 and 12 describes the upper and lower end stops that are very stiff springs, )( <−− xllxk =  uuu Fu  (11) 0 else,

)( −<+− xlxlk =  lu l Fl  (12) 0 else.

35 The damping force in the model of the generator depends on the electrical power. The power depends on the electrical loading circuit, the active area between the translator and the stator and the speed of the translator F = γ x A x x()()( t). em fac  (13)

Equation 14 describes the active area that is defined by the position of the translator relative to the stator. ls is the stator length and lp is the translator length (piston in Paper III).

+≥  ,0 llx sp ))(2/1(  −≤  ,1 llx sp ))(2/1(  A = fac  (14)  11   −+ elsexll ,,)(   sp  ls 2 

36 5. Wave climate measurements at the research site

5.1. Articles with main focus on the wave climate measurement at the research area Wave height data is essential for all research activity at the research site, especially for evaluating WEC performance. Articles that have their main focus on the wave climate at the research site are Paper I and XV. Paper I describes the early stage of the research area and the wave measuring system. Paper XV describes the absorption of the WEC as a function of both wave height and wave period.

5.2. Measurement set up The wave height as well as the wave period has been measured since April 2004 with a Waverider measurement buoy from Datawell oceanographic instruments. A drawing of the buoy and its mooring system is shown in the right side of Figure 6. A small buoy floating at the sea surface moves up and down with the waves. The vertical displacement is determined by measuring its vertical acceleration and integrating it twice. The value from the second integrator is converted to a frequency and transmitted ashore over an HF radio link as an FM on a sub carrier. The signal is received at Kristineberg Marine Research Station, located a few kilometers form the buoy. The frequency is detected and converted to digital values and stored in a computer. Raw data files are created with a sampling rate of 2.56 Hz. An example is shown to the right in Figure 6.

Figure 6: Example of wave data to the left and the Waverider wave measuring system to the right.

37 6. Measurements with the measuring equipment placed inside the buoy

6.1. Background to buoy measurements Measurements of the buoy movement with sensors placed inside the buoy are described in Paper I, Paper III Paper IV, Paper X, Paper XIII and Paper XIV.

When a point absorber is used together with a linear generator in a WEC with limited stroke length, an impulse force is created each time the translator is accelerated upwards from a stationary state and each time the buoy is stopped by the translator reaching the top of the WEC. The maximum forces working on the system are essential as design parameters for any wave energy converter system. To achieve the correct design parameters when designing the first WEC, a force measurement experiment on a full scale buoy connected to a set of springs in parallel with a stopper rope was performed at the research site. This simple model was chosen because it was assumed that the maximum force appears when the WEC is disconnected from the grid because then there is no damping force between the translator and the stator. Additional to the force measurement, accelerometers were placed inside the buoy. The measurement system was initially designed for the dedicated force measurement set-up but the basic design was reused in all of the following experiments. Therefore a thorough description of the first force experiment is made. In later experiments yaw rate gyro sensors were added to the system. To measure the forces in the torus-shaped buoy three force transducers were mounted as brackets for the chains connecting the wire to the generator in the WEC.

6.2. General measurement set-up When measuring inside a buoy moving relative to the seabed, it is not a good idea to transfer the measured data through a cable to a fixed structure placed on the ocean floor. A cable connection would be beneficial for power supply of the system and for the data transmission bandwidth, but the cable would constantly move and soon crumble. Instead the buoy is equipped with a data logging system that stores data for a limited time. Data series are then auto-

38 matically fetched by a computer program and transferred through the GSM network.

To the left in Figure 7 a schematic illustration of the experimental set-up is shown. The buoy is connected to a force transducer. The other end of the force transducer is connected to a rope which in turn is connected to six springs in parallel with one stopper rope. The stopper rope parallel to the springs limits the stroke length to the stroke length of the translator in one of the first simulated models of the linear generator. The picture in the middle of Figure 7 shows the buoy and the force transducer, the package of springs is seen to the right. A sensor cable intended for measurements of the angle towards the foundation follows the rope. The angle sensor gave ambiguous results for a short period; the reason for this was never investigated. The implementation of the sensor was not in vane because the cable was a very good supporting structure for the cable to the force sensor. The force transducer is protected from water leakage by a steel bucket filled with two component resin. The data for the experimental set-up can be found in Table 1.

Figure 7: The experimental set-up used in the first force experiment.

TABLE I DATA FOR EXPERIMENTAL SET-UP

Parameter Value

Buoy radius 1.5 m Buoy height 0.8 m Buoy mass 950 kg Force from the set of springs (7065 * l – 2666) N Mass of spring unit 400 kg End-stop-rope length 2.65 m

39 6.3. Measurement system The measurement system consists of a force transducer: amplifiers, one GSM data logger and three accelerometers. Later yaw rate gyros were added to the system.

6.3.1. The force sensor Directly beneath the buoy the line force is measured with a force transducer placed between the buoy and the line. The force transducer used in experiments with cylinder shaped buoys was HBM U2B 200 kN and the one used on the torus-shaped buoy was HBM 5-200K-Y. Force sensors are usually designed on a metallic structure with well known strain. On this structure strain gauges are placed and are seen from the outside as an ordinary Wheat- stone bridge. However the sensor shown in Figure 8 has eight strain gauges and a few resistors building an unknown internal circuit probably used by the manufacturer to optimize the performance of the sensor.

Figure 8: To the left, Stefan Gustafsson holds the force transducer used in the first experiment. The inside of the force sensor HBM 5-200K-Y used on the torus-shaped buoy is shown in the middle. Close up of one strain gauge to the right.

6.3.2. Accuracy of force measurements The choice of sensor range is important. If the range is much wider than the measured forces, the sensor signal is low and small forces may not be detected because of noise. The limit is the noise that origins in the resistance of the strain sensors. Noise from the amplification is added to the signal. The quality of the first stage of the amplifier circuit is most important for good S/N ratio [47]. If the rated range of the sensor is smaller than the maximum forces, the signal from low forces has a better S/N than if the sensor is over

40 dimensioned. However the peak values will not be measured with the specified accuracy. Most of the force sensors will give an output value outside the specified range, if the rest of the data acquisition system can handle this output without distortion, measured data might be used with a lower accuracy.

Figure 9: Example of the measured forces.

The sampling frequency is also one important parameter for the accuracy of force measurements. The effect of low sampling frequency can be visualized in Figure 9, where the sampling frequency is 8 Hz. If the two neighboring samples in each peak are about the same size, it is likely that there exist a stronger force in-between them. If one of them is much stronger this might be very close to the peak. If the sample before and after the peak value are about the same size it is likely to be the correct peak force. There will be a deviation of the measured value of peak forces of the same size. However if many waves of equal size are measured there will be a distribution of measured peak forces, the correct maximum peak force will be found within the distribution. For detecting rare extreme forces the sampling frequency must be sufficient even if interpolation can be used to anticipate the peak value.

6.3.3. Electronics The electronics consisted of amplifiers and accelerometer sensors. The physical implementation of the accelerometer and amplification circuits is shown in Figure 10 and the schematics in Figure 11.

The accelerometer circuits The experiment was equipped with accelerometers measuring in the x, y and z direction. Calibration of the accelerometers was made by measuring the earth gravity in two directions for each accelerometer.

Figure 10: The amplifiers and the accelerometers used in the force experiment.

The amplification circuit

Figure 11: Schematics of amplification and accelerometer circuit.

The Wheatstone bridge needs a fixed DC excitation voltage and an amplification circuit to produce a voltage that is well suited for input to the data logger. The excitation voltage sets the maximum output voltage of the force sensor; the sensor used has an output of 2 mV/V of its specified range. A high output voltage is beneficial because it gives a higher signal to noise ratio, unfortunately it also results in higher energy consumption. The feeding voltage to the force transducer is one big energy consumer in the designed data logging system hence there is a trade-off between this two design parameters. One 12 volt 340 Ah battery was chosen. The sensor resistance

42 between the connections of the DC source is 345 Ohm. For the first experiment the feeding voltage was chosen to 7.5 V. This voltage was stabilized by a voltage reference. The sensor signal was amplified by two different operational amplifier circuits, one with a gain of 100 and the other with a gain of 400. This was done to increase the resolution of lower forces after digitizing of the data logger, and still be able to measure up to 200kN.

6.3.4. The data logger It was a time consuming task to find a useful data logger to place inside the buoy because of two very important design parameters of equal importance: the data logging system must consume low power and be able to sample fast enough to detect the peak forces. The power consumption demand, led to that no measuring system based on a consumer- or standard industrial computer could be used, solutions where the sampling frequency is not a problem. The other problem is that data loggers with low power consumption are most often used to measure slow changing processes. Temperature monitor- ing is the big market; hence most data loggers are designed to sample with a maximum sampling frequency of 1 Hz. The data logger chosen, shown in Figure 12, has a maximum sampling rate of 16 Hz and a storage capacity of 32000 samples. The A/D converter uses 12 bits. If all 8 channels are used the logger can store 4 minutes of data before data has to be transferred from the logger through the GSM network to a hard disk drive on a personal computer. In the force measuring project the sampling frequency 8 Hz was chosen to extend the time between data transmissions from the logger. It was discovered when processing the data presented in Paper XXIII that 8 Hz are a low sampling frequency for measuring extreme forces originating from ocean waves. The maximum sampling frequency 16 Hz was used when force was measured on real WECs.

Figure 12: The data logger placed inside a water tight box.

43 6.4. Calibration The force transducer, the differential amplifier and the data logger were calibrated all together by applying 15 different forces with an Instron 8516 ser- vihydralic fatigue tester2. The test system has an accuracy of 0.5 % of the indicated load. The data logger has an accuracy of 0.1% of the measurement range. The measurement range is 200 kN. The force transducer has a linearity variation of < 0.1 % and an influence of temperature on sensitivity that is 0.01 %/ºK. Force measurement range is 200 kN. The calibrated measurement system was used to measure the package spring constant by measuring the force at different spring lengths.

6.5. The launch of the first force measurement at the research site The experimental equipment was constructed and built at the Department of Electricity in Uppsala. After assembling at the Lysekil harbor, Figure 13(a) it was towered to the research site, Figure 13(c), and submerged at Lat N 58° 11.7317’, Long E 11° 11.4479’, Figure 13(d). The wave measuring buoy was positioned at Lat N 58° 11.740’, Long E 11° 22.340’. The distance between the measurements was roughly 100 m.

Figure 13: The preparations of the launch of the experiment at the Lysekil harbor are performed by local workers (a). The author stands beside the buoy (b). The towing of the barge (c). Submersion of the foundation connected to the set of springs and the buoy (d).

2 http://www.instron.us/wa/home/default_en.aspx (retrived 2012-07-08)

6.6. Developments and changes of the measurement system in following experiments 6.6.1. Buoy measurements with the first WEC L1 We were not able to capture any force data during the first offshore test of the full scale WEC because of data log failure. The electronic were well sealed inside the buoy, and could not be accessed offshore. In the following experiments the electronics were placed behind a water tight inspection hatch. During the second test period force data was collected during 12 days. The maximum measured force during the period was 148 kN, see Figure 14. The peak was measured at a significant wave height of about 2.5 m (the extreme value was found in the shift between two calculated half hours of higher and lower wave heights). The reason for measurement failure in the second test period was that the signal cable to the force sensor broke. This was probably due to the lack of support of the cable between the buoy and the sensor. Accelerometer data was collected during the whole second test period.

Figure 14: Measured peak force for the cylindrical buoy. The significant wave height was 2.3 - 2.7 m.

A few changes were made before the buoy was reconnected to the generator. In the dedicated force measurement experiment the accelerometers were placed in the same container as the measurement electronics and the battery. The equipment needed to be placed as close to the center as possible because of the acceleration sensors, that had the consequence that it was not possible to access the data logger in the buoy during the offshore experiment because the man hole was positioned off center. To solve this problem for the second

45 test period of L1 the accelerometers was mounted separately in the centre of the buoy and the data logger directly beneath the man hole, the lid that the antenna is attached to in Figure 15(b). The hose that protected the signal cable 2006 is seen in Figure 15(a). Figure 15(c) shows the force sensor and the hose 2007.

Figure 15: The buoy is loaded in Lysekil harbor 2006 (a). Kalle Haikonen is preparing the buoy for reconnection with the generator in 2007 (b). The force transducer, the chain and the hose protecting the sensor cable (c).

6.6.2. The torus-shaped buoy In the torus-shaped buoy the accelerometers are placed together with the data logger inside the torus (off centre) Figure 16 (d). The measuring system was complemented with yaw rate sensors to be able to compensate for the influence of gravity due to tilt of the buoy, which was considered the main parameter disturbing the horizontal accelerometer measurements. It was antici- pated that the rotation of the buoy would be small and the influence of a small rotation on the acceleration measurement could be neglected, this was an incorrect assumption. It became very clear when we processed the measured values used for Paper XIV, that the rotation of the buoy can not be neglected even if it is small. Because the acceleration is measured in the buoy coordinate system and the calculated displacement is presented in the coordinate system of the earth. For example if the buoy is accelerated towards the east and then rotated 90 degrees before it is accelerated back towards the west, the acceleration is picked up by the other sensor, resulting in an unchanged speed from the first measurement. The calculated speed from the measurements will be erroneous booth in speed and direction. Hence both tilt and rotation (or angular position) are needed to compensate for errors when calculating the horizontal movement of a cylindrical point absorber from acceleration measurement.

46 The major change of the measuring system constructed for the torus buoy is the force measurement system. One flexible suspended force sensor is replaced with three sensors fixed to the hull and used as brackets for the chains connecting with the generator, se Figure 16(a, c, e). This configuration eliminates the risk of cable failure due to sensor and sensor cable movement. To calculate the force in the line connected to the generator the angles between the line and the chains must be known. An approximation can be done by assuming that the angle of the wire is fixed perpendicular to the plane of the torus. For the torus buoy no. 2, Figure 16(b), calibration constants in this special case were achieved by hanging a weight in the buoy and measure the force in each sensor. The calibration constants for torus 2 are shown in equation 15.

Figure 16: The author is working with the first torus buoy tested 2007 (a). The second torus buoy tested from October 30 2009 (b). The buoy is connected to the generator wire with chains forming a tripod (c). The measuring equipment in one of the battery boxes (d). The connection between the force sensor and the chain, the protection of the signal cable is seen as a snake on the surface (e). Halvar Gravråkmo (f). The force sensor is mounted on the buoy (g-j).

−+++−= Wire GF 1 G2 G3 63465.06557.019078.0 (15)

47 The correct force in the wire connected to the generator is found by calculating the resulting force of the three forces in the chains forming a tripod structure. The three forces in the chains forming the tripod structure are known. The angles of the chains towards the buoy plane are known. For torus no. 2 the design angles is 45° but the calibration formula indicates that the real angle for G1 is 51.3, G2 is 44.2 and G3 is 40.5 degrees. On torus no 1, the force sensors are mounted with an angle of 55 degrees towards the plane of the buoy.

6.6.3. Cylindrical buoys 2 and 3 Yaw rate sensors were also mounted inside the cylindrical buoys 2 and 3. The specification from the manufacturer was used to calculate the yaw rate from the yaw rate sensor voltage output. The total redesign of the buoys allowed locating space with easy access to the measurement electronics in the centre of the buoys. The cable to the force sensor was protected by a 2 inch armed rubber tubes wired around the sensor.

48 7. The measuring station

7.1. Introduction The measuring station is the onshore connection point of the Lysekil research site wave power test area. It is briefly described in Paper IV, VII, XII, XV and XVI. Historically most experiments with wave power devices have either been performed in a basin or with the measuring equipment and loads placed inside the device where data has been logged for later analysis. One of the base criteria for the Lysekil research project was that the energy should be transferred to the mainland. Of course we realized that some kind of on shore structure was needed but at first we thought that a steel cabinet containing the measuring equipment would be sufficient. When the plans emerged it became clear that a measuring station was necessary to house the measurement equipment and the electric loads. The outside of the station is shown in Figure 17. The meadow at Härmanö before the measuring station was built is shown in (a). The Station during installation of the first loads and the measuring equipment, (b). The station in the spring during the first test period of L1, (c). The station the day after the substation was submerged at the research site 2009, (d).

Figure 17: The meadow (a). The station during installation of the first loads in the winter 2006 (b). The measuring station in the spring 2006 (c). The measuring station in the “spring” 2009 (d).

49 The equipment inside the measuring station has developed and expanded. The first results were measured with a data logger connected to a lap top, Figure 18(a,b). Now the station has specially designed equipment inside the cabinets as shown in Figure 18(c).

Figure 18. The author checking the measurements on a laptop the day when L1 was submerged at the research site (a). Stefan Gustafsson helps with the final connections of the WEC to the loads (b). Deepak Elamalayil Soman, Remya Krishna, Wei Li and the author at the measuring station 2011 (c).

The power cable from the research area is connected to this station. A schematic of the generator and the sea cable is shown in Figure 19.

Figure 19. The circuit model of the generator and cable without the voltage symbol in the generator.

50 The assembling of the loads were in Uppsala, Figure 20 (a). The loads were then transported to the measuring station where they were mounted on the back of the garden shed that has developed to the measuring station.

Figure 20. Cecilia Boström and Erik Leijerskog are building loads, super- vised by the author (a). The first loads mounted on the back of the station (d). The loads are exchanged to wire wound resistors, loads connected to DC to the left (c). High voltage loads on the roof for the loading of the substation (e).

The power produced in the generator during the first test period was converted to heat over a three phase delta-connected resistive load. The first measurements were made with an oscilloscope and a USB data logger NI 6009 controlled by a laptop running a Labview program. It sampled with 128 Hz and measured both voltages and currents. Unfortunately this system was demolished due to short circuits between the generator neutral and Computer neutral. It was replaced with a compactRIO system with better isolation between the measurement ground and mains ground. The control, load and measurement system have since then been successively expanded.

The first loads used in the measuring station were built from semiconductor resistors, Figure 21 and Figure 20(a, b). It was discovered that they could not withstand any overvoltages from the generator and where therefore soon replaces by wire wound resistors, Figure 20(c, d).

Figure 21: The loads at the measuring station, as was, when the first data from the WEC were retrieved.

51 7.2. The first substantial data The data from the day when the WEC was submerged showed that it worked but the wave height was very low, hence no conclusion could be made from the measurements. Internet was not connected to the measuring station until a few weeks after the WEC become operational. The data for the first evaluation of the experimental is seen in Figure 22 and were transfered to the crew in Uppsala by MMS.

Figure 22: The original pictures with the first substantial data on which the first conclusions were drawn and the continuation of the project was decided.

7.3. Measuring Station Control System 7.3.1. The first compactRIO system CompactRIO is an industrial computer designed and manufactured by Na- tional Instruments. It is their version of the industrial computer system that generally is called Programmable Automation Controller (PAC). The PAC from National instruments consists of three major hardware parts: a real-time controller, one Field-Programmable-Gate-Array chip (FPGA) located in a separate backplane and different input and output modules. A FPGA consist of a large number of programmable logical gates. The main advantage of FPGA is that all calculations are made in one clock cycle; hence, the FPGA is very fast. It is also very stable. The drawback is that it

52 only handles binary values. In the CompactRIO system the fast binary data processing in the FPGA is combined with the generality of real time computer system.

Figure 23: The parts of the first CompactRIO system. The real time computer enclosure includes the communication and network connection (a). The back plane where the FPGA chip is located (b). Analog input modules (c and d).

The first CompactRIO system used in the measuring station shown in Figure 23 consisted of: a 4 slot 1MGate FPGA, a 9002 Realtime controller, one 9201 8 channel analog input module and one 9215 4 channel analog input module. The 9201 was used to measure the line voltages from the generator at the measuring station. The line voltages are measured over a resistor network because the module only measures voltages up to 10 Volt and the generator voltage is a few hundred volts. When the measured voltage is damped with the resistor network before measuring with the measurement module, the noise impact from the module noise on the measured signal is increased. Figure 24 shows a voltage measurement of the line voltage close to zero. The digital peak to peak noise voltage in the module is 25 mV but as seen in the figure it multiplies with the damping factor and becomes about 1 volt of the measured real voltage. Luckily the RMS noise is about seven times lower. Hence the impact of digital noise on average power calculated from measurements with this measuring set-up is not as big problem.

Figure 24: Noise on measurement data

53 In the 20 second example of the voltage output in Figure 25 the noise is not visual. The problem with the noisy data has been to detect the zero crossings of the generator voltage when calculating the translator speed.

Measured Output Voltage @ Hs=4 m

150 100 50 Phase A 0 Phase B Phase C -50 Voltage (v) Voltage -100 -150 02468101214161820 Time (s) Figure 25: Example of output voltage from WEC connected to resistive load.

7.4. Non-linear load During the second test period, the measuring station was upgraded with a rectification bridge and a huge capacitor bank paralleled with resistive loads. The total resistive values that can be selected are 9.17, 13.75, 18.34, and 27.5 ohm. This system was installed in the measuring station during the au- tumn 2006 and spring 2007. Electrical circuit is shown to the left in Figure 26 and the physical implementation to the right.

Figure 26: The measuring set-up during the tests with non-linear load.

54 The upper colored part of Figure 27 shows the voltage from the generator before rectification. The horizontal lines show the capacitor voltage. Beneath the current is presented, current only appears when the generator voltage is higher than the capacitor voltage plus the diode voltage.

Figure 27: Voltage and currents from the WEC when it is connected to the non-linear load.

7.5. The measuring station during the park experiments 2009 During the wave park experiments the measuring station was fully remotely controlled. The control was performed by remotely logging in to a personal computer placed inside the station. With a Labview program the station PAC was controlled. Another computer was assigned the task to communicate with the CompactRIOs in the substation; it was accessed via the first computer. Other peripheral instruments as oscilloscopes were also connected to the local network and could be remotely accessed from Uppsala. New loads designed for the loading of the substation were mounted on the roof of the station. The communication structure is shown in Figure 28. Short periods of data could be transferred through the internet but large amount of data needed to be transport by a car.

Figure 28: The communication structure of the research area and the data transfer to Uppsala

55 8. The Marine Substation

The marine substation is described in Paper VI, Paper VIII, Paper XI, Paper XVI and Paper XVIII.

8.1. The enclosure The marine substation is designed for installation at 25 m water depth at the Lysekil Research site. The base material for the enclosure is a 3 bar pressure vessel with a volume of 3 m3. This has been equipped with connectors and mounting plates welded to the enclosure. Electronics are mounted on these plates

Figure 29: The enclosure is bolted and the substation is ready for transport in the left picture. The inside of the substation is presented in the middle and to the right the switchgear is transported to harbor for service.

8.2. The electrical system 8.2.1. The electrical main circuit The electrical main circuit of the substation is shown in Figure 30. The schematic is shown in (c). The contactors and the transient voltage surge suppressors are shown in (a). It is possible to connect all the generators via rectifiers (b) or bypass the whole substation and connect one generator to the main output. The rectified currents are collected in capacitor storage (d). The DC voltage is transformed to AC with the inverter (e) and sent to the transformer (f).

Figure 30: The main circuit of the first substation.

8.2.2. Auxiliary Power Supply System The substation is powered by 24 volt DC. The distribution system is shown in Figure 31. The batteries are charged from the AC output of the substation via a transformer and of-the-shelf battery chargers. The distribution circuit does not contain any fuses because it is not possible to replace them. Instead the power is distributed via diodes and resistors calculated to withstand a short circuit. Critical resources are powered from all three 24 volt systems.

Figure 31: The schematics of the 24 volt battery system.

57 9. The control system of the marine substation

The control system is described in Paper VI and briefly mentioned in Paper XVIII and Paper XIX. In Paper VI the substation is denoted Low Voltage Marine Substation (LVMS).

The substation control and measurement system consists of three PACs inside the substation and one land based PAC in the measurement station (section 7.3). The measured data is stored on the hard disk drive of a land based personal computer. The communication structure is shown in Figure 32.

Externsl hard disk PC PAC 4 drive

Mesuring station Switch

4 Pair SHDSL Modem

Substation 4 pair SHDSL Modem

Switch

PAC 1 PAC 2 PAC 3

Figure 32: The communication structure including the point to point communication between the substation and the measurement station.

PAC 1 is a safety and relay control system. PAC 2 forms together with the switching transistors and the drivers the inverter, which converts the DC voltage to AC voltage. PAC 3 is a dedicated data acquisition system that logs WEC data and environmental data from sensors inside the Substation. PAC 4 controls the resistive power loads placed outside the measurement station. It also measures the voltages and currents in these loads.

Figure 33: One PAC (a). Interface card (b). PAC1(c, h). PAC2(d, g). PAC3 (e, f). Supply system (i). Modem (j).

9.1. The safety and relay control system. PAC 1 uses only the FPGA section of the PAC. The real-time computer is not used to increase system stability. A real-time program consists of many processes that depend on each other. There is always a risk that one process blocks another process, if this occurs hardware reset might be the only way to solve the problem. The contactors controlled by PAC 1 are shown in Figure 30(a, c) and Figure 31. The PAC is programmed so that the WECs are either switched to rectification, or one WEC is connected directly to the output of the substation and the other is disconnected. The third alternative is that all WECs are connected to individual resistive loads. From the beginning the protection algorithm in PAC 1 was designed to measure the voltage and current in neutral conductor of the WEC. The WEC should was to be disconnects from the substation if the current exceeded a pre defined value. When the WEC is connected to an unlinear load as a DC voltage, the loading of the generator is not symmetrical and the current in the neutral conductor is much higher than for a three phase generator connected to a symmetrical load. It was decided that the neutral conductor should be disconnected. With PAC 1 it is possible to choose one of the five transformer outputs. It also controls the power supply to the two real-time systems. In this way, it is possible to make a manual reset of these systems in case of deadlock. It measure temperatures on the rectifier modules and on the IGBTs, located inside the inverter, see Figure 30(e). If the temperature reaches a critical temperature, the WECs are switched off by PAC 1.

59 9.2. The control of the inverter PAC number two controls the conversion of the DC voltage to 50 Hz AC voltage. The inverter inside the Substation consists of a PAC, IGBT driver electronic and 6 IGBTs, the drivers and the IGBTs are shown in Figure 34. The PAC measures the voltages and currents on the DC bus, the IGBT outputs and the transformer output. Based on the measured voltages and currents, it sends pulses to the IGBTs. It also sends the measured values to the land based PC that stores the data on a hard disk drive.

Figure 34. The drivers to tle left and the IGBTs beneath the DC supply copper bars to the right.

9.3. Dedicated data acquisition system The third system is a dedicated data acquisition system that measures the voltages and currents from each WEC and from sensors placed inside two of the WECs and inside the substation. The sampling rate is 256 Hz. The frequency of 256 Hz is chosen because the generator measurements could more easily be synchronized with measurements of the buoy described in chapter 6. 256 is a multiple of the sampling frequency of data logging system inside the buoy. The dedicated data acquisition system uses two 16 channel differential analogue input modules to measure the three phase voltages and currents from the three WECs. Two modules are used to measure the sensors in WEC 2 and one module is used to measure the sensors in WEC 3.

60 10. Sensors and measurements during the park test

Measurements inside the generators are made with the dedicated measurement system placed inside the substation. The analog output from the sensors was signal-conditioned inside the generator and then sent through a 70 m twisted pair cable to the substation where it was digitized. The collected data was then transferred 3 km with a point to point copper link from the substation to the measuring station, where it was stored on a hard disk drive. In both WEC 2 and WEC 3: the position of the translator, the magnetic flux at the stator teeth, the temperature on the stator segments and water leakage are measured. WEC 2 is also equipped with strain gauge sensors placed on the metal structure and laser sensors that measure the horizontal movement of the translator. Temperature, pressure and humidity inside the substation are measured by PAC 3.

10.1. Position of the translator 10.1.1. Measurement set-up Translator position is measured with a standard wire sensor from Micro Ep- silon. The translator is seen (resting on a board) in Figure 35(i) and the sensor in Figure 35( f). The sensor is attached to the upper end stop plate of the inner mechanical framework via a 50 mm thick rubber mat to protect the sensors from vibrations, see Figure 35( f). A hole is drilled in the plate for the wire that is fastened to the top of the translator, Figure 35(g). The wire is attached to the translator, Figure 35( h).

Figure 35: Measurement of the position of the translator in L2. The picture is adapted from Paper XIX.

10.1.2. Calibration Two different measurements were the basis for the calibration of the position of the stator in the linear generator L2, the stator position relative to the translator position, Figure 35(a-c) and the distance from the top of the rod piston to the top of the seal housing, Figure 35(d, e). Paper XIX discusses the two calibration measurements the Paper suggests that the most accurate curve fit is based on the measurement of the length of the piston outside the generator. The maximum difference between any of the three linear trend lines presented in the paper is 35 mm.

In Paper XIV an error estimation of the calibration of L3 is made. The calibration is based on two measurements; one with the top of the translator and the stator in line with each other and one with the translator lowered 500 mm. The estimated error is ± 7.17 mm plus an additional 5.18 % of the movement of the translator from the midpoint of the measurements, which is 55 mm below the point where the translator and the stator centre are in line with each other.

62 There are two main reasons for the big error estimation. First the sensor was calibrated with a separate voltmeter and not the measuring system inside the substation; because of this both the error from the voltmeter and the measurement system must be added to the error estimation. Second the endpoint positions of translator could not be measured because of the limited height of the overhead crane in the assembly hall. The error with which the calibration points were measured will propagate with increasing distance from the calibration points. The first value ± 7.17 mm is the sum of all inaccuracies that are independent of position of the stator. It includes: the linearity of the sensor ( ± 3 mm), accuracy of the voltmeter and the measurement system, the impact of a 20 degrees temperature shift on the amplification circuit and the measurement system.

10.2. Strain measurements Strain was measured on the structure and the enclosure of L2.

Figure 36: Strain gauge placement.

The inner framework structure consists of four corner pillars connected by twelve cross bars distributed between the capsule bottom plate and the upper end stop. WEC L2 is provided with strain gauge circuits both for bending strain measurements and for uniaxial strain measurements. The placements of the sensors are shown in Figure 36 Ch 13 and 14 measure bending strain of the capsule wall. The data are used in Paper XXI, most of the other sensors are used for Paper XX. The strain gauge amplifying electronics has two major design parts. First one discrete amplifier placed close to the strain element amplifies the differential signal from the measurement bridge. This signal is sent to an instrument amplifier that increases the gain further before the signal is sent to the substation, where it is sampled.

63 10.3. Search-Coil sensors Basic induction or “search” coils are inductive sensors composed of coils based on Faradays law. Air coils are very stable and linear, but their sensitivity is limited. A low noise preamplifier is usually mounted in close proximity to the induction coil, Figure 37(e). The search coil used is designed as squares on a two layer printed circuit board (PCB) with ten turns on every side, Figure 37(b, d). This design enables two ways of calculating the magnetic field. By summing the contribution from each conductor the magnetic flux can be calculated by using the formula for a straight wire. To use the method of the straight wire, the length of every conductor must be known. The other method is to calculate the total area of the coil and from that calculate the magnetic flux. The placement of the sensors are shown in Figure 37(a, c).

Figure 37: The search-coil measurements. The placement of the search coil as seen from above (a). The sensor (b). View from the side (c). Picture of the mounted sensor (d). The amplifier circuit (e).

64 10.4. Temperature measurements The stator temperature is measured with the precision centigrade sensor LM35. The sensor has an output of 10 mV/degree Celsius. Time multiplex- ing has been used to measure eight different temperatures. In this way only one conductor, between the WEC and the substation, has been used for the stator temperature measurement. The stator temperature and the air temperature inside the WEC are presented in Paper XVIII . It was found that

10.5. Water level and leakage detection Sensors were placed inside the WEC to detect water leakage, if a leakage found it also measures the water level inside the generator. The water level indicator circuit is a rather simple design with four wires installed on one of the corner pillars. The wires are cut at four different heights, 25 cm, 50 cm, 75 cm and 100 cm above the capsule bottom plate. Water contact results in a current in the conductor between a voltage source and ground, this current that can be detected with an amplification circuit. The leakage detector sensors were mounted close to the sealing for early detection of leakage.

10.6. Piston and seal housing displacement The piston and seal housing displacement was measured with laser sensors. This is described in detail by Strömstedt in Paper XIX.

65 11. Buoy and translator movements

Paper I, Paper III, Paper IV, Paper XIV, Paper XVII, Paper XXIII describes different measurement methods of the buoy motion. The Lysekil project is all about converting kinetic energy to electric energy adapted to be distributed in the electric grid. The first stage is to harvest the energy in the ocean wave and transfer it to buoy movement. During the experimental period described in this thesis several types of measurements of the position of the buoy have been performed. In the first force experiment with a buoy connected to a set of springs, described in Paper I and Paper XXIII accelerometers were placed inside the buoy. Speed can be calculated by integration of the measured acceleration. With a second integration position is calculated. Another method to find the movement of the buoy in the first experiment is to use the force sensor used in the experiment. The length of the springs could be calculated from the measured force and the spring constant, this method can not be used with experiments when the generator is connected to the buoy [35]. Instead the three phase voltage can be used to calculate the position of the translator. From the number of zero crossings and the generator pole width the distance can be calculated. The phase order gives the direction of motion. If the translator position is known the buoy position relative to the generator is known when the rope is stretched, translator position calculated with this algorithm is presented in Paper IV and Paper VII. An observation tower equipped with a camera was installed at the research site to monitor the area and by image analysis track the buoy movement, Paper XIV, [48]. The fourth method used at the research site to measure buoy position was to attach strain gauges to the enclosure of the wave energy converter, Paper XXI.

11.1. Time and “unnecessary measurements” One important task that is easy to overlook when designing individual measurement systems, is the time perspective and time correlation between other measurements. This became surprisingly clear when the force measurements should be plotted in the same diagram as the generator voltage in Paper IV. The curves did not match when we plotted them based on the stored time

66 data of each data series. The reason was that the clock inside a computer is not very precise. The time error of a computer clock increases with seconds every day. It was impossible to guess how to fit the force measurements to the voltage measurements by only using these two measurements. Luckily acceleration was measured and logged with the same measuring system as the force. Hence the buoy speed could be calculated and compared to the speed of the translator.

11.2. Accelerometer measurements and influence of gravity and rotation By integrating acceleration twice position is calculated. Integration will in- duce drifting errors but with a high pass filter the movement can be balanced around zero. It is shown in paper XIV that the filter parameters can be found that results in good accuracy of the vertical displacement of the buoy. Calcu- lating the position from horizontal acceleration measurements is much more complicated for a buoy moving on the ocean surface. The big problem is how to compensate for the disturbance from gravity and how to keep track of the two coordinate systems, i.e. the coordinate system of the buoy and the earth coordinate system The outputs from the accelerometers do not purely reflect the movement of the buoy in surge and heave. The measurements are also affected by the incline relative to the earth. If the sensor is directed toward or from the earth centre the accelerometer will be 100 % influenced by earth gravity. If it is shifted exactly 90 degrees there is no influence of gravity. The impact of the angle variation is not the same in the horizontal plane as it is in the vertical plane. The sensor output due to gravity is a sinus function of the angle towards the earth. Hence the influence of the pitch variation is small around the z direction but contribute substantially to the output of the accelerometers in the x and y direction. To compensate for this the buoy angle towards the sea surface can be measured. One method is to measure the yaw rate with yaw rate gyros. Yaw rate gyros were implemented but the usefulness of these could not be evaluated because of a second disturbance parameter, the slow and small rotation of the buoy (maximum 180 degrees). It was antici- pated that the rotation of the buoy could be neglected when calculating horizontal measurements but Paper XIV shows that it can not be neglected. The paper also shows that the accuracy of the vertical acceleration measurement of buoy movements in this application is good.

67 11.3. Measurement of the buoy position with strain gauges placed on the capsule Paper XXI describes a very interesting method of measuring the position of the buoy. The strain gauge mounted on the generator capsule seen in Figure 38, measures the bending strain of the capsule.

Figure 38. The placement of the strain gauges for bending strain.

The horizontal force on the funnel induces tension in the enclosure of the WEC. Hence the horizontal force from the wire can be calculated. Together with the measured line force between the line and the buoy, the angle of the line can be calculated. If the position of the translator is measured, it is possible to calculate the x, y and z position of the buoy.

11.4. Optical measurements The optical measurements of buoy movement with a camera placed in the surveillance tower are thoroughly described in [22] and Paper XIV. An overview of the system is given in Paper IV.

11.5. Comparison of the different measurements The presented measurements with the strain gauge sensors in Paper XXI, Figure 39(a, b) was performed during a wave climate of 1.3 m on a buoy with a 3 m diameter. The presented measurement in Paper XIV, Figure 39(c) was measured in a wave climate of about 2 m with a buoy of 4 m diameter. It is then from the presented data in the articles not possible to quantitatively compare the measurement accuracy of the two measuring methods from the published results, but the measurements with the smaller waves measure less movement than the measurement performed in a wave climate with higher

68 waves. The maximal horizontal movement as seen from the position of the camera is about 4 m and the maximum movement measured by the strain sensors is about 2 m. In Paper XIV the accelerometer measurements are evaluated and compared with the optical measurement and the translator motion. The calculated vertical motion shows good agreement with the other measurement methods. The calculated horizontal movement from accelerometers over estimates the motion mainly because of the rotation of the buoy that could not be compensated for.

Figure 39: The measured position with strain gauges @ Hs=1.3 m (b) and the line angle (a). The horizontal position as seen by the camera @ Hs=2 m (c).

69 12. Comments on presented power data from one single generator

Power output from one generator is mainly presented in Paper II, Paper VII, Paper XII (Journal version of Paper B) and Paper XXIV.

12.1. Power absorption with resistive linear load

Figure 40: The absorbed power, presented in Paper II (Applied Physics Let- ters, 90:034105, 2007).

The data presented in App. Phys. Lett, Paper II and Figure 40 called the “absorbed power” is calculated from the measured line voltages at the measuring station. The in addition to the produced power, resistive losses originating from the winding resistance and the sea cable is included in “absorbed power”. All other losses, as for example the friction between the stator and translator and losses in the magnetic circuit are not included. Hence the power absorbed by the WEC is higher than what is presented in Figure 40.

70 12.2. Power absorption with a unlinear load In Paper VII, Paper IX and Paper XII the WEC connected to a rectifier and then loaded with a smoothing capacitor in parallel with a resistive load. The measurement set-up is described in section 7.4. Figure 41 presents the absorbed power. The phase voltages and currents are measured before rectification inside the measuring station and power is calculated from these measurements. The electrical losses are calculated from the measured currents and the known resistance of the sea cable and the generator winding. Out of the tested loads, most power is absorbed when the one with the lowest resistance is connected in parallel with the capacitor. More data is needed to draw a conclusion about an optimum load in high energy states. The benefit of the capacitor storage is that the power peaks from the generator can be absorbed and smoothened. The drawback is that no power is absorbed when the voltage from the generator is lower than the capacitor DC voltage. A common DC bus is also a simple way to aggregate generators. The first time I saw the voltage and currents originating from the generator connected to the unlinear load, it became very clear that Newton’s third law “To every action there is always opposed an equal reaction” is superior and governs any other application formulas. The force in the rope ends are the same and there is a balance between force from the wave working on the buoy and the force in the generator governed by Lenz’s law “An induced electromotive force (emf) always gives rise to a current whose magnetic field opposes the original change in magnetic flux”. The instant voltage on the DC bus will limit the voltage from the WEC as seen in Figure 27. To match the speed of the buoy and the force working on it the current is varied instead of the voltage, as seen in Figure 25.

Figure 41: The power from the generator in the non-linear load experiment.

Figure 42: Absorbed power from the generator at DC load, relative to significant wave height.

12.3. Generator power output In the first papers focus is on power absorption of the available energy in the waves. The achievements that the early papers describe have had a very important impact on the general view on wave power as one of the sustainable energy sources for the future. When the technology matures, the focus must change towards system design to achieve efficiency, durability and service- ability. The ultimate goal is long term energy production. In Paper XXIV the data from the first experiments is presented with focus on the produced- instead of the absorbed power. Generator output is related to significant wave height instead of the energy flux. The delivered power at the generator output is presented in Figure 43. In a commercial wave power park the losses in the cable from the generator to the substation should be low because the WEC is located close to the substation. Therefore the 3 km cable to the measuring station is seen as part of the load.

Figure 43: The output power from the generator without any losses added (x on purple squares).

For a brief comparison between the figures in this chapter, the transfer function between significant wave height and energy flux can be approximated with equation 16. It is based on the measured wave climate at the research site in March 2007. ≈ 2 +− fluxEnergy March 2007 5569H s 5879H s 2105 (16) The power output in Figure 43 increases up to a significant wave height of almost 2 m corresponding to an energy flux of 12 kW/m. Above this level the absorbed energy shown in Figure 40 continues the smooth reduction of power absorption. However the output power in Figure 43 shows a much steeper change in the output above a wave height of 2 m. The simulation results presented in Figure 43, show the behavior of the WEC if the buoy would follow the motion of the water surface. Simulations are done with and without inductance and limited stroke length. The magnetic and other losses are omitted in the simulation.

73 12.4. Power calculations from L2, L3 When the generators are loaded with resistive loads different methods can be used to calculate the power. If the output voltage can be assumed to be a balanced three phase voltage the load can be delta-Y transformed and the phase voltages can be used. This is accurate for a perfect sinusoidal and balanced three phase voltage, but if the three phase output is not perfect the calculation introduces errors. The correct method would be to calculate the voltage over each load or by use phase currents and phase voltages. The correct method must be used if one phase is shifted 180 degrees as the case was with L2 and L33. The importance of this phenomenon is shown in Paper XXII were the damping factors for the three WECs are calculated when they are connected to equal resistive loads. “The damping factors applied to the WEC generators of L1, L2 and L3 during the second experiment of 19.7 days (linear damping) were 16.2, 7.4 and 7.3 kNs/m, respectively, all with the accuracy of +5%. These were calculated from 12 1-min raw data files sampled at 256 Hz”.

3 WEC L2 and L2 were faulty assembled.

74 13. Power smoothening effects from two or more WECs.

The power smoothing effect of many WECs in a wave power park described in chapter 2.2 is evaluated in Paper XXII. The effect of power smoothing is described with two generators connected to the DC bus of a submerged substation. The effect is also evaluated with three generators connected to individual resistive loads. The measurements were done with the measuring system inside the substation.

It was found that the electrical power from an array of directly driven point absorber WECs calculated from 88 min of data in a 4.6-kW/m sea state was found to be smoother than that from individual array members. The standard deviation of electrical power was reduced by 30% as a mean for converters L2 and L3. Results from a 19.7-day record, taken in mixed seas of mean intensity 2.24 kW/m (a maximum 20.6 kW/m), show that the reduction in standard deviation of electrical power was 80% as a mean for an arbitrary device in the three-WEC array.

13.1. Discussions about power smoothening effects and the outline of the collecting system The outline of the collecting system was described by Thorburn [13], [29]. Theoretically the specification of this voltage could be the same as for the grid connection point, but that has some drawbacks. Most often the grid connection point is an alternating current system. There is then a need of one inverter for each single generator to convert the varying frequency from the individual generator to one fixed common frequency. This inverter has to be designed to handle the maximum power from each generator, or some of the energy is lost. If the power from all generators could be collected on a common bus before the voltage is adapted to fit the grid voltage, the rating of the inverter is based on the peak power that should be transferred from the DC bus. The rating of the inverter connected to a common DC bus may be lower than the total rating of inverters connected to individual WECs.

75 14. Results from force measurements

Results from the dedicated force measurements are presented in Paper XXIII. Trend-lines for the average and peak values during the experiment period is presented in Figure 44. The slope of the peak line is about ten times the slope of the average line. The highest measured force was 84.5 kN, that is 26 % above the trend at Hs=1.94 m. The highest measured force with a WEC connected to the buoy shown in Figure 14 was 148 kN at a wave height of about 2.5 m.

Figure 44: Trend-lines for half hour bins of force data.

Figure 45 shows the number of peaks at different wave heights. When the wave height is low there are many peaks because the frequency of the waves is high. When the wave height increases the force of the peaks increases.

Figure 46 shows a bubble diagram that displays the number of peaks as a function of both wave height and wave period. Three regions can be clearly distinguished. In the lower left, the wave height is low and the energy period is short. In the upper left only forces lower than 10 kN are detected. In this region Te is longer and Te peak is of the same magnitude, hence this region consists mainly of swell waves. Between the two areas there exist wave cli-

76 mates with both swell and wind generated waves. Waves resulting in the maximum forces have a high wave height and energy period between 5 and 6 meters. Waves generated purely by local winds are probably distributed along a line connecting the lowest energy periods for each measured wave height. Swell and combinations of swell and wind generated waves are likely to be above this line.

Figure 45: Number of peaks at different wave height.

Figure 46: Number of peaks as a function of wave height and wave period.

77 14.1. Peak and average power The peak force and the repeating medium forces are very important as input parameters for the mechanical design of the WEC. The design parameters for the power collecting system are the currents and the voltages from the WEC. Peak power and average power are the key parameters. These parameters are the most important for the design of the collecting substation but all the connections between the WECs and the onshore grid connection point are influenced. Paper XXIV tries to evaluate the first WEC as a power source. In Figure 47 the power output from the generator are presented in relation to the wave height divided in bins of 0.1 m. Peak and average output power are calculated and displayed. The relation between peak and average is also displayed. At low wave heights the peak power can be more than 50 times the average power, this is not very important because the power is low. What is very important is that the relation between average and peak power is high when the energy state in the sea is high. For the bin with the highest wave height measured the relation is 17.

In Figure 48 simulation of the generator following the vertical motion of the waves is presented. Different generator parameters are excluded to show the impact of the parameter. The simulation may help to interpret the results from offshore measurements presented in Figure 47 and Figure 43. The tendency of the results from the simulations was expected but it was a sur- prise that it was this clear. The inductance will decrease the peak to average relation. The stroke length will increase the peak to average relation starting a significant wave height 30 percent lower than the stroke length. The simulation with the same stroke length and inductance as the real generator show good agreement regarding the peak to average value.

Figure 47: Measured peak and average power output from WEC.

Figure 48: Simulated peak to average output with parameters from the first WEC.

79 15. Summary of Papers

Paper I Experiments at Islandsberg on the West Coast of Sweden in Prepara- tion of the Construction of a Pilot Wave Power Plant The author contributed to the building of the experimental set up together with the first author. The author designed and built the measurement system. The author also was been responsible for the collecting of the measured data. Paper II Experimental results from sea trials of an offshore wave energy system This is the first paper that was published on the results of the WEC L1. The information covers two and a half month of data from the WEC operation in a range of sea states and with various loads connected to the generator. The paper was the first to verify the studied concept for wave energy conversion. It showed energy absorption dependency on load and demonstrated voltage and power output. The results indicate that optimal load does not vary with wave climate. The author was a member of the design team that carried out the experiment. The author was involved in the design discussions regarding the WEC and the assembling the WEC. The author played one of the leading roles in building the onshore infrastructure for securing the measurement data that resulted in this (pioneering) paper. Paper III Wave power absorption: Experiments in open sea and simulation In this paper, experimentally collected data of energy absorption for different electrical loads were used to verify a model of the wave power plant that includes the interactions of wave, buoy, generator, and the external load circuit. The wave-buoy interaction was modelled with linear potential wave theory. The generator was modelled as a non-linear mechanical damping function that is dependent on piston velocity and electric load. The results showed good agreement between experiments and simulations and potential wave theory was found to be well suited for the modelling of this point absorber in normal operation as well as for the design of future converters. The author was a member of the design team that carried out the experiment and had extensive reasonability for the collection of experimental data.

80 Paper IV Wave energy from the North Sea: experiences from the Lysekil research site This paper was the first paper that provided a detailed and thorough overview of the Lysekil research site. One of the objectives of the paper was to present the different research areas in the group within the same paper to give an overall picture of the project. The initial studies of wave height at the research site started 2004. Almost a full year of sea states, as measured by the wave measurement buoy, is presented with the conclusion that the average energy flux during 2007, excluding August, was 3.4 kW/m. The initial force measurement experiment is described. The installation of the WEC in the spring of 2006 is illustrated, and the paper includes a discussion on environmental studies at the site and some results on the artificial reefing effects of the installed WEC. Time synchronised data from measurements inside the buoy (vertical buoy speed and line force) are presented together with voltages and power from the WEC connected to a resistive load. The vertical buoy speed is calculated from measured acceleration. Furthermore, power and voltage when the WEC is connected to a non-linear load is presented. The author contributed to the written material: the text about the measuring station, the text about the WEC output and forces, the time synchronization of the measurement data, the text about the force measurement and being involved in the discussion section and the conclusions. The vertical buoy speed was essential to synchronise the measured force with the output from the generator, the author did this calculations. The author also contributed to the article with pictures and participated in the work with the experimental set-up at the Lysekil research site. Paper V Catch the Wave to Electricity - The Conversion of Wave Motions to Electricity Using a Grid-Oriented Approach The paper describes the challenges in wave power research and the different verification stages in the process that resulted in conversion of wave motion to electric energy. The author played one of the leading roles in the second stage of the verification process, the measurements of the forces in the line between a buoy with a diameter of 3 m attached to a set of powerful springs located on a concrete foundation on the ocean floor. The author was also a member of the team that carried out the successful experiment that converted wave motion to electricity.

81 Paper VI Description of the control and measurement system used in the Low Voltage Marine Substation at the Lysekil research site This paper describes the design of the control and measurement system used in the substation. The control system was developed on The National In- strument CompactRIO platform consisting of a real-time controller with real-time operating software, one FPGA chip and different input and output modules. Three CompactRIO systems were placed inside the substation. Each system is dedicated to one task. One system controls the contactors in the substation and disconnects the WECs if there is a faulty current. The next system logs: voltage, current and sensor data. The last system controls the switching of the IGBT that convert a DC voltage to alternating voltage. The result shows how the control system can manage to control the inverter in a laboratory experiment. The author is the main author of the paper and has written the paper and constructed the signal conditioning hardware and written all application software code. The paper was presented orally by the author at the 8th Euro- pean Wave and Tidal Energy Conference, EWTEC, 7–10 September, Upp- sala, Sweden, 2009. Paper VII Study of a Wave Energy Converter Connected to a Nonlinear Load This paper focuses on how the generator operates, i.e. how it moves and produces power when it is connected to a rectifier and a filter. The results show that a smooth power can be achieved with only one direct driven linear generator. The absorbed power will increase when the resistive value of the load is decreased. (The paper has been on the IEEE Journal of Oceanic En- gineering “top ten accessed articles”-list from June to October 2009 (except for August) and in January 2010.) Besides assisting in the expansion of the equipment in the measuring cabin to perform the nonlinear loading of the generator, the author has contributed to the article by a thorough review of the text. Paper VIII Laboratory experimental verification of a marine substation In this paper, the initial testing of the interconnection of several PMSGs on a common DC-bus is made by means of an offshore underwater substation. The paper presents laboratory measurements from the interconnection and the results are compared with simulations in MATLAB Simulink. The electrical system of the substation in explained and the way the interconnection will work in offshore operation is discussed. It is emphasized that the DC- bus voltage level will determine at which LG translator speed the buoy motion of the WEC will start to be damped..

82 The author was one of three people that had a leading role in the design and implementation of the tested substation. This paper was presented by the author at the 8th European wave and tidal energy conference, EWTEC, 7-10 September, 2009. Paper IX Design proposal of electrical system for linear generator wave power plants Paper IX presents two different marine substation designs, one built and developed by Uppsala University and one built and developed by Seabased Industry AB. Power data from L1 are presented during a three month period and the difference between a linear load and a non-linear load is discussed. The author worked with the design and implementation of one of the described substations. Paper X Determining the service life of a steel wire under a working load in the Wave Energy Converter (WEC) The service life of the steel wire appears to be a very important characteristic for the WEC. In order to prevent metal-to-metal contact between the steel wire and a funnel, the steel wire was impregnated in a black high density (HD) jacketing compound. The paper describes the result from a full-scale experiment of the dynamic behaviour of the steel wire under a working load in the WEC. The author has contributed with the presented force, voltage and power data and has thoroughly discussed the data presented in the figures with Dr Savin. Paper XI Offshore underwater substation for wave energy converter arrays The paper takes an engineering approach and gives as such a description of most of the important aspects on the design, implementation, deployment, recovery and maintenance of the substation. It elaborates on various design considerations and explains the various systems in the substation in detail. Results from initial laboratory measurements using two PMSG are compared with simulations. Results from interconnection of two WECs in offshore operation in the Lysekil wave energy research site are shown and the difference between these and the laboratory measurements are discussed. The author was involved in building and testing the substation.

83 Paper XII Experimental Results From an Offshore Wave Energy Converter In this paper, the performance of the WEC is investigated at different nonlinear loads. The results show what impact the damping of the WEC has on the power production. The highest absorption reached was about 26% and the average absorption was 5–10%. The author contributed to the experimental set-up at the Lysekil Research site. Paper XIII Description of a Torus Shaped Buoy for Wave Energy Point Absorber The paper discusses the dynamics of the WEC system. Added mass for a sphere buoy and a torus buoy are presented and discussed. The author designed the three point force measuring system used on the torus buoy. Paper XIV Wave Buoy and Translator Motions—On-Site Measurements and Simulations In the process of improving the concept in the Lysekil project, it is important to be able to quantify and predict several different entities in the wave energy converters. This paper has an engineering focus, and its aim is to compare different ways of measuring buoy motion, as well as relating these measurements to simulations made in Simulink. Buoy motion was measured using a land based optical system, and a buoy based accelerometer system. The measurements were found to correlate well in the vertical direction. However, in the horizontal direction the difference was substantial. The main reason for this was that the buoy rotation about its axis of symmetry was not measured. In a first comparison, the simulations and the measurements showed good agreement. Further on translator position inside the generator was measured and compared with simulations and buoy motion. The author has been very much involved in the discussions with the first author and contributed with: text, data processing and a figure about the accelerometer measurement. The author has also contributed with the measurement of translator position. Paper XV Ocean wave energy absorption in response to wave period and amplitude – offshore experiments on a wave energy converter The ability of a wave energy converter to capture the energy of ocean waves has been studied in offshore experiments. This study covers 50 days during which the converter was subjected to ocean waves over a wide range of frequencies and amplitudes as well as three different electrical loads.

84 The results present the wave energy converter’s energy absorption as a function of significant wave height, energy period and electrical load. It is shown that the power generated overall continues to increase with wave amplitude, whereas the relative absorption decreases towards the highest periods and amplitudes. The absorption reached a maximum of approxi- mately 24% with the used combination of buoy, generator and electrical load. Absorption to cover for iron and mechanical losses has not been included. A brief study of the nature of the electromagnetic damping force has also been included in the study. The wave energy converter is of the technology that is being researched at Uppsala University and experimented on off the Swedish west coast at the Lysekil wave energy research site. Published in IET Renewable Power Generation Received on 15th July 2010 The author has contributed to the experimental set-up. Paper XVI Lysekil Research Site, Sweden: A Status Update The paper presents a status update of the Lysekil project since 2008. The paper includes also important results and areas of further research. It includes a short description of 11 WECs, of the floating buoys connected to the WECs, the different connection lines as well as information about the marine substation, the observation tower and the measurement results. The author was deeply involved in the development and construction of the sensor system mounted inside WEC L2 and WEC L3 as well as in the development and construction of the measurement and control system placed inside the substation. The author was also main responsible for collecting and securing data from the research site. Paper XVII Sensors and Measurements inside the Second and Third Wave Energy Converter at the Lysekil Research Site The paper presents an overview of the sensor system and interactions of the measurements in the WEC L2 and L3. The WECs were instrumented with sensors for the measurement of the translator position, magnetic flux, temperature, rod piston vibrations and structure strain. Signals from the sensors were converted to electrical signals, amplified inside the WECs and sent via a cable to the substation. In the substation, the analogue signals were digitized. A correlation between different signals was found. The author was the main author of the paper and the poster. The paper was presented in a poster session at the Ninth European Wave and Tidal Energy Conference EWTEC, 2011.

85 Paper XVIII Temperature measurements in a linear generator and marine substation for wave power The temperatures in the WEC and in the marine substation are studied in the paper. Two different experiments are presented. In the first experiment, the temperature was measured in the WEC during 34 hours. The sea state was quite good during the experiment, 15 kW/m. In the second experiment the temperatures were also measured in the substation during 158 minutes. No drastic temperature increase was found in the two experiments. The author has implemented the measurement system and the sensors that have med used. The author has also written the part about the calibration. Paper XIX A Set-Up of 7 Laser Triangulation Sensors and a Draw-Wire Sensor for Measuring Relative Displacement of a Piston Rod Mechanical Lead-Through Transmission in an Offshore Wave Energy Converter on the Ocean Floor The wave energy converter in contains a piston rod mechanical lead-through transmission for transmitting the absorbed mechanical wave energy through the generator capsule wall while preventing seawater from entering the capsule. A set-up of 7 laser triangulation sensors has been installed inside the WEC to measure relative displacement of the piston rod and its corresponding seal housing. A draw-wire sensor has also been set up to measure translator position and the axial displacement of the piston rod. The paper gives a brief introduction to the Lysekil research site, the WEC concept, and the direct drive of WEC prototype L2. A model of operation for the piston rod mechanical lead-through transmission is given. The paper presents sensor choice, configuration, adaptation, mounting, and measurement system calibration along with a description of the data acquisition system. Results from 60 s measurements of nominal operation two months apart with centered moving averages are presented. Uncertainty and error estimations with sta- tistical analyses and signal-to-noise ratios are presented. Conclusions are drawn on the relative motions of the piston rod and the seal housing under normal operating conditions, and an assessment of the applicability of the measurement system is made. The author contributed with the measurement system; from the analog output of the laser sensor until the measurement data is secured on a hard disk drive. The author also contributed with the implementation of the translator position measuring system.

86 Paper XX Estimation of Stress in the Inner Framework Structure of a Single Heaving Buoy Wave Energy Converter This paper presents a method for strain measurements in the inner framework structure of the single heaving buoy Wave Energy Converter (WEC). The study focuses on estimation of stress in the inner framework structure of the WEC using strain measurements in material. Among the results the time histories of stresses and forces and bending moments. The objective is to extrapolate this study to other forms of Wave Energy Converters. This paper is a step towards the future design of wave energy devices in terms of material aspect, survivability in a hard wave climate and cost-effective renewable energies. The author played a leading role together with the first author of the paper in the installation of sensors for strain measurements on the inner framework structure of the WEC L2. One very important task performed by the author was the design of the low noise amplifiers used to amplify the signal from the strain sensors. A large and time-consuming effort were put into finding the right gain of each sensor before the analog signal was transferred to the substation. Paper XXI Azimuth-inclination angles and snatch load on a tight mooring System This paper presents a method for the measurement and evaluation of the normal force on the guiding system. The experimental data allow us to de- fine the inclination and azimuth angle between the generator and the floating buoy. The inclination angle is one of the key parameters for the design and the construction of the outer structure. The azimuth angle allows the evaluation of the direction of the loading on the guiding system. The inclination and azimuth angles allow the reproduction of the position of the floating buoy on the water surface. The author played a leading role together with the first author of the paper in the installation of sensors for strain measurements on the capsule of WEC L2. The main structure of the measurement method and how it could be implemented arose in discussions between the first author of the article and the author of this thesis. The author has also helped with line force data and time synchronisation between the line force data and the strain measurements. Paper XXII Experimental results from the operation of aggregated wave energy converters Wave energy comes in pulses and is unsuitable for direct conversion and transmission to the grid. One method to smooth the power is to deploy arrays of wave energy converters (WECs), the geometrical layout and damping

87 optimisation of which many have studied analytically and numerically, but very few by experiments at sea. In this study, the standard deviation of electrical power as function of various parameters is investigated. Two offshore experiments have been conducted. During the longer run, three WECs were operated in linear damping during 19.7 days. It is shown that the standard deviation reduces with the number of WECs in the array up to three WECs. The reduction compared to single WEC operation was found here to be 30 and 80% with two and three WECs, respectively, as a mean for an arbitrary array member. It is found that in sea states above 2 kW/m, the standard deviation is independent of sea state parameters. This is contradictory to a previous study on the same device. The results are, however, in accordance with numerical results of the SEAREV device but show larger reduction in standard deviation with number of WECs. This could be because of suboptimal damping conditions The author contributed to the design of the marine substation and the collection of data. The author also investigated the individual damping factors of the three WECs used in the experiment. Paper XXIII Peak Force Measurements on a Cylindrical Buoy with Limited Elastic Mooring This paper investigates the line force of a moored floating buoy. The experiment was the first experiment made at the Swedish wave energy research area located close to Lysekil on the Swedish west coast. The Lysekil project is run by the Swedish Centre for Renewable Electric Energy Conversion at Uppsala University. The experimental set-up consists of a cylindrical buoy with a diameter of 3 m and a height of 0.8 m. The buoy is moored with a line connected to a set of springs in parallel with a rope. The rope in parallel with the springs represents the limited stroke length of a linear generator type wave energy converter. The measurement system consists of a force transducer between the buoy and the rope, a three axis accelerometer inside the buoy and a data logger remote operated through the GSM network. The peak forces related to the significant wave height showed a trend of 33 kN/m. Trends were also calculated in 10 kN bins. The data could be used in fatigue simulations of similar devices. The result was very important to set the mechanical design parameters of the first wave energy converter that was deployed at the Lysekil Research Site in March 2006. The author is the main author of the paper and has written the paper and retrieved the data that it is based on.

88 Paper XXIV A Study of the Possible Power Extraction from a Point Absorbing Wave Energy Converter Today there exist several different ongoing wave power projects around the world. One of them is carried out by researchers at Uppsala University, Sweden. The concept is based on a point absorbing wave energy converter (WEC). A buoy is placed on the ocean surface and connected to a direct driven linear generator placed on the seabed. When the buoy moves with the waves, the moving part of the generator will move and a voltage is induced in the stator windings. The speed of the moving part will vary between zero and a peak value twice every wave period. The maximum speed and also the power depend on the wave height, the wave length and the damping of the generator. Moreover, the maximum speed of the real sea wave will be different for every wave period. Thus, the electrical system must be designed to handle the peak power that the generator generates even though the average power is significantly lower. This paper aims to investigate the difference between the peak power and mean power further. An analytical investigation of the possible maximum and mean power extraction of the WEC is performed. A comparison and discussion between the analytical results and results from experiments are presented. Moreover, the ongoing research activities at the university's research site are presented. The author is the main author of the paper. The author has written the text and produced the figures. The author also made the simulations and processed the experimental raw data.

89 16. Conclusions

This thesis describes how the experiments at the Lysekil Research Site for wave power have been carried out. Collected data from the experiments were used to verify the theoretical models simulated in previous work at the department. The theoretical and practical results indicate that wave power has the potential to be implemented as a sustainable energy source for the future. The key finding is that it has been shown possible to transform the motions of ocean waves into electrical energy and distribute it to solid ground. This process includes many conversion stages; each conversion must be carried out effectively and with durability, for the technology to be long time profitable. The author has been involved in developing technology to perform these steps. During my time as a graduate student, we have transferred energy to the on shore measuring station in two ways: The first way it was implemented in was by connecting the output of one generator directly to a cable whose other end was connected to the loads placed at the measuring station. The second implementation was by rectifying the AC voltage from each of three generators and connect all three to a common DC bus in a substation placed on the ocean floor close to the WECs. The aggregated energy is inverted to alternating voltages and transformed to a higher voltage before it is distributed to the on shore measuring station.

16.1. The control system I have had special responsibility for measurement and control system of the substation. The control system implemented in the substation was more than a simple system that controlles the on and off of the generators in the main circuit. It also included control of the transformation of voltage from the DC voltage of 50 Hz AC voltage. The computer used; a CompactRIO manufactured by National Instruments, forms together with transistors and filters an inverter. This system has become "de facto standard" at the department of Electricity when inverting voltages.

90 16.2. Measurements inside the buoy The first measurement arrangement was designed to measure the force in a rope between the buoy and a package of springs. The package of springs affects the buoy in the same manner as a WEC if it does not dissipate any power. This measurement circuit consists of a force transducer, a signal amplifier, a data logger and a GSM modem. In addition to the force in the line, buoy acceleration was measured. The basic design of this measurement set- up is reused in full-scale WEC experiments. The main reason that it is problematic to convert the energy in ocean waves into other energy forms is that energy is randomly dispersed in time. This has the consequence that the average power of the waves is much smaller than the maximum power.

16.2.1. Force measurements The results of test of force shows the variations between peak and average force. The measured force data from the first experiment was divided in half- hour bins. The mean and the maximum measured force was calculated within these bins. The maximum measured force in the line, when the buoy motion is limited by a stiff stopper rope is ten times the average force in that particular sea state, if compensation is performed for the static lifting force of the buoy. The frequency of the peaks, as well as the strength of each peak are investigated. At half-meter-high waves, it is a peak roughly every three seconds. When the wave height increases, the number of peaks increases. If only the wave lengt was concidered the peaks shuld become fewer as the wavelength increases. One reason that the number of peaks does not decrease is oscillations when reaching the top limit. Another is that the real water waves contains many different frequencies and peaks are detected even if it is not the highest peak whitin the wave.

16.2.2. Measurements of buoy movement One area that is studied and measured in detail is the buoy motion. An important early measurement result was the vertical acceleration of the buoy, this measurement gives, after integration, buoy velocity and position. When the voltage data from the generator is to be interpreted it helps a great deal if the direction of the buoy is presented. The translator position in two of the WECs have been measured, it gives almost the same information but there is a problem. Measuring a quantity on the inside of a wave power plant is a rather expensive and difficult operation, because the signal must be transmitted to the shore or to an external mast for radio transmission. An accelerometer measurement inside the buoy is implemented relatively

91 simple. It is also durable because all the parts: accelerometer, logger, antenna, battery, do not move relative to each other. This has however been a problem when the line force was measured, because the force transducer is constantly changing its angle towards the buoy. The sensor cable is repeatedly flexing and if it is not properly designed it will wear out and break. On the inside of one of the generator housing strain gauges are mounted. Together with the measured force in the rope we were able to calculate the buoy position in the horizontal plane, which is not possible with a simple accelerometer measurement. Accelerometer measurements in the horizontal plane are much more influenced by buoy motions in other directions than the measured direction; than vertical measurements are. Angular displacement of the buoy affects the influence of Earth's gravity on the measurement. Buoy rotation affects the measured value picked up by the sensor.

16.3. Relation between peak energy and average energy In the latest article, the average power from a WEC is compared with the maximum power. For small waves the ratio is large, up to 50 times. This is not very important as energy is small. What has much greater significance is that the peak power is up to 20 times greater than the average power in a wave height about three meters. A simulation in the same article examines how the alternator stroke and generator inductance affects the relationship between the maximum output power and average power. When the waves are higher than the stroke length of the translator the ratio increases because the energy is delivered in a shorter amount of time. When the inductance increases, the ratio is lowered because the WEC delivers less power at high frequencies because of the change in impedance.

92 17. Svensk sammanfattning

Den här avhandlingen berättar om hur experimenten vid Lysekils forsk- ningsområde för vågkraft har utförts. Insamlade mätdata har använts för att verifiera teoretiska samband som modulerats vid Elektricitetslära, Uppsala universitet. De teoretiska och praktiska resultaten har visat på att vågkraft har förutsättningarna att implementeras som en hållbar energikälla för fram- tiden.

17.1. Varierande energikälla Vi har länge känt till att det finns mycket energi i havsvågor. Under de se- naste århundradena har många kommit med idéer om hur vågenergi ska tyg- las, men endast mycket små energimängder har kunnat omvandlas till att utföra ett arbete. Den första som lyckades omvandla vågenergi till användbar energi var Yoshio Masuda som 1965 kunde utvinna energi från vågorna och använda denna energi för att driva en navigationsboj. Den största anledningen till att det är problematiskt att omvandla energin i havsvågor till en annan energiform, är att energin är slumpmässigt utspridd i tiden. Detta får konsekvensen att medeleffekten från vågorna är mycket mindre än den högsta effekten. Det är så inom varje våg, eftersom varje typ av vågenergi-omvandlare som hittills konstruerats endast omvandlar rörelse energi eller läges energi i vågen. Det är också så inom en tidsperiod av ett antal vågor att medeleffekten är lägre än effekten i den största vågen, eftersom varje våg har olika höjd och längd. När tidsspannet som analyseras ut- ökas till att gälla månader och år skiljer sig medelenergin även mellan olika timmar på grund av att vinden som är upphovet till vågor, blåser från land eller mot land och ibland inte alls.

17.2. Mitt arbete Min tid på Elektricitetslära inleddes med att bygga ett experiment för att mäta kraften i vågorna vid experimentområdet utanför Lysekil. Jag var med om byggandet av det första vågkraftverket som testades. Jag har haft ett speciellt ansvar för mätstugan som tar emot energin från testerna som utförs ute till havs. Jag har varit med i det team som konstruerat det första ställverket

93 som placerats under vattenytan med uppgift att sammanlagra energi från flera vågkraftverk. I detta projekt har jag haft speciellt ansvar för mät och styrsystemet av anläggningen. De generatorer som testades tillsammans med ställverket var försedda med många olika givare. Alla dessa givare och mät- data från dem var jag på något sätt inblandad i.

17.3. Mätningar med sensorer och datalagring inuti bojen Den första mätuppställningen som byggdes, var avsedd att mäta kraften i linan mellan bojen och ett fjäderpaket fastsatt på havsbotten. Fjäderpaketet påverkar bojen på liknande sätt som en våg-energi-omvandlare som inte levererar någon effekt. Denna uppställning består av en kraftgivare, en mät- förstärkare, en datalogger och ett GSM modem. Förutom kraften i linan, mättes bojens acceleration. Grunddesignen av detta mätsystem användes i de följande experimenten. Mätningarna på den ringformade bojen modifierades genom att tre stycken kraftgivare användes istället för en .

17.4. Mätstugan Mätstugan på Härmanö var det andra stora projektet. Mätstugan byggdes föra att kunna göra olika tester på de generatorer som testas på området. Den innehåller flera olika lastmöjligheter och möjligheter att mäta spänning. Det- ta kan numera styras från en annan plats via datakommunikation.

17.5. Ställverket och mätningarna i generatorerna Det största projektet under min tid som doktorand var konstruktionen av ett ställverk som placerades på havsbotten. Ställverksprojektet bestod at tre stora delar där jag hade ansvaret för kontrollsystemet; Magnus hade ansvaret för inneslutningen och Cecilia för det elektriska systemet. Det kontrollsy- stem som implementerades i ställverket var inte bara ett system som kontrol- lerade påslag och avslag av generatorerna i huvudkretsen. Det innehöll också styrning av omformningen av spänningen från DC spänning till 50 Hz AC spänning. Den dator som användes var en CompactRIO tillverkad av Natio- nal Instruments, bildar tillsammans med transistorer och filter en växelrikta- re. Detta system har blivit ”de facto standard” vid avdelningen för Electrici- tetslära när omformning av spänning ska ske. I två av vågkraftverken placerades sensorer för translator position: det magnetiska flödet genom statortänderna, temperaturen på statorsegmenten

94 och i luften samt indikation på vattenläckage. I det ena av dessa monterades även givare som mätte töjningen i strukturen samt sensorer som mätte rörel- sen av kolvstången.

17.6. Resultat Det viktigaste resultatet är att vi har visat att det går att omvandla havsvågors rörelse till elektrisk energi som i sin tur kan överföras till fast mark. Denna omvandling innehåller många olika steg där varje steg måste utföras effek- tivt och utan att omvandlingsapparaten går sönder. Först ska rörelsen i vå- gorna påverka bojen på ett sådant sätt att bojens rörelse överförs till translatorn. Sedan ska translatorns rörelse omvandlas till elektricitet. Den elektriska energi som skapats måste transporteras till land där den kan användas av elkonsumenter. Under min tid som doktorand har vi överfört energi till land på två sätt: dels genom att koppla utgången på generatorn som står på havsbotten direkt till en kabel vars andra ände varit ansluten till laster, dels genom att likrikta växelspänningen från tre generatorer och sammanlagrat energin i en stor gemensam kondensator bank. Denna energi omformas sedan till växelspän- ning och transformeras upp till en högre spänning innan överföringen till land.

17.6.1. Bojens rörelse Ett område som jag studerat och mätt mer ingående är bojens rörelse. Bojens rörelse har mätts på flera olika sätt. En viktig tidig mätning som gjorts har varit den vertikala accelerationen av bojen, denna mätning ger efter integration bojens hastighet och position. När spänningsdata från generatorn ska tolkas underlättar det mycket om rörelseriktningen av bojen är visualiserad. I vågkraftverk två och tre mättes translatorns position. Att mäta inuti ett vågkraftverk är en ganska dyr och besvärlig operation eftersom signalen måste överföras till land eller till en extern mast för radioöverföring. En accelerometer-mätning inuti bojen implementeras däremot förhållandevis en- kelt. Den är dessutom hållbar eftersom alla delar: accelerometer logger, an- tenn, batteri, inte rör sig i förhållande till varandra. Detta har däremot varit ett problem när kraften i linan mätts. Kraftgivaren ska mäta kraften i linan, därför ändrar den hela tiden sin vinkel gentemot bojen. Detta resultera i att kabeln mellan givaren och bojen går av på grund av de konstanta vickningarna. På insidan av en av generatorns hölje placerades även töjningsgivare. Tillsammans med den mätta kraften i linan har vi kunnat beräkna bojens position i horisontalplanet, vilket inte är möjligt med en enkel accelerome- termätning. Accelerometer mätningar i horisontalplanet påverkas mycket

95 mer av bojens rörelser; vinkel förändring av bojen påverkar influensen av jordens gravitation och bojens rotation påverkar själva mätningen av accelerationen.

17.6.2. Kraften i linan Resultaten av kraftmätningen visar verkligen på variationer. Kraftdata från det första experimentet indelat i halvtimmes perioder har databehandlats. Medelvärdet samt den högsta uppmätta kraften har presenterats. Den maximala kraften som mäts i linan under en halvtimme är ca 10 gånger större än medelvärdet av kraften. Hur ofta det är en uppkommer en peak och hur star- ka de är har också undersökts. Vid halvmeter höga vågor är det en peak un- gefär var tredje sekund. När våghöjden ökar blir peakarna fler. Man skulle kunna tänka sig att de blev färre eftersom våglängden ökar med ökad våg- höjd. En anledning till detta är att translatorn studsar i det övre ändstoppet varje gång den slår i den. En annan är att riktiga vattenvågor innehåller många olika frekvenser.

17.6.3. Medelenergin i förhållande till den maximala energin från vågkraftverket I den sista artikeln har medeleffekten från ett vågkraftverk studerats och jämförts med den högsta effekten. Vid små vågor är förhållandet stort, upp till 50 gånger. Detta är inte så viktigt eftersom energin är liten. Det som har mycket större betydelse är att effekten ut från vågkraftverket är upp till 20 gånger större än medeleffekten vid en våghöjd kring tre meter. Simuleringar i samma artikel undersöker hur generatorns slaglängd och generatorns induktans påverkar relationen mellan den högsta uteffekten och medeleffekten. När vågorna är så höga att translatorn begränsas av generatorns längd ökar förhållandet eftersom energi levereras kortare del av tiden. När induktansen ökar minskar förhållandet eftersom lägre effekt kan lämna generatorn vid höga frekvenser.

96 18. Acknowledgements

Först vill jag tacka min handledare Mats Leijon, det har varit en ynnest att få vara med på resan att utveckla vågenergi; från det att nästan ingen trodde det var möjligt att tygla resursen, till idag när det ses som ett viktigt bidrag i den framtida energiförsörjningen. Alla våra finansiärer förtjänar ett stort tack. Min lön har under doktorand- tiden betalats via Vågkraft projektet i Lysekil. De största finansiärerna till projektet är: Energimyndigheten, Vetenskapsrådet, Swedich Research Coun- cil grant no. 621-2009-3417, Vattenfall AB och Fortum. Därtill följer en rad finansiärer Göteborg energi, Ångpanneföreningen, Sweco, Statkraft, Fal- kenberg energi, Wallenius Stena, Beijer samt Uppsala universitet. Stiftelsen J. Gust. Richert har bidragit med ett resestipendium. Tack till alla i våggruppen framförallt Cecilia och Magnus, vi som kon- struerade ställverket. Till alla studenter som hjälpte till med det praktiska arbetet. Tobias som jobbade dedikerat med att bygga elektronik (till och med på julafton) och körde bilen till mätstugan på Härmanö (där han också hitta- de ett viktigt fel). Hjalmar och Lisa som byggde stora delar av ställverket och bidrog med många bra konstruktionslösningar. Rickard som hjälpte till med ställverket en natt i Lysekil och som sedan tog över och nu utvecklar en ny version av kontrollsystemet och ställverket. Rumskamraterna i utflyttande ordning; Remya för att du vidgat mitt inter- nationella perspektiv och alltid är en intressant diskussionspartner. Diskus- sionerna med Simon som har givit ett större djup i analyserna, tyvärr spelade vi alldeles för lite gitarr. Magnus och Rafael och Stefan som var med om att bygga det första vågkraftverket, oj vad vi har diskuterat och vänt och vridit på alla detaljer. Stefan för den första tiden i Lysekil. Jag vill också tacka Jan Sundberg för allt stöd nere i Lysekil speciellt för att du lärde en tvättäkta landkrabba att köra båt så att han nu rör sig obehind- rat i farvattnen mellan Lysekil och Härmanö. Lars Tegnér, Valeria Castellucci och Marie Johansson Little som har granskat hela manuset (i olika stadier), samt alla andra kamrater som granskat enskilda kapitel. Alla på avdelningen: Thomas och Mikael som fixar alla datorproblem. Ulf, Gunnel, Maria, Elin och Christina. Till slut vill jag tacka mamma och mina syskon.

98 19. Bibliography

[1] S. H. Salter, Wave power, Nature Volume 249, p.720.S, 1974. [2] D. Bernoulli, Hydrodynamica, 1738. [3] F. J. Gerstner, 1802. Theorie der Wellen. Abhand. Kön. Böhmischen Ge- sel. Wiss., Prague. Also in Weber & Weber, 1825. [4] M. Faraday, On a peculiar class of acoustical figures; and on certainforms assumed by groups of particles upon vibrating elastic surfaces, Philos. Trans. R. Soc. London 121, pp 299-340, 1831. [5] G. Stokes, Report on recent researches in hydrodynamics. Rep. 16 th Brit. Assoc. Adv. Sci., pp. 1–20, 1846. [6] H. Lamb, Hydrodynamics. 1th Edition. Cambridge, University press, 1895. [7] D. V. Evans, A theory for wave-power absorption by oscillating bodies. J: Fluid Mech Volume 77, pp 1-25.S, 1976. [8] Y. Masuda, Wave-activated generator. Int. Colloq Exposition Oceans. France: Bordeaux; 1971. [9] A. F. O. Falcão, Wave energy utilization: A review of the technologies. Renewable and Sustainable Energy Reviews, 14, pp. 899-918, 2010. [10] K. Budal, J. Falnes, A resonant point absorber of ocean wave power Na- ture Volume 256, p 478, 1975. [11] M. Faraday, Experimental Researches in Electricity, Vol I, II & III 1837, 1844, 1855. [12] I. Ivanova, Simulation of Linear Permanent Magnet Octagonal Generator for Sea Wave Energy Conversion , Licentiate thesis, Uppsala 2004. [13] K. Thorburn, Electric Energy Conversion Systems: Wave Energy and Hy- dropower, Doctoral thesis, Uppsala University, Sweden, 2006. [14] O. Danielsson, Wave Energy Conversion, Linear Synchronous Permanent Magnet Generator, Doctoral thesis, Uppsala University, Sweden, 2006. [15] M. Eriksson, Modelling and Experimental Verification of Direct Drive Wave Energy Conversion. Buoy-Generator Dynamics, Ph. D Thesis, Upp- sala University, Uppsala, Sweden, 2007. [16] R. Waters, Energy from Ocean Waves. Full Scale Experimental Verifica- tion of a Wave Energy Converter, Ph. D Thesis, Uppsala University, Upp- sala, Sweden, 2008. [17] O. Langhamer, Wave energy conversion and the marine environment: Colonization patterns and habitat dynamics, Ph. D Thesis, Uppsala Uni- versity, Uppsala, Sweden, 2008.

99 [18] M. Rahm, Ocean Wave Energy: Underwater Substation System for Wave Energy Converters, Ph. D Thesis, Uppsala University, Uppsala, Sweden, 2010. [19] R. Krishna, Multilevel Inverter for Wave Power Conversion, Licentiate thesis, Uppsala 2011. [20] H. Gravråkmo, Buoy for linear wave energy converter, Licentiate thesis, Uppsala 2011. [21] C. Boström, Electrical systems for wave energy conversion, Ph. D Thesis, Uppsala University, Uppsala, Sweden, 2011. [22] J. Engström, Hydrodynamic Modelling for a Point Absorbing Wave En- ergy Converter, Ph. D Thesis, Uppsala University, Uppsala, Sweden, 2011. [23] S. Lindroth, Buoy and Generator Interaction with Ocean Waves, Ph. D Thesis, Uppsala University, Uppsala, Sweden, 2011. [24] B. Ekergård, Electromagnetic Energy Converters - Rotating Motors and Linear Generators, Licentiate thesis, Uppsala 2011. [25] A. Savin, Experimental measurement of lateral force in a submerged single heaving buoy wave energy converter, Ph. D Thesis, Uppsala Univer- sity, Uppsala, Sweden, 2012. [26] K. Haikonen, Environmental Impact from Wave Energy Converters - Un- derwater Noise, Licentiate thesis, Uppsala 2012. [27] M. P. Tulin, On the Transport of Energy in Water Waves. Journal off engineering mathematics, Volume 58, pp. 339-350, 2007. [28] K. Michel, Ochi., Ocean waves :The Stochastic Approach. Cambridge University Press 2005. [29] K. Thorburn, H. Bernhoff, and M. Leijon, Wave energy transmission system concepts for linear generator arrays, Ocean Engineering, 31(11-12), pp 1339 – 1349, 2004. [30] O. Danielsson, K. Thorburn, E. Sjöstedt, and M. Leijon, Simulated response of a linear generator wave energy converter, ISOPE-2004, Toulon, France, 2004. [31] M. Eriksson, K. Thorburn, H. Bernhoff, and M. Leijon, Dynamics of a linear generator for wave energy conversion, 23rd International Conference on Offshore Mechanics and Arctic Engineering, Vancouver, Canada, 2004. [32] M. Leijon, H. Bernhoff, O. Ågren, J. Isberg, J. Sundberg, M. Berg, K. E. Karlsson, A. Wolfbrandt, Multiphysics Simulation of Wave Energy to Electric Energy Conversion by Permanent Magnet Linear Generator, IEEE Transactions on Energy Conversion, Volume: 20 , Issue: 1, pp. 219 – 224, 2005. [33] I. A. Ivanova, H. Bernhoff, O. Ågren and M. Leijon, Simulated generator for wave energy extraction in deep water, Ocean Engineering, Volume 32, Issues 14-15, pp. 1664-1678, 2005. [34] I. A. Ivanova, O. Agren, H. Bernhoff, M. Leijon, Simulation of Wave- Energy Converter With Octagonal Linear Generator, IEEE Journal of Oceanic Engineering Volume 30, Issue 3, Page(s):619 – 629, 2005.

100 [35] M. Eriksson, J. Isberg, and M. Leijon, Hydrodynamic modelling of a direct drive wave energy converter, International Journal of Engineering Science, Volume 43, Issues 17-18, pp. 1377-1387, 2005. [36] O. Danielsson, M. Leijon, K. Thorburn, M. Eriksson, H. Bernhoff, A Di- rect Drive Wave Energy Converter – Simulations and Experiments" Pro- ceedings of OMAE 2005: 24th International Conference on Mechanics and Arctic Engineering, Halkidiki, Greece, 2005. [37] M. Leijon, O. Danielsson, M. Eriksson, K. Thorburn, H. Bernhoff, J. Is- berg, J. Sundberg, I. Ivanova, E. Sjöstedt, O. Ågren, "An electrical approach to wave energy conversion", Renewable Energy, Volume 31, Issue 9, Pages 1309-1319, 2006. [38] M. Eriksson, J. Isberg, M. Leijon, Theory and Experiment on an Elastically Moored Cylindrical Buoy, IEEE Journal of Oceanic Engineering, Volume 31, Issue 4Page(s):959 – 963, 2006. [39] H. Lamb, Hydrodynamics. 1th Edition. Cambridge, University press, p. 8, 1895. [40] S. Eriksson, Direct Driven Generators for Vertical Axis Wind Turbines, Ph. D Thesis, Uppsala University, Uppsala, Sweden, 2008 [41] K. Thomas, Low Speed Energy Conversion from Marine Currents, Ph. D Thesis, Uppsala University, Uppsala, Sweden, 2008. [42] B. Bolund, Electric Power Generation and Storage Using a High Voltage Approach, Ph. D Thesis, Uppsala University, Uppsala, Sweden, 2006. [43] E. Segergren, Direct Drive Generator for Reneable Power Conversion from Water Currents, Ph. D Thesis, Uppsala University, Uppsala, Sweden, 2005. [44] M. Ranlöf, Electromagnetic Analysis of Hydroelectric Generators, Ph. D Thesis, Uppsala University, Uppsala, Sweden, 2011. [45] J. Santiago, FEM analysis applied to electric machines for electric vehicles Ph. D Thesis, Uppsala University, Uppsala, Sweden, 2011. [46] J. Falnes, Optimum control of oscillation of wave-energy converters, Pro- ceedings of the Eleventh (2001) International Offshore and Polar Engi- neering Conference, Stavanger, Norway, June 17-22, 2001. [47] W. M. Leach, Fundamentals of low-noise analog circuit design, Proceed- ings of the IEEE Volume: 82 , Issue: 10. Page(s): 1515 – 1538, 1994. [48] S. Tyrberg, H. Gravråkmo and M. Leijon, "Tracking a Wave Power Buoy Using a Network Camera: System Analysis and First Results", Proceed- ings of the 28th International Conference on Ocean, Offshore and Arctic Engineering (OMAE 2009), Honolulu, Hawaii, 2009.

101 Acta Universitatis Upsaliensis Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology 957 Editor: The Dean of the Faculty of Science and Technology

A doctoral dissertation from the Faculty of Science and Technology, Uppsala University, is usually a summary of a number of papers. A few copies of the complete dissertation are kept at major Swedish research libraries, while the summary alone is distributed internationally through the series Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology.

ACTA UNIVERSITATIS UPSALIENSIS Distribution: publications.uu.se UPPSALA urn:nbn:se:uu:diva-179098 2012