Home

MSc dissertation - School of Computing

1. Type of terrain Roughness length Zo m Cities forests 0 7 Suburbs wooded countryside 0 3 Villages countryside with trees and hedges 0 1 Open farmland few trees and buildings 0 03 Flat grassy plains 0 01 Flat desert rough sea 0 001 Table 2 Roughness length values Burton et al 2001 p 19 2 2 6 Electrical systems It is clear from Figure 7 that the main electrical element of interest is the generator however the type of generator and the requirement for other electrical components depends on the detailed turbine design Firstly it is important for later discussions to note that an electric generator is simply an electromechanical machine EM used in generating mode Gross 2007 p 79 The same physical device can be used either as a generator where a kinetic input is converted to an electrical output or as a motor where an electrical input passes through the same circuitry to produce a kinetic output Although other energy conversion applications typically rely on synchronous permanent magnet EMs the natural operation of a wind turbine introduces periodic fluctuations into its output which degrade the power quality Using an induction generator on the other hand introduces a damping action which reduces the unwanted fluctuations to acceptable levels Burton et al 2001 p 364 Whereas with a permanent magnet EM the rotor is locked to the rotational speed of the magnetic field in the stator the ro
2. DC bus connection OD D Broking resistor connection Control power supply and interfaces between power and control Amesso ries i Expansion Slot 1 white Encoder Interface Slat 2 yellows P D1 io D4 Analog e Inputs SC amp COMM 2 AIT and A anyus Slot 6 Digital Outputs DO RLI o DOs PLS 101 Appendix K System specification 1 Introduction The information in this document constitutes a starting point for the design of a control system for the Gaia Wind turbine emulator The development approach will be based on agile development principles and therefore further design details will emerge as the project progresses 1 1 Overview The system design is based on the central concept of a Test object which is made up of a sequence of nstructions The main components of the system are shown in the block diagram below Inverter communications Test module management module User interface Controller communications module External files scripts and logs The system will be implemented in C which means that the main system components will be represented by a set of C classes The user interface will be built from standard interface controls provided by Visual Studio and the remaining classes will be designed independently based on the UML analyses below Communications with the hardware components have not been fully specified at this time because there are several opt
3. IGBT inverter Generator SCIM model D Figure 19 Block diagram for WTE torque control 2 3 4 Input data The final element required for a WTE is a source of wind speed data The three main options here are e Directly controlled wind speeds e Artificially generated time sequence e Logged time sequence The choice of input type is determined by the type of test being conducted Many research teams use the first option because it facilitates the observation of system behaviour under controlled conditions Bouscayrol et al 2005 Hsu Wen Ko 2010 Kojabadi et al 2004 Lopes et al 2003 Mansouri et al 2003 Munteanu et al 2010b de Oliveira et al 2007 Seman et al 2006 Teodorescu et al 2003 Weihao Hu et al 2008 Weiwei Li et al 2007 Direct control over wind speed 42 allows for steady state observations step changes in wind speed and steadily ramped speed changes Where testing is focussed on aspects of turbine operation with a strong stochastic element such as fault diagnosis or the characterisation of effects due to turbulence time series data is preferred In these situations wind models such as those described by Diop et al 2007 can be used to generate synthetic time series or real logged data can be used Chinchilla et al 2004 Examples of both approaches are found with Crabtree 2011 preferring synthetic data and Dolan et al 2005 opting for logged data Monfared et
4. Offline P0683 Efficiency Slip Figure 29 HMI showing Auto tab and numerical feedback The numbered features in Figure 29 are described in Table 8 65 Label Description 1 The Auto tab allows the operator to specify a series of commands to the inverter 2 The toolbar allows the operator easy access to commands to run and stop a test for example and to navigate from one stored test to the next The screenshot in Figure 29 shows that there are two tests currently stored in the internal database 3 Output from the system is shown in one of three tabbed panels on the right Figure 29 shows the panel which reports numerical values from the inverter and the turbine When the WTE is active these values are updated automatically at a polling period defined by the sampling period specified by the operator 4 The operator can choose to log operational data to an external file in Microsoft Excel format Again the rate is determined by the sampling period set by the operator The access to Excel files in achieved using the C project ExcelPackage Tunnicliffe 2007 which is protected under the GNU Public Licence 5 The HMI has a large animated indicator to show the state of the WTE hardware to minimise the possibility of operator error This section of the display also shows summary calculations performed by the system The values of inverter parameters are displayed on the left of the feedback panel Tur
5. The variation in performance of the WTE can be assessed by collecting data over a range of speeds and creating a box plot of each of the logged quantities This will provide a performance envelope that defines the accuracy of the values currently displayed by the HMI For this test the generator speed is raised just above synchronous speed and held long enough for the turbine controller to close the grid connection Thereafter the inverter reference speed is incremented by 1 rpm from 1005 to 1020 rpm Each speed setting is maintained for 30s and the speed reading from the turbine controller is logged every second 53 4 Results This section summarises the output from the activities described in 3 4 4 1 Hardware configuration The sections below summarise the important characteristics of the individual components of the WTE and the overall construction 4 1 1 Overview The Gaia Wind turbine is a two bladed fixed pitch fixed speed unit which is designed according to the Danish model Its principal design parameters are summarised in Table 6 Where two values are given the first corresponds to the European version of the turbine which is designed to be connected to a 50Hz distribution network and the second refer to the American version designed for a 60Hz network Characteristic Value Rotor diameter 13m Rotor weight 200 kg Hub height 18 3 m Rated electrical power 11 kW Rated wind speed 9 5 m s
6. All current designs rely on more or less rigid rotors based on the principles of aerodynamics and usually have two or three rotor blades mounted on a horizontal axis Vertical axis designs are also in use but they are less popular due to their lower efficiencies and higher cost Gross 2007 p 104 The explanation for the design convergence in horizontal axis machines requires reference to some basic theoretical concepts As for any other energy conversion device it is important to have a means of expressing the efficiency of a wind turbine Fundamentally this is ratio of the actual power production of the machine to the total energy available and gives a dimensionless power coefficient Cp whose basic formula is given by Eq 1 By optimising the formula for Cp it can be shown that the maximum possible value is 0 593 known as the Betz limit Because the Betz limit is derived from theoretical principles rather than with reference to any particular turbine design Burton et al 2001 p 45 it provides a fixed upper bound for the efficiency of any turbine The actual value of the power coefficient is affected by the geometry of the rotor blades 14 and also varies with wind speed The interaction of these quantities can be captured with reference to another proportional characteristic known as the tip speed ratio A The tip speed ratio is defined as the tangential velocity of the spinning rotor divided by the velocity of the uninter
7. 56 0 dB A global Rated voltage 380 4004 15 660 690V ve Rated om 22 9 22 3 22 3132 129 A Tor Deag L R Amperes 460 156 156 92 3 90 5A 0 81 0 79 0 76 896 8956 8956 ivin 70 0 76 0 72 0 68 896 896 896 No load current 11 7 13 0 14 26 74 7 54 A kanata n Rated torque 108 108 108 Nm Bearing Quantity lubricant Locked rotor torque 210 240 260 Front 6309 C3 139 Breakdown torque 240 270 290 Rear 6209 Z C3 9g Design N Insulation class F Lubrication interval 20000 h Temperature nse BOK Grease Polyrex EM ESSO Locked rotor time 13 13 13 s hot Service factor 100 Duty cycle 1 Ambient temperature 20 C to 40 C Altitude 1000 ma s Notes Temperature detector type Klixon The figures given herewith are regarded as guaranteed values and applied to sinusoidal power supplied motors within permissible tolerances under IEC 60034 1 Noise level wih tolerance of 3 dB A Weight value can be changed without previous notification Effkiencies according to the indirect method of IEC 60034 2 1 2007 with stray load losses determined fom measurement Checked Revision Approved AUTOMATICO Nr Oo Dap 25 APR 2010 91 Appendix F Weg 60Hz SCIM datasheet Product code Product line Rated current L R Amperes Hsin No load current Rated torque Locked rotor torque Breakdown torque Design Insulation class Temperature nse Locked rotor time Service factor Duty cycle Ambient temperature
8. S Roye D Guiraud J 2010b Hardware in the Loop based Simulator for a Class of Variable speed Wind Energy Conversion Systems Design and Performance Assessment Energy Conversion IEEE Transactions on vol 25 no 2 pp 564 576 June 2010 doi 10 1109 TEC 2010 2042218 URL http ieeexplore ieee org stamp stamp jsp tp amp arnumber 5437239 amp isnumber 5467345 Munteanu l Bratcu A I Cutululis NA Ceanga E 2008 Optimal Control of Wind Energy Systems Towards a Global Approach Springer NREL 2010 Gaia Wind Avaliable online at http www nrel gov wind smallwind gaia_wind html Accessed 4 August 2010 Neammanee B Sirisumrannukul S Chatratana S 2007 Development of a Wind Turbine Simulator for Wind Generator Testing International Energy Journal Volume 8 Issue 1 Ofgem 2011 Feed in Tariff Adjusted tariff rates 1 April 2011 Available online at http Awww ofgem gov uk Pages Morelnformation aspx docid 16 amp refer Sustainabilit y Environment fits Accessed 23 July 2011 Ovando R I Aguayo J Cotorogea M 2007 Emulation of a Low Power Wind Turbine with a DC motor in Matlab Simulink Power Electronics Specialists Conference 2007 PESC 2007 IEEE vol no pp 859 864 17 21 June 2007 doi 10 1109 PESC 2007 4342101 126 URL http ieeexplore ieee org stamp stamp jsp tp amp arnumber 4342101 amp isnumber 4341940 Parekh R 2003 AC induction motor fundamentals Application note AN887 Microchip Tec
9. measured from top dead centre of the swept area 25 For downwind turbines the Powles 1983 model which is also used in GH BLADED provides a general solution The affected zone is the same as for the upwind model as shown in Figure 9 and the wind speed V at a point within the sector is given by Eq 6 For other blade azimuthal angles the same correction factor is made as for the upwind calculation V U Acos a where A is the maximum velocity deficit directly behind the tower as a fraction of U W is the width of the tower shadow as a proportion of the tower diameter d Bossanyi and Quarton 2003 Wind shear whereby wind speeds tend to increase with height above ground level constitutes a further source of torque oscillation in addition to tower shadow Burton et al 2001 p 233 The result of wind shear is that a rotor blade pointing upwards will experience a higher wind speed than one pointing downwards thus exacerbating the effect due to tower shadow The problem can be further compounded if the frequency of the resulting oscillation is close to the resonant frequency of either the blades of the tower Burton et al 2001 p 267 Wind shear can be modelled as a logarithmic function given by log VK zo h log OF z where V is wind speed at height h V is wind speed at a reference height ho Zo is the ground roughness length which takes values from Table 2 Bossanyi and Quarton 2003 Vh Vo 26
10. the PMSM also has the advantage that there are no associated field losses Gross 2007 however torque control is more difficult because its relationship with the three phase sinusoidal current is complex Using a mathematical operation known as Park s Transform the three AC phase currents 37 can be reduced to two DC components in a rotating frame of reference Lee et al 1984 In the rotating frame of reference a direct current component d rotates at a 90 phase angle from a quadrature component q The torque in the PMSM can be shown to be proportional to the quadrature current Pillay and Krishnan 1988 Once the required current in the q axis has been calculated the equivalent AC supply to the PMSM can be found by applying the inverse Park transformation This method for controlling a PMSM is known as field oriented control FOC and can be accomplished relatively easily using a microprocessor Gabriel et al 1980 A third group of WTE studies use a squirrel cage induction machine SCIM to provide the required torque usually citing the SCIM s rugged design and lower cost as benefits Teodorescu et al 2003 Kojabadi and Chang 2004 Ming Qiao et al 2007 Neammanee et al 2007 de Oliveira et al 2007 Yaoqin Jia et al 2007 The torque produced by a SCIM can be controlled using FOC in the same manner as for a PMSM As a consequence of choosing this type of drive a reliable variable voltage source is also required and
11. 1 03 0 08 0 93 1 52 0 08 0 29 2 01 0 08 0 12 2 5 0 08 0 07 3 0 10 0 04 3 49 0 11 0 03 55 55 5 82 125 65 1199 88 10 83 0 00 0 88 3 99 0 31 0 06 55 57 5 82 125 70 1200 38 9 48 0 01 2 47 4 49 1 15 0 16 55 60 5 82 125 76 1200 89 8 43 0 02 9 14 4 99 2 28 0 23 55 62 5 82 125 81 1201 39 7 59 0 03 18 12 5 49 3 67 0 27 55 64 5 83 125 86 1201 90 6 90 0 04 29 16 5 99 5 00 0 29 55 67 5 83 125 92 1202 40 6 33 0 05 39 71 6 49 6 27 0 28 55 69 5 83 125 97 1202 90 5 84 0 05 49 77 7 7 57 0 27 55 71 5 83 126 02 1203 41 5 42 0 05 60 07 7 49 8 70 0 25 55 74 5 84 126 07 1203 91 5 07 0 05 69 01 7 99 9 80 0 24 55 76 5 84 126 13 1204 42 4 75 0 05 77 70 8 49 10 77 0 22 55 78 5 84 126 18 1204 92 4 47 0 05 85 35 9 11 67 0 20 55 81 5 84 126 23 1205 42 4 22 0 05 92 45 9 5 12 36 0 18 55 83 5 85 126 28 1205 93 4 00 0 04 97 87 10 13 12 0 16 55 89 5 85 126 41 1207 13 3 80 0 04 103 79 10 49 13 69 0 15 55 94 5 86 126 54 1208 33 3 63 0 04 108 19 11 14 15 0 13 56 00 5 86 126 66 1209 54 3 47 0 04 111 71 11 49 14 59 0 12 56 05 5 87 126 79 1210 74 3 32 0 04 115 07 12 14 80 0 11 56 11 5 88 126 91 1211 95 3 18 0 03 116 61 12 49 14 90 0 09 56 16 5 88 127 04 1213 15 3 06 0 03 117 29 13 15 00 0 08 56 22 5 89 127 17 1214 35 2 94 0 03 117 96 13 49 14 93 0 07 56 28 5 89 127 29 1215 56 2 84 0 02 117 29 13 99 14 80 0 07 56 33 5 90 127 42 1216 76 2 74 0 03 116 15 14 48 14 60 0 06 56 39 5 90 127 54 12
12. 2011 11 17 09 1005 100 5 1003 65 436 65 100 18 08 2011 11 17 10 1005 100 5 1005 0 0 100 18 08 2011 11 17 11 1005 100 5 1005 65 436 65 100 18 08 2011 11 17 12 1005 100 5 1005 0 0 100 18 08 2011 11 17 13 1005 100 5 1003 0 0 100 18 08 2011 11 17 13 1005 100 5 1003 0 0 100 18 08 2011 11 17 15 1005 100 5 1003 0 1 0 100 18 08 2011 11 17 15 1005 100 5 1003 0 2 0 100 18 08 2011 11 17 16 1005 100 5 1003 0 3 0 100 18 08 2011 11 17 18 1005 100 5 1003 0 3 0 100 Table 10 Logged data with errors Incorrect drive power logged For each row in Table 10 the value in the Drive power column is the value for Drive RPM divided by ten The cause of this is a simple error in which the wrong inverter parameter was used The reason for the division by ten is that all inverter values are supplied as integers even if the actual value has decimal places The stored value therefore needs to be divided by the appropriate factor to give the actual value Negative generator power values incorrectly logged Several values in the Generator power column in Table 10 are large positive numbers where small negative values are expected Negative numbers are typically 73 UL A conversion using 4 byte represented at a low level in two s complement forma two s complement has been implemented but further test results would be required to ensure that the problem is resolved Error in efficiency calculation In resolving the more obvious probl
13. FA E5 01 04 95 24 04 5 The format of the reply packet 60132 The reply packet will have the following format Source Packet Data CRC EOT type count EAE4 Rotor RPM Gen RPM Event no Mode local remote wind speed Act power Total Prod No of SC SCH SCH 99 The size of the packet will depend on the number of active status codes In the above table Source is the node number of the answering controller Data count will depend on active status codes and will be 12 No of SC The CRC calculation is the same and applied to the same fields as in the request packet except that the Data field should also be included here Below is an example of a communication from a PC sending as node number 1 to an IC 1000 controller with node number 2 TX 01 02 01 EA E5 00 01 OF DO 04 RX 01 00 02 FAE4 14 0024 0294 0019 0001 003A Start To From type count Rotor RPM Gen RPM event mode wind speed 36RPM 660RPM F W G1 Rem 8 5 8m s 00002012 00000016 00 00 12A5 04 act Power Acc Power no of SC SC s CRC stop 8210W 22kWh Rotor RPM 0024h 36 RPM Generator RPM 0294h 660RPM Event 19h event 25 Freewheeling G1 Mode 1 Remote mode Wind speed 003Ah 58d 5 8m s Power prod 00002012h 8210W Total prod 00000016h 22kWh Numb of SC 2 0000h 0 active status codes 100 Appendix J Weg CFW 11 block diagram
14. IEEE International Symposium on vol no pp 2140 2145 June 30 2008 July 2 2008 doi 10 1109 ISIE 2008 4677307 URL http ieeexplore ieee org stamp stamp jsp tp amp arnumber 4677307 amp isnumber 4676877 Burley H R Savino J M et al 1979 Some Techniques for Reducing the Tower Shadows of the DOE NASA MOD O Wind Turbine Tower Report DOE NASA 20370 79 17 NASA TM 79202 121 Burton T Sharpe D Jenkins N and Bossanyi E 2001 Wind Energy Handbook John Wiley amp Sons Chichester New York Chan S M Cresap R L and Curtice DH 1984 Wind turbine cluster model IEEE Transactions on Power Apparatus and Systems vol 103 pp 1692 1698 Chinchilla M Arnaltes S Rodriguez Amenedo J L Laboratory set up for Wind Turbine Emulation 2004 IEEE International Conference on Industrial Technology ICIT Crabtree C J 2011 Condition Monitoring Techniques for Wind Turbines Doctoral thesis Durham University Available at Durham E Theses Online http etheses dur ac uk 652 de Oliveira R G Parma G G Silva S R 2007 Development of a wind turbine simulator for wind energy conversion systems experimental results In Proc of 9th Brazilian Power Electronics Conference COBEP 07 Blumenau Santa Catarina Brazil ISBN 978 85 99195 02 4 October 2007 DECC 2010 National Renewable Energy Action Plan for the United Kingdom Available online at http www decc gov uk assets decc what 20we 20do uk 20energqy 20supply
15. Masters degree in Energy and Environmental Engineering My intention is to complete the MSc dissertation over the summer and am looking for an external partner organisation with a practical requirement that could form the basis of a project My main interest is in the control and instrumentation of renewable energy devices but would also be interested to discuss other possibilities have attached my CV for reference and would be grateful if you could forward this message to the appropriate person in the company Thanks very much Brian Davison Programme Leader BSc Information Technology Edinburgh Napier University 0131 455 2373 84 Appendix B Initial company brief extracted from an email from the Gaia Wind engineer Project has been reduced to just providing a Piece of Software to Control the Test bed You would plan to be here from 27 June for three continuous weeks With regards to the project I d like you to use C Express to develop the application for a these reasons it ll make it easier for me to work with the project later already have some Modbus and some Gaia Comms code written which hopefully will give you a head start It will also mean should be able to help more plus have books etc which might be useful d like to break the project down into a few separate outcomes which we should aim to complete consecutively so that if the time line slips we at least have the early stuff complete Stage
16. Start up speed 2 5 m s Cut in wind speed 3 5 m s Cut out wind speed 25 m s Gear ratio 18 1 21 6 1 Table 6 Main characteristics of the Gaia Wind 11kW turbine Gaia Wind 2009 The overall turbine design shown in Figure 22 is relatively simple and provides no opportunity to alter the aerodynamic properties of the rotor via pitch control 54 Nacelle Houses the turbine drive train and speed sensors Anemometer Rotor Downwind design provides passive yaw control Control box Houses controller electronics and voltage and current sensors Figure 22 Gaia Wind turbine structure Gaia Wind 2009b The WTE hardware is constructed from two turbine nacelles with the rotors removed which are connected at the low speed shaft LSS as shown in Figure 23 The SCIM of the nacelle on the left is treated as the HUT and is connected to a standard Gaia Wind controller The SCIM of the second nacelle is used in motoring mode to drive the HUT It is fed by a variable frequency inverter and the two units together constitute a variable frequency drive VFD as described in 2 2 6 55 High speed Gearbox SCIM shaft Low speed Driving nacelle Inverter generator Disc shaft with cover closed brake Figure 23 Gaia Wind WTE hardware layout The general structure is a pragmatic one that takes advantage of the availability of standard equipment However the arrangement is not ideal since it includes energy conversion steps
17. Vector control 22 5 beet octet escebetsbandedeseteoedynined otecsd ebolybipeddecahededented 77 5 3 2 Steady torque control EE 78 5 3 3 Variable torque ue EE 78 5 3 4 Advanced control investigation cccccceeceeeeeeeeeeeeeeeeeeeeeeeeeenneeeeeeees 78 5 4 Laboratory based simulation sssnseseesseeesnnetrnneeseetnnnnnnnrnssertrnnnnnnnneseeennne 78 5 4 1 Weg inverter Simulation 2 ccnwaneweienaunnenneahawauewaines 79 542 Other SIMUANON EE 79 6 EE 80 6 1 Comparison with aims and objectives cette ee eeeeeeeeeeeeeeeeeeeeeees 80 6 1 1 Review of current relevant literature un 80 6 1 2 Hardware investigation ccceececeeeeeeeeeeeeeeeeeeenaeeeeeeeeeeeeeeeeenneeeeeeees 81 6 1 3 Software development 81 6 1 4 Evaluation of the project outcomes 81 6 1 5 Definition of programme of future work 82 6 2 Personal Geller akc cic hed Gal gett EE 82 Appendix A Email to potential partners en 84 Appendix B Initial company brief ENEE 85 Appendix C Project proposals un 87 Appendix D Project EE 90 Appendix E Weg 50Hz SCIM datasheet 91 Appendix F Weg 60Hz GCIMdatasheet unn 92 Appendix G NREL data EE 93 Appendix H NREL analysis E 94 Appendix l M NET protocol E 99 Appendix J Weg CFW 11 block dagram 101 Appendix K System specification EE 102 Ne tee EE 102 Tel e TEE 102 1 2 Manual and automatic operation sseeeseeeenenenneeeseerrrrnnnrnnserrrnnnnnneeeeet 103 2 GlASS CIAGQl ee 104 3 Interface design
18. WTE projects covered in the literature review The evaluation therefore consists mainly in verifying that the HMI is operating as specified However the data available from the system provides a means for evaluating the reliability of the values delivered by the inverter and controller interfaces 47 3 5 Timescale The initial project plan can be found in Appendix D The overall duration was already fixed as discussed previously given the target deadline of August 25 During negotiations with Gaia Wind it became apparent that it would be advantageous to have access to the hardware during the main software development period and that an intense period on site at the company premises in Glasgow would be necessary A mutually convenient period of three weeks was identified from 27 June 15 July 2011 This essentially split the project into three phases as described in Table 5 The gap in the timeline from 16 28 July was due to a holiday Phase Activities Dates 1 Literature review 10th June 26th June Familiarisation with Visual C and Microsoft Visual Studio Initial work with equipment manuals First draft of system design 2 Detailed familiarisation with hardware 27th June 15th July Software development 3 Evaluation 29th July 25th August Write up Table 5 Overall project structure and schedule 3 6 Investigation of hardware configuration The hardware for the WTE had already been assembled
19. Weg Industrias Available online at http Awww weg net files products WEG cfw 1 1 programming manual 0899 5620 2 0x manual english pdf Accessed 20 August 2011 Weg 2010b Serial CFW 11 Communication Manual Weg Industrias Available online at http Awww weg net files products WEG cfw 11 rs232 rs485 serial communication manual 0899 5741 manual english pdf Accessed 14 August 2011 Weihao Hu Yue Wang Xianwen Song Zhaoan Wang 2008 Development of wind turbine simulator for wind energy conversion systems based on permanent magnet synchronous motor Electrical Machines and Systems 2008 ICEMS 2008 International Conference on vol no pp 2322 2326 17 20 Oct 2008 128 URL http ieeexplore ieee org stamp stamp jsp tp amp arnumber 4771136 amp isnumber 4770630 Weiwei Li Dianguo Xu Wei Zhang Hongfei Ma 2007 Research on Wind Turbine Emulation based on DC Motor ndustrial Electronics and Applications 2007 ICIEA 2007 2nd IEEE Conference on vol no pp 2589 2593 23 25 May 2007 doi 10 1109 ICIEA 2007 4318881 URL http ieeexplore ieee org stamp stamp jsp tp amp arnumber 43 18881 amp isnumber 4318335 Yang W Tavner P J Wilkinson M R 2009 Condition monitoring and fault diagnosis of a wind turbine synchronous generator drive train Renewable Power Generation IET vol 3 no 1 pp 1 11 doi 10 1049 iet rpg 20080006 URL http ieeexplore ieee org stamp stamp jsp tp amp arnumber 4694978 amp isnumber 4694977 Ya
20. able to withstand the bending moment due to gravity when the blades are horizontal However it is likely to be prone to buckling under the out of plane aerodynamic force The force of gravity on a rotating blade is also a source of fatigue since the direction of the force reverses with each rotation Burton et al 2001 p 236 A further major source of fatigue is the variation in angle of attack which occurs when the plane of the rotor is not perpendicular to the direction of the wind All wind turbines rotate in the horizontal plane to bring the rotor to face into the prevailing wind direction a motion known as yaw Under real operating conditions a wind turbine cannot react quickly enough to variations in wind direction to maintain perfect orientation and therefore experiences variable aerodynamic forces most of the time Burton et al 2001 p 96 Steel plate has been used in rotor construction in early commercial turbines but its low strength to weight ratio makes it less than ideal Most modern turbines use glass fibre carbon fibre or wood composites Because of their layered construction these materials have a relatively high strength to weight ratio and also lend themselves to moulding which again is not possible with steel Although carbon fibre gives the best performance fibreglass is usually preferred on cost grounds Figure 6 shows the typical construction of a fibreglass rotor in cross section showing an internal structural web wh
21. al 2007 and Yaogin Jia et al 2007 use time series data but do not specify the type and Neannamee et al 2007 use all three types of input signal because of their interest in testing the capabilities of their WTE 2 4 Conclusion This chapter reviews the essential aspects of wind turbine design that are needed to specify and construct a WTE As noted in 2 3 the specific design of aWTE necessarily follows from its purpose and the type of tests that it is intended to support Each topic covered here could therefore be explored in much greater detail with respect to a specific WTE That was not the purpose of the review however and a broad background was intended Chapter 2 reviews compares and summarises previous work on WTE design construction and evaluation and presents a general model for a WTE This will be used to underpin the rest of this report 43 3 Methodology 3 1 Introduction The purpose of this chapter is to describe the methods used during the completion of the project The following sections cover the main stages of the project process in chronological order 3 2 Project selection An initial set of parameters was defined for the project which provided a context for the selection of the actual project Each criterion was personally motivated and the list can be summarised as follows e Focus on control engineering applied to renewable energy technology e A highly practical element e Collaboration
22. b Documentation for hardware components available from manufacturers c Interface hardware to be supplied by Gaia Wind Tom Grassie Contingency proposal Student Name Title Summary of work to be undertaken during project Deliverables Resources required Proposed Supervisor Matriculation 02014147 Number Brian Davison Reactive control of a micro wind turbine Micro wind turbines are typically located in urban environments where local turbulence interferes considerably with their efficient operation Turbulence effects occur over a small time scale and to compensate for them a control approach would need to react to rapidly changing conditions This contrasts with more stable environments in which control strategies are geared to achieving and maintaining a steady state in relation to a given set point This project would investigate the range of available data that could be used to drive a reactive control strategy and would use a programmable logic controller PLC to provide yaw control via a brake on the turbine mount Evaluation would be carried out by comparing the output from the turbine with and without active control under similar wind conditions Consider the academic challenges that you will face and what solutions you proposed to deliver as part of the project and consider what will make the project a success There should be no more than 5 N B Approved deliverables will be compared with your final
23. by Gaia Wind It was therefore important to become familiar with the components that were already in place The central component was the Weg CFW 11 IGBT inverter Weg 2008 and familiarisation largely consisted of working systematically through the programming manual to identify the relevant operational parameters that would need to be manipulated The detailed results of this work can be found in the nverter Communications section of the system specification document in Appendix K The other main pieces of equipment that were included in the investigation were the turbine itself and the Weg SCIM generator whose datasheet can be found in Appendix E Turbine details were obtained from test documentation from the Unites States National Renewable Energy Laboratory NREL relating to the Gaia Wind turbine NREL 2009 and from various Gaia Wind publications Gaia Wind 2008 2009 2009b 48 The familiarisation phase did not really have a defined end since new and revised information was identified throughout the development period 3 7 Software development The requirements for the system were reasonably clear from the outset a first draft of the system specification having been sent to Gaia Wind on 23 June The development approach adopted was an agile one in which the features of the system were prioritised so that less important details could be omitted if delays occurred A further feature of agile development is the regular d
24. deliver a steady motor speed as in this first stage project Because of this requirement the simplicity of V f control and because V f control is the default strategy provided by the inverter it was adopted for this project The Weg manual claims an accuracy of 1 of motor rated speed when the inverter is used in V f mode 62 l Torque and Voltage Base Speed Frequency Figure 28 Torque speed characteristic of VFD using V f control Parekh 2004 In addition to the V f strategy the CFW 11 provides vector control in which the motor current is divided into two components using the Park transformation as described in 2 2 7 In this mode the CFW 11 offers direct control over the developed torque Vector control can be used as an open loop configuration which relies solely on an accurate description of the motor via the relevant parameter values Weg quotes an accuracy of 0 5 of rated speed in sensorless mode whereas in a closed loop configuration using feedback from a shaft encoder accuracy increases to 0 01 The fourth control strategy Voltage Vector Weg VVW offers a level of performance which lies between V f and sensorless vector control The second stage of the Gaia WTE development would require the use of vector control to recreate the reference torque specified by the turbine model At that stage the hardware configuration could also be improved by the addition of a shaft encoder to provide feedback for clos
25. ensures that the torque produced corresponds to the reference torque This arrangement shown in Figure 15 is ideal for examining the control of the generator and the conditioning of the output power Because it combines virtual elements with real ones it is often referred to as hardware in the loop HIL Munteanu et al 2010b or power hardware in the loop PHIL Ayasun et al 2007 Strictly speaking PHIL designates a system in which real power is passed between real and virtual components whereas in a HIL system only information signals are exchanged Grid Generator e High speed shaft Figure 14 Schematic representation of wind turbine Turbine model Grid Reference torque Generator Electrical drive Figure 15 Schematic representation of wind turbine emulator 36 Clearly the schematic in Figure 15 represents a very general model of a WTE and additional hardware components would be required to emulate more complex designs which include say a DFIG generator Likewise an appropriate computer model would be required for the particular turbine design being emulated It is also worth noting that given an appropriate model of the prime mover the arrangement shown is not restricted to wind turbine emulation and could equally well be applied to other forms of electricity generation This generic approach is the one taken by Mauri et al 2008 and Munteanu et al 2010 A distinction can be made between
26. features 9 1 Graphical output The design includes a scrolling graphical trace of the speed and power from the two devices This is implemented using the C library by Stephan Zimmermann which is available from http www codeproject com KB miscctrl GraphPlotting aspx 9 2 Excel export The user needs to log test statistics to a file ideally in Excel format This functionality is provided by the Office Open XML library available from http excelpackage codeplex com wikipage title Creating 20an 20Excel 20spre adsheet 20from 20scratch amp referring Title Home Overall software design The system requires a number of activities to be carried out simultaneously without interfering with each other A multi threaded design is therefore required which allows background tasks to be completed without preventing the user interacting with the interface Both the inverter and the turbine controller are treated as subclasses of a generic serial device which is associated with its own thread This allows each device to poll continuously for status information independently of any other activities When a scripted test is being run it also needs an independent thread so that the user can interrupt it if necessary and the monitor function which updates the user interface with status information likewise runs independently Along with the user interface itself a total of five threads can be running simultaneously which can lead to timing and inter thre
27. in this way Neammanee et al 2007 on the other hand use a SCIM in their WTE and use the speed of the generator shaft as a surrogate for torque Exactly how this is done is not explored in the paper De Oliveira et al 2007 who also use a SCIM rely in contrast on the relationship between torque and slip illustrated in Figure 10 Finally torque can be measured directly using a torque transducer Crabtree 201 1 takes this approach and also collects a wide range of other sensor readings because of his interest in fault diagnosis Dolan et al 2005 also measure torque directly as mentioned earlier in order to compensate for motor inertia Only extremely simple WTEs in the literature rely on open loop control Martinez et al 2009 All others implement a form of closed loop control in which the torque delivered by the emulator is compared to the reference torque value supplied by the turbine model as shown in Figure 19 Note that the diagram shows the rotational speed of the drive generator shaft being used to calculate actual torque using the relationship between torque and slip This element could equally well be replaced by a torque transducer to measure torque directly 1 See for example http AWwww datum electronics co uk how to measure torque aspx 41 Wind Qre reference torque speed Qoorr corrected torque Qer torque error w rotational speed of drive generator shaft Turbine model Variable Frequency Drive
28. logfiles generated for a single test which is run on different occasions and the database maintains the link between logfiles and test The user interface can therefore list all logfiles for a particular test on request and the user can view or delete them 103 2 Class diagram Classes marked with an asterisk have corresponding database tables Speedtest testid stepNo lt gt description action ui duration execute state stopTimeout MNet queue rotorRpm modbus comPort generatorRpm portName eventNo baudRate mode uiDownPeriod dataBits windSpeed parity actualPower stopBits accumulatedPo readTimeout init enable getStatus 0 1 O sendMessage 0 1 1 1 1 1 1 1 buildMessage run buildMessage stop uiUpPeriod 104 3 Interface design The general layout corresponds to that of a standard Windows application with drop down menus and a toolbar in automatic mode In the sketches below only four icons are shown on the toolbar but in fact there would be one for each menu option Menu items are described in the following table and the subsequent section provides some additional detail on the layout 3 1 Menu items Menu Item Description File New Clears screen for new input Open Dialog allows user to select an existing test from the database Save Saves current test to the database Import Allows the user to read test details from an external XML file
29. of the turbine under test can be observed Data from the onboard turbine controller will be used to provide process feedback Based on this initial development a plan for further improvements to the control algorithms can be constructed The main challenges are therefore a To identify the capabilities of the available hardware components b To design and develop appropriate communications interfaces between the hardware components and a master PC application c To identify and measure appropriate variables for analysis and control d To ensure that the basic system takes into account the longer term goals of the development The first part of the project will be to produce a review of current literature on turbine testing and test rig instrumentation and control The practical phase of the project will include the following activities a Familiarisation with the hardware components and the selection of appropriate technologies for communications and control b Development of appropriate communications and human machine interfaces c Evaluation of the basic control system and the identification of promising options for further development Deliverables Resources required Proposed Supervisor 1 Literature review 2 Specification of initial control system 3 Communication and control software and documentation 4 Evaluation of results 5 Forward plan for development a Physical test rig at Gaia Wind Ltd facility in Glasgow
30. outside the scope of this project For several of the machine characteristics the datasheet provides three different values corresponding to standard voltages in different contexts The relevant voltage in this situation is 400V which means that where three values are shown it is the second which is relevant As a six pole machine connected to the UK distribution network at 50Hz the synchronous speed of the generator will be 1000 rpm as explained in 2 2 6 The slightly lower speeds shown on the datasheet take the slip into account under the related load conditions The SCIM datasheet also provides values for power factor and efficiency at different loads However the rated power of the SCIM is chosen to match the rated power of the wind turbine which is only achieved at the rated wind speed of 9 5 m s The NREL duration test report shows that rated wind speed was achieved about 26 of 59 the time over a 9 5 month period at the NREL test facility in Boulder Colorado Huskey et al 2010 This is a fairly typical pattern which suggests that to replicate such a wind regime the SCIM of a WTE would be working with a small load Even in this first stage project the power input will be fairly low and drive efficiency is likely to be around 88 As well as important information on electrical performance the datasheet also quotes a figure for the moment of inertia of the SCIM rotor which could be used in Eq 11 for controlling the delive
31. polling period is greater than 1s If a smaller polling period is used the x axis labels are incorrectly displayed This is because the points to be labelled are identified using the modulus function and any fractional values are effectively rounded up Apart from the visual effect the display and logging works correctly Issue logging and version control will be implemented following the end of the academic project in order to identify and resolve any further bugs 4 2 6 Evaluation The software was developed over a very short period of time and as usual with short project only a minimum amount of time was available for testing All of the major functions have been shown to operate correctly however a comprehensive examination of all possible error conditions has not been carried out It is to be expected therefore that there will be unhandled exceptions from time to time when the software is in use The company will log any such errors for fixing later The current class hierarchy was developed in an incremental way and it does not therefore capture as accurate a representation of the hardware as possible The inverter and turbine were identified as subclasses of a generic serial device however the same was not achieved for the two protocols used in the application Ideally a generic protocol class should be constructed with Modbus and M NET as subclasses This would also be an opportunity to accommodate the protocol 71 message an
32. prevented their completion e Serial communications issues absorbed a great deal of the available time e The need for a PC with a native serial port rather than the author s laptop meant taking that PC away from its normal use e There were competing priorities regarding the use of the WTE hardware during the latter stages of the project e Access to a turbine controller with the extended version of the M NET protocol was difficult to arrange e An error in the logging feature of the software system meant that the first tests did not provide adequate data e The Gaia engineer was unable to help with running tests in the last few days before the deadline 72 At the time of writing a number of errors in the logging function of the software have been resolved and an updated version of the software has been sent to Gaia Wind on 22 August However it has not been possible to run any tests since that time and Appendix L contains email communications between the author and the Gaia engineer regarding the logging function To illustrate the issues with the earlier version Table 10 shows an extract of actual data logged by the system during a test attempt run by the Gaia engineer on 18 August The four main problems are briefly discussed below Drive Drive Generator Generator Efficiency Timestamp RPM power kW RPM power kW Slip 18 08 2011 11 17 08 1005 100 5 1003 65 436 65 100 18 08
33. project Part of the reason for this was the simple nature of the software because the turbine model had been excluded there appeared to be insufficient grounds for comparing performance of the WTE with any other installation The design of the software makes use of object oriented principles but this was not previously the author s primary area of expertise A number of technical lessons were learned along the way and the final result is less than perfect Some generalisations such as the relationship between the two communications protocols were not identified until late on in the development and were therefore not captured There was also a reliance on simple software applications previously developed at Gaia Wind which were not based on object orientation This has led to some inconsistencies in the construction of the software that might cause some initial confusion for future developers working with the current code The current version performs adequately however and the experience has provided insights into many aspects of object oriented development and the use of the NET framework 83 Appendix A Email to potential partners Dear XXXXXX hope you don t mind me contacting you out of the blue but you are listed as the main contact for XXXXXXX on the Scottish Renewables Web site am a member of the teaching staff in the School of Computing at Edinburgh Napier University but also over the last two years have been doing a
34. report 1 Literature review 2 Model of the relationship between selected parameters and brake activation 3 Hardware solution including sensor PLC and brake 4 PLC control program 5 Final evaluation of results a Access to micro wind turbine at Edinburgh Napier b Hardware components sensor wiring PLC brake etc c PLC programming software Tom Grassie 89 Appendix D Project plan dn Sunum eu stuouidojenep amin 10 stesodoid wass Hoi ul suonenun 40 uoneounusp seyau saanu Jo zuawudoj naq Qs Asojouyoe yum uonesuemwes D voten yo uonenjergiuoissnosia D uojsieA 7 weld moines amesa n mes amesaini adoos riede oneno n 5 pum sen yum rar et REEL Doro Soe some SORT somo cove cone cos como coc race POs POI FORO bny ze Dny Si Ony 80 Oy LO L INF Sejt INF Bt IN LEE Ne POT UNF Ze unr Oe unr Eer unr 90 f ABW OF ABW Ec Aen 91 ABW 60 Aw zo 10W Se 10W 84 10W Lb dv 70 90 Appendix E Weg 50Hz SCIM datasheet Date 26 APR 2010 istri WEG Ind strias S A EE DATA SHEET Three phase Induction Motor Squirrel Cage Customer Product code Product line 1E2 HIGH EFFICIENCY Frame 160L Enclosure IPSS TEFC Output 11kW Mounting B3T Frequency 50 Hz Rotaton Both Poles e Aprox weight 133kg Rated speed 970 970 975 mm Moment of inertia 0 1760 kgm Slip 3 00 3 00 2 50 Sound Pressure Level
35. rotating at 33 r p m 500 Blade pitch below rated increased 1 4 degrees compared i i uA with stall regulated machine Stall regulated machine with TR Ki blades rotating at 30 r p m 400 e 300 Power kW Rotor diameter 40 m 100 0 5 10 15 20 25 30 Wind speed m s Figure 13 Effect of pitch regulation on performance Burton et al 2001 p 342 Active manipulation of the aerodynamic behaviour of the blades requires sensors such as anemometers to provide information about prevailing conditions and digital controllers to select the appropriate settings for blade pitch actuators This lays the basis for closed loop control in which operating parameters are adjusted in response to changes in a process variable such as shaft rotation speed to bring the system behaviour closer to a known optimum The desired value of the process variable is known as the set point and in the case of wind turbines where external changes account for the majority of adjustments the set point is determined by reference to a mathematical model of optimum system performance Closed loop control is also essential in turbines with yaw drives to maintain the optimum orientation of the rotor plane Burton et al 2001 p 477 While aerodynamic control is primarily protective the goal of generator control is concerned with maximising the efficiency of the power conversion This is most obvious in the case of variable speed turbine
36. the turbine side high speed shaft for example would provide a direct measurement of the developed torque Crabtree 2011 and Hus Wen Ko 2010 describe the use of different type of sensor for the detection of various kinds of fault condition The scope of the WTE could easily be expanded in this way to support different kinds of test 5 3 Inclusion of turbine model The turbine model has been researched as part of this project and its implementation is the obvious next stage of the WTE development However there are several identifiable steps towards this goal 5 3 1 Vector control Taking advantage of the facilities provided by the Weg inverter the type of control could be changed from V f to sensorless vector without altering the overall functionality of the current software This would require a similar range of activities to the current project and would provide the basis for moving forward 12 http www mathworks com products matlab 77 5 3 2 Steady torque control Replicating the torque produced by a turbine rotor entails direct control over the torque developed by the VFD Allowing the user to provide target torque values via the HMI is not particularly intuitive however Instead this step would require a minimal turbine model which would provide a reference torque value based on a wind speed value In the first instance the WTE would allow steady state operation with respect to a particular wind speed input and tran
37. the output power and synchronous speed as shown in Eq 9 where Qis torque P is power W is synchronous speed The simplest electrical arrangement for a turbine is for the generator to be directly connected to the distribution network and in that situation the network frequency determines its rotational speed This arrangement is very common and is known either as the Danish model or more descriptively as a fixed speed turbine Muller et al 2002 The synchronous speed is given by the formula n 120 10 where fis the network frequency N is the number of poles in the EM Parekh 2003 Thus for a 6 pole EM connected to the UK electricity grid which runs at 50 Hz the synchronous speed is 1000 rom When the torque generated by the blades is sufficient to turn the EM rotor at more than 1000 rpm therefore the turbine will export power to the network At lower speeds the turbine will actually draw power from the network unless it is isolated The major disadvantage of the fixed speed design is that the turbine will operate sub optimally at wind speeds that do not correspond to the peak tip speed ratio Burton et al 2001 p 360 In order to allow a turbine to operate at variable speed the generator must be decoupled from the network by interposing a frequency converter however this significantly increases the complexity of the design and 29 hence the cost Muller et al 2002 An appropriate frequency converter would b
38. the set of physical WTE components and those that replace elements of the real environment such as the prime mover Testing usually focuses on the behaviour of the former which in PHIL terms is designated the hardware under test HUT 2 3 1 Drive selection Several teams of researchers have suggested using a DC electric motor as an alternative source of torque in a WTE Chinchilla et al 2004 Lopes et al 2005 Monfared et al 2007 Weiwei Li et al 2007 Mauri et al 2008 Martinez et al 2009 Moore and Ekanayake 2010 Munteanu et al 2010 2010b Crabtree 2011 The attraction of a DC machine is that the torque it produces is directly proportional to the armature current and is therefore very easy to control Lopes et al 2005 In fact Martinez et al 2009 propose a WTE for fixed pitch turbines which consists of nothing more than a DC motor connected in series with a resistance and a variable voltage source However no detailed evaluation is presented and the main justification for the design is that the power curve of the DC motor is roughly the same as that of a fixed pitch turbine Permanent magnet synchronous machines PMSM have also been proposed for use in WTEs Dolan et al 2005 Weihao Hu et al 2008 The main benefits of a PMSM are that it offers higher impulse torque and power density and faster speed of response in comparison to a DC machine Dolan et al 2005 Because it does not require an excitation system
39. worldwide installed capacity reached 196 63 GW in 2010 WWEA 2011 This strong growth is illustrated in Figure 1 World Total Installed Capacity MW 300 000 250 000 oe 200 000 Soe 150 000 Y Du a d o A Fa D AS 100 000 dr SS Y L d s Sg e Ti aaa 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 Figure 1 World wind turbine installed capacity WWEA 2011 Focussing on the situation in Europe a recent report by RenewableUK previously the British Wind Energy Association shows a very uneven deployment of turbines Whereas Denmark and Germany have relatively high turbine densities at eleven and six turbines per hundred square kilometres respectively the UK has just over one RenewableUK 2011 The UK therefore currently lags behind other European countries despite its richer wind resource however the UK government s target of 12 producing 15 of its electricity from renewable sources by 2020 DECC 2010 would require among other things that onshore turbine density increases to around 2 44 turbines per hundred square kilometres RenewableUK 2011 The market for wind energy in the UK is therefore promising in the short term with its underexploited capacity and positive economic pressures There are wide differences in the deployment of turbines among the constituent nations of the UK Scotland already sources 15 of its electricity from wind and accounts for 61 of UK installed capacity
40. 1 Inverter Control Module Modbus Code Module to Start Motor Stop Motor Set Acceleration Set Deceleration set Speed It should be able to read Motor Power and RPM Plus a basic Test Application Stage2 Gaia Query Module Module to Get the Power Data from the controller Our current version of control firmware has Wind Power Energy and errors available The firmware in development also has Generator and Rotor RPMs already a small application in C though it might need a little tidying up Stage 3 Overall Application User Interface to Display Live Drive Motor Speed and Power Generator Speed lf available and power There should be two modes Manual mode and Auto Mode In Manual Mode you can set Speed Acceleration Deceleration and Start Stop the Inverter 85 In Auto Mode there should be a start button that runs a script Script should be text based so it can be edited easily XML perhaps but each entry should be something like this bool Run Acc Value Decc Value Speed Time Example1 Start Accelerate over 10 seconds Decelerate don t care Speed 1000 60 seconds The time wouldn t start counting down until actual speed set speed 1 10 10 1000 60 Example2 Bring the Motor to a stop over 10 seconds Time 0 O 10 10 0 0 Example3 Test Run 1 10 10 1000 60 A Enable Drive Accelerate up to 1000 rpm over 10s then spend 60 seconds at speed 1 2 10 1010 60 Enable Driv
41. 17 96 2 65 0 02 114 47 15 14 40 0 05 56 44 5 91 127 67 1219 17 2 56 0 02 112 79 15 48 14 49 0 05 56 50 5 92 127 80 1220 37 2 48 0 02 113 38 16 03 14 42 0 04 56 55 5 92 127 92 1221 57 2 40 0 02 112 72 16 5 14 15 0 04 56 61 5 93 128 05 1222 78 2 34 0 02 110 50 16 99 14 24 0 04 56 67 5 93 128 17 1223 98 2 27 0 02 111 10 17 5 14 13 0 03 56 72 5 94 128 30 1225 18 2 21 0 01 110 13 17 99 14 08 0 03 56 78 5 95 128 43 1226 39 2 15 0 01 109 63 18 48 13 91 0 03 56 83 5 95 128 55 1227 59 2 09 0 01 108 20 18 99 14 00 0 03 56 89 5 96 128 68 1228 79 2 04 0 01 108 80 19 45 14 11 0 02 56 94 5 96 128 80 1230 00 1 99 0 01 109 55 19 92 13 74 0 02 57 00 5 97 128 93 1231 20 1 95 0 01 106 57 93 Appendix H NREL analysis In 2009 and 2010 NREL carried out a series of tests on the Gaia Wind turbine according to the International Electrotechnical Commission IEC standards The reports are now published and constitute a reliable description of the operation of a real turbine The most relevant report is the result of the power performance test Huskey et al 2009 which includes binned wind speed data with corresponding instantaneous power and power coefficient as shown in Appendix G Combining this information with the published turbine details Gaia Wind 2009 a number of other measures can be derived including angular velocity of the high speed shaft and load torque However there are two main limitatio
42. 3 37 Methodology EE 44 SCH tee HIE ee ee ee ee ee Een 44 3 2 dee o EE 44 39 9 Bee 45 EN ee 46 3 4 1 Evaluate hardware sein 46 3 4 2 Establish communication with VFD ssssssssssssseeeeeeeeerrerereeerrrrrerreeeeeeeeene 47 3 4 3 Establish communications with turbine controller ccseeeeeeeees 47 344 e Eder D e EE 47 eP OP dE ee 47 25 WIMeSCale tee 48 3 6 Investigation of hardware Confouraton 48 3 7 Software development ee 49 BES AS VSION VCO SION EE 49 3 9 Evaluation Of results csc eke innnan een ae 51 3 9 1 Comparison with real installation cccceecceeeeeneeeeeeeeeeeeeeeeenneeeeeeees 51 3 9 2 EIER 52 3 9 3 EE 53 GE 54 4 1 Hardware configuration ice 2s52 2252 en coehensnss ites eneheheeennetecensiheenmetieied teeta 54 Blels QVETWCW EE 54 4 1 2 Instrumentation and control ei cocccvcsseccecsives couziveiavedemneceusioniverdivecemedacecs 57 r N e 59 4 174 et ee 60 4 1 9 NED ere io ans ata ia Caan a nln tenia hia thin ti aia 62 4 16 EENS cee ad eae cette 63 4 2 Software EE 64 423i NEE 64 4 2 2 HMI and general system features AEN 64 42 3 Resolved EE 67 4 2 4 Ell ere 68 4 2 5 Residual EE 71 APG SEV de 71 GEI 72 EOI A GROAN EE 75 5 1 Limitations of current mplementaton 75 5 2 Alternative implementations EEN 76 en LEE 76 5 2 2 Weg Ladder Programmmg ENEE 76 5 2 3 Real time operating Systems EE 76 524 EE 77 5 2 5 Increased instrumentation EE 77 E Inclusion Be r TAT 77 5 3 1
43. A guide for beginners 3rd Edition JUMO GmbHCo KG Fulda Germany ISBN 3 9357 42 01 0 127 S rensen P Hansen A D Andre P Rosas C 2002 Wind models for simulation of power fluctuations from wind farms Journal of wind engineering and industrial aerodynamics 90 pp 1381 1402 S rensen P Hansen AD lov F Blaabjerg F Donovan MH 2005 Wind farm models and control strategies Technical Report RIS R 1464 EN RIS National Laboratory Roskilde Denmark Teodorescu R lov F Blaabjerg F 2003 Flexible development and test system for a 11 kW wind turbine In Proceedings of 34th IEEE power electronics specialists conference PESCO03 vol 1 June 2003 p 67 72 Tongguang Wang and Coton F N 2001 A high resolution tower shadow model for downwind wind turbines Journal of wind engineering and industrial aerodynamics 89 pp 873 892 Tunnicliffe J 2007 ExcelPackage Office Open XML Format file creation Available online at http excelpackage codeplex com Accessed 14 August 2011 Weg 2007 WLP software manual Weg Industrias Available online at http www weg net files products WEG wlp software manual 10000051 171 8 7x manual english pdf Accessed 12 August 2011 Weg 2008 CFW 11 User s guide frequency inverter Weg Industrias Available online at http catalogo weg com br files wegnet WEG cfw 1 1 a users quide 10000063093 manual english pdf Accessed 14 August 2011 Weg 2010 CFW 11 Programming Manual
44. Alttude DATA SHEET Three phase Induction Motor Squirrel Cage WEG ELECTRIC MOTORS UK LTD 1E2 HIGH EFFICIENCY 160L Enclosure 12 5 KW Mounting 60 Hz Rotation 6 Aprox weignt 1170 1170 mm Moment of inertia 2 50 2 50 Sound Pressure Level 440 480V 22 421 9A 157 153 A S 11 013 0A 102 102 Nm 210 250 Front 240 280 Rear N E Lubrication interval 20000 h 80K Grease Polyrex EM ESSO 13 13 s hot 1 00 1 20 C to 40 C 1000 ma s IP55 TEFC B3T Both 133 kg 0 1760 kgm 59 0 dB A global Power factor 0 81 0 75 0 76 0 68 0 63 0 55 Bearing 6309 C3 6209 Z C3 Efficiency 89 6 90 4 89 6 90 1 89 0 88 5 Quantity lubricant 139 9g The figures given herewifh are regarded guaranteed values and apgiled to sinusoida power supplied matos within permissible tolerances under IEC 60034 1 Noise level with tderance of 3 dB A CH Weight vauecan be changed witiout previous nofification Effidencies according to the indirect method of IEC 60034 2 1 2007 with stray load lasses deternined from measurement Revision Approved Nr 0 Date 14 MAY 2010 92 Appendix G NREL data Wind speed Power kW Cp Q rpm O rad s Ouss rad s Qyss rpm A Ca Q 0 49 0 08 8 43
45. Development of a wind turbine emulator control system Brian Davison 02014147 Supervisor Tom Grassie MSc Energy and Environmental Engineering Gaia Wind premises in Glasgow with pylon in the background Abstract Wind turbine design and development is currently a very dynamic industry and designers need appropriate tools and resources to experiment with innovations and to evaluate the effect of modifications However access to installed turbines in situ for testing and monitoring is typically difficult Hardware in the loop simulation offers a solution to this problem A wind turbine emulator uses a real or scaled down drive train but typically replaces the prime mover with an electric drive Based on a review of previous work on wind turbine emulators this project proposes a control system designed to reproduce a variable wind profile taking into account torque oscillations due to wind shear and tower shadow Initial software development is undertaken to provide a basic human machine interface to emulator hardware The software does not attempt to implement the turbine model at this stage Results from evaluation and software testing identify limitations of the current configuration A series of follow on projects is outlined Acknowledgements am indebted to Derek Robertson the Gaia Wind electrical engineer who was instrumental in agreeing the project at the beginning and who was patient and supportive throughout the
46. Export Allows the user to save the current test as an external XML file Delete Deletes current test from the database and clears screen Closes the application Test Validate Check current sequence of instructions against validation rules Runs the current test Stop Interrupts the current run Log Directory Allows the user to select a directory for storing logfiles 105 Notes Not available while test is running Not available while test is running Not available while test is running Instructions are validated before saving and if errors are found the save operation fails Not available while test is running Not available while test is running If test has not previously been stored Delete has the same effect as Open Not available while test is running User is prompted to save current test if changes have been made Only available if changes have been made Not available while test is running If changes have been made the instructions are validated before running If errors are found the run operation fails Only available while test is running Brings the drive to a graceful stop and disables Not available while test is running Log Toggles logging on and off The name of the logfile is generated automatically and includes the id number of the test and a timestamp Open Allows the user to select a logfile for opening with an appropriate Windows application eg Excel Help Manual Opens the
47. Kojabadi and Chang 2004 propose an inverter based on insulated gate bipolar transistors IGBT This type of inverter is typically composed of two main components as shown in Figure 16 The rectifier converts the standard AC supply to DC and the inverter component whose basic structure is shown in Figure 17 uses a set of IGBTs under microprocessor control to produce a series of pulses in three phases The resulting output in each phase uses pulse width modulation PWM to synthesise a sinusoidal voltage of the required frequency as shown in Figure 18 The inverter therefore converts the problem of frequency and voltage magnitude control into one of IGBT switch timing Gross 2007 Three phase PWM K constant V Three phase variable v Three phase constant f Se variable f i 8 diode rectifier induction balanced three phase motor AC source inverter e o Wim a i S at a db e me pe aw pa a ne ets Ne i e Motor drive Mechanical load Figure 16 General configuration of PWM motor drive Gross 2007 38 To balanced three phase load Figure 17 The three phase IGBT circuit diagram Gross 2007 ail TH d ki j M Wil Figure 18 Approximation of a sine wave using PWM Adapted from Gross 2007 2 3 2 Turbine model The purpose of the turbine model is to provide a reference torque for the WTE drive It therefore consists of a static element whi
48. RPM 1 RPM Target speed must be equal to current speed 0 lt Duration lt Duration cannot be negative Con 7 Durationmax Stop Inverter state on Test must start with a Start instruction RPM 0 Previous instruction must bring the drive to Stop standstill n onl r as last instruction Stop N Nmax Can only occur as last instructio 4 3 Changing speed In cases where the acceleration and deceleration ramps are used the next instruction would not be processed until the required speed is achieved This may add to the overall duration of the test 108 Where an instruction requires a change of speed over a period longer than the natural response lag the software would artificially extend the change to produce a smooth ramping of the speed over the specified period 4 4 XML file format The format proposed follows the sequential nature of the instructions in a test Each test is represented by a lt GaiaSpeedTest gt lt GaiaSpeedTest gt element and each instruction in the test is a child node Each lt instruction gt lt instruction gt element must have an lt action gt lt action gt element and may have lt duration gt lt duration gt and lt rpm gt lt rpm gt elements The XML example below corresponds to the following set of instructions Start drive Accelerate up to 1000 rom over 10s then spend 60 seconds at speed Accelerate up to 1010 rom over 2s then spend 60 seconds at speed Accelerat
49. Thanks Derek 120 References Akpolat Z H Asher G M Clare J C 1999 Dynamic Emulation of Mechanical Loads Using a Vector Controlled Induction Motor Generator Set IEEE transactions on industrial electronics 46 2 Anaya Lara O Jenkins N Ekanayake J Cartwright P Hughes M 2009 Wind energy generation Modelling and control Jonn Wiley amp Sons Ltd Ayasun S Fischl R Vallieu S Braun J Gadirli D 2007 Modelling and stability analysis of a simulation stimulation interface for hardware in the loop applications Simulation Modelling Practice and Theory 15 pp 734 746 Bakshi U A and Bakshi V U 2009 Electrical technology Technical Publications Pune India Bindner H Rosas P A C Teodorescu R Blaabjerg F 2004 Stand alone version of the 11kW Gaia wind turbine Ris R 1480 EN Ris National Laboratory Roskilde Denmark Bouscayrol A Guillaud X Delarue Ph 2005 Hardware in the loop simulation of a wind energy conversion system using energetic macroscopic representation Industrial Electronics Society 2005 IECON 2005 31st Annual Conference of IEEE vol no pp 6 pp 6 10 Nov 2005 doi 10 1109 IECON 2005 1569302 URL http ieeexplore ieee org stamp stamp jsp tp amp arnumber 1569302 amp isnumber 33243 Bouscayrol A Delarue P Lemaire Semail B 2008 Graphical description for Hardware in the loop simulation ndustrial Electronics 2008 ISIE 2008
50. ad communication issues 117 Appendix L Communications regarding data collection Sent 18 August 2011 11 21 To Davison Brian Attachments 2011_08 18 TestRun1 xlsx 77 KB Open as Web Page 2011_08 18 TestRun2 xlsx 76 KB Open as Web Page Hi Brian Had to tweak the Run sequence a little to work with the new inverter set up but other than that the test run went well 1 Start 2 Acc 10 1005 3 Con 60 1005 4 Acc 2 1015 5 Con 10 1015 6 Acc 2 1025 7 Con 10 1025 8 Acc 2 1035 9 Con 10 1035 10 Dec 5 1000 11 Con 15 1000 12 Dec 5 986 13 Con 5 986 14 Dec 10 200 15 Stop First let me explain a few changed to the turbine behaviour with the new set up 118 When the inverter detects a spinning turbine it enables at the correct frequency to match rotation but at a lower voltage then ramps up the voltage This feature makes it easier to test Section 3 is there to give the turbine plenty of time to Ramp Up When this system is running at 50Hz and Power is lt 500W the generator frequency is dropped to 49 5Hz by running at a lower speed we hope to try and squeeze a little bit more power out the turbine at the low end So when slowing down stage 11 1000rpm is to allow the turbine to enter it s lower freq mode you ll see a jump in power when that happens as this has increased the slip Then when in low frequency mode if the turbine is producing lt 500W the turbine will start to brake using the inverter Section 13 is
51. ameters Up ramp time Name Description Down ramp time Name Description Speed Name Description Name Description Name Description Name Description Status Name Description Speed Name Description 5 1 Notes A The user can set three parameters directly up and down ramp times and speed The ramp times are used to calculate the appropriate ramp parameter setting see Acceleration section below B The status section highlights the main status values for the drive These duplicate those in the parameter list C The values in this list are read directly from the inverter 6 Test management module When the user starts a test many of the functions in the user interface are disabled and control is passed to the test management module which loops through the set of instructions The pseudocode below describes the main operation of the module 111 for each instruction compose appropriate inverter control message s send inverter message s while instruction not complete poll inverter status The Weg inverter provides a large range of operating parameters which can be set and read over a serial channel Some of these allow a ramp facility for acceleration and deceleration actions which means that the C program only needs to wait for the instruction to complete rather than actively managing the duration of the action See the Acceleration section below for details 7 Inverter communica
52. and a CV was included as an attachment Between 22 April and 31st May 11 replies were received out of which 5 were polite refusals This left 6 possibilities Negotiations were undertaken by email telephone and face to face until a final decision was made on 9 June 2011 to select the current project in cooperation with Gaia Wind Ltd This decision was taken slightly later than anticipated but the negotiation period allowed a clear project brief to be developed which is provided in Appendix B Prior to the final project selection two alternative project proposals were lodged with the dissertation Module Leader one related to the Gaia Wind project and another as a backup which did not rely on any external collaboration This was a contingency in case of problems during the negotiation process The two initial proposals can be found in Appendix C along with a second version of the Gaia Wind document which was the result of scoping discussions with the company 3 3 Literature review Early literature searching focussed on previous examples of WTE projects Science Direct and Google Scholar were the most useful search tools and a large number of sources were found spanning the period 2003 2011 Given the abundance of examples a representative set was selected that covered the indicated time period but which concentrated for the most part on more recent work It was clear from the outset that some theoretical subjects needed to be i
53. at such issues please refer to Burton et al 2001 The cross section of a rotor blade at any given point along its length can be modelling in two dimensions as a standard aerofoil The characteristic shape of an aerofoil gives rise to the phenomenon of aerodynamic lift when air flows over it ata given angle The lift force acts at right angles to the actual air flow while a second component of the total force on the aerofoil known as drag acts in the same direction as the air flow Figure 3 summarises the aerodynamic forces on an aerofoil set at a given angle a to the incident air flow This is known as the angle of attack and variations in a have a large effect on the relative sizes of the lift and drag forces Specifically once the angle of attack reaches a critical value of around 10 to 16 the behaviour of the air flow leads to the drag force outweighing the lift force and the aerofoil is said to be in stall Burton et al 2001 p 166 Lift L Se E Aerofoil zero KS Ip line Z Wind direction Figure 3 Aerofoil lift and drag forces Adapted from Burton et al 2001 p 61 17 Because the blades of a wind turbine are rotating an air flow is also induced in the opposite direction to the rotation which is to say perpendicular to the air flow due to the wind It is common therefore to talk about the vector sum of these two flows as the air flow experienced by the rotor blade It is the angle of attack relative to the resul
54. bine controller values are displayed on the right of the feedback panel A turbine reset button which was not in the original design was included on request so that the operator can recover from error conditions such as rotor overspeed This duplicates the function described in the user guide Gaia Wind 2008 Table 8 Notes on HMI layout Figure 30 also shows one of the alternative feedback panels which gives a graphical indication of the speed of both drive and generator 3 The third tab not shown offers a similar display comparing drive and generator power Drive power is obtained from the inverter parameter P0010 which is the calculated power output of the VFD i e the power input to the WTE and generator power is the measured electrical power from the generator Le overall output power These values are the basis for the calculation of system efficiency The graphical displays are achieved using an open source C project Zimmermann 2009 4 http www opensource org licenses gpl license php 66 e a Formi File Test Conf Sampling period s 1 00 gt Visible s 20 d Log Ato Manual Values Speed Power Settings Up time 15 Down ramp time 15 RPM 200 Equipment Ee Jg ee Status Speed pm Torque Power kW Figure 30 HMI showing Manual tab and graphical display 4 2 3 Resolved issues The system requires a number of activities to be carried out simultaneousl
55. cal model of turbine behaviour Inverter control software a el Evaluation of results 87 Resources required Proposed Supervisor a Physical test rig at Gaia Wind Ltd facility in Glasgow b Wind and turbine data to be collected from a live site Some data may also already be available c Appropriate software for Weg inverter programming available from Weg Tom Grassie Gaia Wind project version 2 Student Name Title Summary of work to be undertaken during project 88 Brian Davison Matriculation 02014147 Number Development of basic control system for a wind turbine emulator The Gaia Wind 11kW is a two blade fixed pitch fixed speed turbine for small applications Gaia Wind Ltd is currently developing a test rig which consists of two turbines nacelles connected hub to hub so that one can be used as a motor to drive the other The motor element is fed by a Weg CFW 11 inverter which has the facility for programmatic control The ultimate goal is to be able to read a file of logged wind data and for the inverter to accurately recreate the equivalent forces via the motor so that the behaviour of the passive turbine can be observed This project will be concerned with the initial stage of the development which is to develop a basic control system for the inverter This will allow the user to specify a sequence of steady state inputs to be fed to the motor and the duration of each state so that response
56. ch describes the torque produced by the rotor as a function of wind speed using Eq 3 and a dynamic element that describes the torque oscillations produced by wind shear and tower shadow using Eq 7 and either Eq 5 or Eq 6 The static element requires an understanding of the turbine s performance at different wind speeds As noted by Teodorescu et al 2003 the distribution of torque coefficient with tip speed ratio is more useful in this case than the power coefficient The torque speed characteristic can be obtained from test data where available or derived from the turbine manufacturer s published data Yaoqin Jia et al 2007 for example provide a procedure for deriving Co from the parameters shown in Table 4 39 Symbol Quantity R Rotor radius m J Inertia kg m3 PW Rated power kW wW Rated rotating speed rpm VW Rated wind speed m s wM Maximum rotating speed rpm VS Startup wind speed m s VIN Cut in wind speed m s VOUT Cut out wind speed m s TS Startup torque Nm Table 4 Wind turbine parameters Yaoqin Jia et al 2007 To compensate for the difference in inertia between the rotor and the electrical drive Kojabadi and Chang 2011 develop Eq 11 based on Eq 4 above Given a description of the mechanical torque the only additional piece of information required is the generator shaft speed as noted by Lopes et al 2005 Q Qmech LU eter AW genera
57. cting control programs using the Weg ladder logic authoring tool and uploading them directly to the internal CPU would appear to be an obvious and efficient approach however it would also have introduced difficulties in terms of the integration of data from the HUT as well as making the development of the HMI more complex In addition the company brief specifies Microsoft CH as the development language C programs running on a PC have several theoretical drawbacks such as runtime interpretation of bytecode and communications and operating system overheads However the additional overhead was not deemed significant for this first stage project given a sufficiently powerful host PC and a dedicated communications link The operation of the inverter is controlled by setting the values of around 900 operating parameters The programming manual Weg 2010 describes their use and range of values but in summary there are essentially three categories e static parameters which need to be configured in order for the inverter to perform in the particular hardware configuration and whose values remain constant e dynamic parameters which must be constantly monitored and adjusted by the control program to achieve the required behaviour e read only parameters that provide information about the state of the machine Only some of the available parameters were relevant to this project and much of the early familiarisation was concerned with identifyin
58. d expensive components The industry is also developing very quickly with larger models and new designs frequently being produced Efficient methods are therefore required for the evaluation of existing design performance and the experimental testing of new designs Although this is true for most industries wind turbine designers and manufacturers are also faced with the additional problem that turbines are typically installed in locations that are difficult to access thus making in situ evaluation difficult Static models of turbine behaviour however cannot capture the real wind regime at a turbine location or the short term dynamics of drive train components reacting to oscillating torque resonance and other operational phenomena Rabelo et al 2004 This dilemma has led several teams of researchers to propose a range of turbine emulators that can be used to recreate realistic loads on a turbine in a laboratory situation Different motivations have been invoked including those shown in Table 3 34 Motivation References Development of control algorithms and Teodorescu et al 2003 techniques Chinchilla et al 2004 Kojabadi et al 2004 Rabelo et al 2004 Moore and Ekanayake 2010 Munteanu et al 2010b Analysis of the dynamics of drive train Kojabadi et al 2004 components Helsen et al 2010 Investigation of fault conditions Mauri et al 2008 Moore and Ekanayake 2010 Condition monitoring and fault diagnosis Yang e
59. d response variables as properties of the protocol class Currently these are simply program variables declared within the inverter or turbine classes The selection of Microsoft C was made for practical reasons of maintainability Although this is acceptable for this first stage project the additional overheads incurred by running a C application on a PC might be significant when trying to replicate a wind regime accurately It would be useful at some future date to compare the performance of this application with an equivalent written in ladder logic and uploaded directly to the onboard programmable logic controller PLC of the inverter Such a solution might still require the creation of an external HMI however An intermediate solution might recreate the current application in a compiled language rather than one which relies on runtime interpretation of bytecode A further alternative design would be to make use of the customisation features of Microsoft Excel Because Excel already has a very robust user interface and provides for easy input output and manipulation of data the features of the current development could be added as part of an Excel based application written in C This would retain the maintainability of the software but would simplify the interface and eliminate the requirement for writing external Excel log files 4 3 Test results Although the evaluation tests had been fully specified a number of problems have
60. d velocity at reference height h0 m s Wind velocity a height h m s Width of tower shadow Linear distances m Aerofoil zero lift line Ground roughness length Angle of attack rad Blade pitch angle rad Maximum airstream velocity deficit Efficiency Drive efficiency Drive gearbox efficiency Generator efficiency Generator gearbox efficiency Efficiency factor for unknown losses tip speed ratio Density of air 1 225kg m3 at sea level Wind direction relative to rotor plane rad Azimuthal turbine blade angle rad Angular velocity of turbine rotor rad s Angular velocity of drive train component rad s Angular velocity of generator rotor rad s Angular velocity of EM rotor rad s Angular velocity of stator magnetic field rad s Abbreviations Units AC CPU CRC DC DFIG EM EMI FITS FOC HAWT HIL HMI HSS HUT IEC IGBT LSS NMEA NREL PC PHIL PI PID PLC PMSM PWM SCIM STP USB Vi VFD VVW WLP WTE Alternating current A Central processing unit kWh Cyclic redundancy check m Direct current N Doubly fed induction generator rpm Electromechanical machine s Electromagnetic interference V Feed in tariff scheme W Field oriented control Horizontal axis wind turbine Hardware in the loop Human machine interface High speed shaft Hardware under test International Electrotechnical Commission Insulated gate bipolar transistor Low speed shaft National Marine Electronics Association National Ren
61. development period am also grateful to Jonnie Andringa Gaia Wind CEO for allowing me to spend time with the company would also like to thank my supervisor Tom Grassie and the Module Leader Alan Edgar for their flexible approach to administrative processes during my extended negotiations with possible partner companies Last but not least would like to thank my partner Bridget Webster for picking up the domestic slack while was otherwise occupied Symbols and abbreviations Symbols A Swept area of turbine rotor m2 CP Power coefficient CO Torque coefficient D Aerodynamic drag force d Direct component of stator current A f Distribution network frequency Hz generator Moment of inertia of generator rotor kg m2 Jrotor Moment of inertia of turbine rotor kg m2 L Aerodynamic lift force n Gear ratio n 1 N Number of poles in EM ns Synchronous speed P Power W Q Torque Nm q Quadrature component of stator current A Qoorr Corrected torque Qarive Torque associated with EM drive Nm Qerr Torque error Qgenerator Torque associated with generator Nm Qish Mechanical torque Qef Reference torque Qerotor Torque associated with turbine rotor Nm R Radius of turbine rotor blade length m S Slip T Tower diameter m U Uninterrupted wind speed m s Vo Vh X Y Z z0 gt D Na Ndgb Ng Nggb E DG 90 D gt Wgenerator wr WS Actual velocity of air stream m s Win
62. dynamics of the blades to the software that controls their operation is paramount Designers need appropriate tools and resources to experiment with innovations and to evaluate the effect of modifications Because wind resource is typically most abundant in remote and inaccessible places monitoring and testing in situ is difficult Simulation offers a solution to this problem but although software simulation is becoming quite sophisticated building an adequate software model can be very time consuming and expensive An alternative is to use the hardware in the loop HIL approach in which real or scaled down turbine hardware is used with a simulated wind resource Such a system can be referred to as a wind turbine emulator WTE This project focuses on the requirements for an effective WTE through an exploration of the relevant background literature and a practical development exercise 1 2 Aims and objectives The aim of this project is to establish a basic control system for a WTE In order to do this the following objectives will be addressed e A review of current relevant literature e An investigation of the capabilities of a specific hardware configuration e The specification and development of a software application to provide a human machine interface HMI and communications with the hardware components e An examination of the limitations of the developed application e The definition of a series of future projects that build up
63. e it is the drive train that is of most interest in terms of dynamic behaviour For larger turbines where blade flexibility is significant Anaya Lara et al 2009 propose a three mass model of the drive train which includes the generator hub and blades as independent masses connected by torsional springs However they derive an effective two mass 22 system similar to that presented by Hansen et al 2003 which is sufficient for most studies In the two mass system shown in Figure 8 a large rotating mass representing the rotor is connected to a smaller mass representing the generator via a low speed shaft an ideal 1 n gearbox and a high speed shaft The moments of inertia of the shafts and gearbox are deemed negligible Some authors go a step further and represent the turbine as a single mass system by ignoring the dynamic characteristics of both shafts Neammanee et al 2007 Munteanu et al 2010 Kojabadi and Chang 2011 In this case the dynamic behaviour of the turbine can be represented by dW generator Qmech rotor fjera dt Qgenerator n n where Qmech is mechanical torque nis the gear ratio Jrotor IS the rotor inertia UgeneratoriS the generator inertia Wgenerator iS the rotational speed of the generator Qenerator 1S the generator torque Kojabadi and Chang 2011 Wgenerator generator speed r shaft Gearbox Qgeneratot High speed i shaft Jrotor Rotor inertia 8 Jgenerator Gen
64. e Accelerate up to 1010 rpm over 2s then spend 60 seconds at speed 1 5 10 1050 60 Enable Drive Accelerate up to 1050 rpm over 5s then spend 60 seconds at speed 0 10 10 0 0 A Stop the drive and disable decelerate to zero over 10 seconds You should also be able to set a sample period from 1 300s This can be part of the script or part of the UI Every time this period elapses a data entry should be made in an XL or Tab delimited file TI RPM kW Rpm kW 8 5 2011 1020 1010 5 5 6 5 100 1010 1020 100 8 45 86 Appendix C Project proposals Gaia Wind proposal version 1 Student Name Title Summary of work to be undertaken during project Deliverables Brian Davison Matriculation 02014147 Number Simulation of wind regime and wind turbine response in a physical test rig The Gaia Wind 11kW is a two blade fixed pitch fixed speed turbine for small applications Gaia Wind Ltd is currently developing a test rig which consists of two turbines without rotors connected so that one can be used as a motor to drive the other The turbine acting as a motor is fed by a Weg CFW 11 inverter which has the facility for programmatic control The ultimate goal is to be able to provide a file of logged wind data and for the inverter to accurately recreate the wind forces via the motor so that the behaviour of the passive turbine can be observed The main challenges are therefore a To produce an accurate model of real tu
65. e a back to back inverter In this device an alternating current AC of arbitrary frequency is first converted to a direct current DC and then into an alternating current which matches the frequency of the network as shown in Figure 11 Gearbox ee aaa ey A a ke Grid Synchronous Generator Figure 11 Inline adjustable speed generator M ller et al 2002 With this arrangement any rotational fluctuations in the power output can be smoothed by the frequency converter and an induction EM is no longer required However when the frequency converter is in line with the generator it has to carry the full power rating of the system which makes it expensive Miller et al 2002 An alternative arrangement shown in Figure 12 is to use a doubly fed induction generator DFIG Here the rotor is of the wound construction rather than the simpler squirrel cage design and is fed with a variable voltage supplied by the frequency converter Control of the rotor current compensates for variations in speed and the output can remain at the required frequency The main advantage over the in line design is that the frequency converter only needs to carry around 25 of the total system power and is therefore significantly cheaper M ller et al 2002 30 Gearbox Grid Figure 12 Adjustable speed generator with DFIG M ller et al 2002 As discussed earlier more complex turbine designs which include yaw drives or active pitch control
66. e costing around 40000 and operating at a modest 15 capacity factor the FITS payments would mean a financial payback period of approximately 10 years This makes small scale wind generation an attractive investment for small landowners such as farmers Table 1 summarises RenewableUk s classification of turbine sizes and the corresponding FITS rates for a single turbine Note that the FITS scheme has additional bands with lower rates which allow for the installation of multiple turbines The largest capacity allowed under the scheme is 1 5 MW for which the rate is 4 7 p kWh Category Rated power kW Installed cost k FITS rate p kWh Micro 0 1 5 0 5 5 36 2 Small 1 5 15 2 50 28 Small medium 15 100 50 250 25 3 Table 1 Turbine categories and corresponding FITS rates RenewableUK 2011b Ofgem 2011 2 2 Turbine design issues The effective design of wind turbines relies on detailed knowledge of several distinct subject areas which are briefly summarised in the sections below 2 2 1 Conceptual design All wind turbines operate by using a proportion of the kinetic energy in an air flow to develop torque in a shaft Early European windmills drew upon the seafaring experience of the local cultures and used sails to drive the shaft Although arrangements with many sails were possible the typical design was based on four sails to simplify the construction of the supporting frame Burton et al 2001 p 340
67. e design which set the content requirements of the rest of Chapter 2 From several points of view therefore the literature review has been successfully delivered There are two main ways in which the literature review could be improved from its current version Firstly a more consistent treatment of the central theme of the replacement of turbine rotor torque by a simulated prime mover would improve the focus Currently these issues are discussed but not tied together as well as they could be 80 The second improvement that could be made would be to draw the relevant equations together into a single consistent formula for WTE control Eq 3 and Eq 11 form the basis of such a model but the final combination was not presented This could be done relatively easily but would require more time that is available to the academic project On the other hand since the current development does not include a turbine model such a synthesis can legitimately be left to the follow up development 6 1 2 Hardware investigation The capabilities of the hardware components of the Gaia Wind WTE were successfully characterised as demonstrated by the operation of the software This investigation was necessarily pragmatic in that it focussed on the requirements of this particular project rather than aiming to extract an optimum configuration Such an approach might for example have led to the use of sensoreless vector control rather than V f control bu
68. e up to 1050 rom over 5s then spend 60 seconds at speed Decelerate to zero over 10 seconds Stop the drive and disable lt GaiaSpeedTest gt lt id gt 15 lt id gt lt description gt A sample test lt description gt lt instruction gt lt action gt Start lt action gt lt instruction gt lt instruction gt lt action gt Acc lt action gt lt duration gt 10 lt duration gt lt rpm gt 1000 lt rpm gt lt instruction gt lt instruction gt 109 lt action gt Con lt action gt lt duration gt 60 lt duration gt lt rpm gt 1000 lt rpm gt lt instruction gt lt instruction gt lt action gt Acc lt action gt lt duration gt 2 lt duration gt lt rpm gt 1010 lt rpm gt lt instruction gt lt instruction gt lt action gt Con lt action gt lt duration gt 60 lt duration gt lt rpm gt 1010 lt rpm gt lt instruction gt lt instruction gt lt action gt Acc lt action gt lt duration gt 5 lt duration gt lt rpm gt 1050 lt rpm gt lt instruction gt lt instruction gt lt action gt Con lt action gt lt duration gt 60 lt duration gt lt rpm gt 1050 lt rpm gt lt instruction gt lt instruction gt lt action gt Dec lt action gt lt duration gt 10 lt duration gt lt rpm gt 0 lt rpm gt lt instruction gt lt instruction gt lt action gt Stop lt action gt lt instruction gt lt GaiaSpeedTest gt 110 5 Manual mode File Test Log Help Auto Actual par
69. ed loop vector control 4 1 6 Summary There are some features of the Gaia hardware configuration that are likely to introduce inefficiencies into the WTE operation which will have to be compensated 63 for This is a good motivation for attempting to quantify the overall efficiency of the system in the evaluation The theoretical or quoted efficiencies of the gearbox and SCIM could provide a way of estimating the expected overall efficiency of the WTE 4 2 Software application The main product from this project is the software application which provides an HMI for the WTE hardware Its design is discussed in 3 8 and this section describes the final version and discusses its limitations 4 2 1 Overview The design which was produced in advance of the on site development period proved to be appropriate with no changes being requested by the company The original plan for development needed to be changed because the WTE hardware was not accessible during the first week The approach adopted was therefore a top down process which began with the HMI itself with the two communications modules being developed later While access to the hardware was restricted communications output was viewed in a console window This allowed communication processes to be developed step by step according to the design As greater detail was included in the communications the structure of the messages shown in the console window could be compared for examp
70. egree of camber in the aerofoil cross section Although it is the aerodynamic construction of the blades that determines the torque they generate the mathematical complexity involved is considerable In practice the torque developed by a rotor is typically calculated on the basis of empirical measurements 2 2 3 Materials The choice of materials for a machine of any kind is driven by several potentially conflicting goals The performance of the machine should be maximised which in the case of a wind turbine means for example that the blades should be as light as possible At the same time it should be robust enough to resist damage under extreme conditions Burton et al 2001 p 377 A trade off is therefore required to balance these requirements In the case of the rotor this has a significant effect on the behaviour of the machine The rotor is subject to a range of forces some of which are required for its successful operation Other forces such as the component of the aerodynamic force perpendicular to the plane of rotation are actually undesirable but nevertheless have to be accommodated In general there are two significant categories of forces that need to be taken into account bending moments which arise from the long thin shape of the typical rotor and fatigue forces which arise from typically small but persistent variations in operating conditions A rotor of 19 light construction has a low moment of inertia and is better
71. elivery of prototypes Three main prototypes were envisaged corresponding to the three main modules of the system The initial intention was to complete the inverter interface first followed by the turbine controller interface and then the HMI This is shown in the project plan in Appendix D In fact at the beginning of the development period the plan had to be changed because the hardware was in use for other purposes Consequently a top down approach was taken and the HMI was addressed first followed by the inverter interface and the turbine controller interface In general the development plan was successful and a main prototype was completed at the end of each of the three weeks Regular backups were taken throughout the development period Because of the agile approach the system specification was continually updated throughout the development period to reflect the evolving design The final version can be found in Appendix K 3 8 System design Conceptually the software system consists of four main components as shown in Figure 21 Two interfaces are provided one to the VFD via the Weg inverter and another to the controller of the turbine under test The operation of these pieces of equipment are summarised in 4 1 The test management module concerns the construction and operation of scripted tests while the HMI provides the means for 3 See for example http en wikipedia org wiki Agile software development 49 the
72. ems an error in the efficiency calculation was discovered The software inserts an Excel formula into the Efficiency column in the log rather than a simple numerical value It was found that the Excel ROUND function was applied at the wrong point leading to all decimal places being lost Error in the slip calculation The slip calculation used at the point that this data was logged was that supplied in the original company brief In fact that calculation incorrectly used the drive angular velocity rather than the synchronous speed to calculate generator slip This issue was raised with the Gaia engineer who agreed the correction 10 http en wikipedia org wiki Two s_ complement 74 5 Forward plan This project was intended as the first stage of a longer development Those aspects of the fully featured WTE which were not included here naturally provide the basis for further work In addition the practical work of this project has revealed the potential of the hardware components and the limitations of the present approach This section presents a series of potential projects that could follow on from the work done here 5 1 Limitations of current implementation As agreed with the company during the scoping of the current project the main limitation is the absence of a turbine model This means that the range of tests and simulations is limited to those involving a steady generator speed reference This is adequate for benchmark tes
73. en ergy 20mix renewable 20energy ored 25 nat ren energy action plan pdf Accessed 23 July 2011 DECC 2011 Feed in tariffs FITS Available online at http www decc gov uk en content cms meeting energy renewable ener feedin_tarif f feedin_tariff aspx Accessed 23 July 2011 Diop A D Ceanga E Retiveau J L Methot J F llinca A 2007 Real time three dimensional wind simulation for windmill rig tests Renewable Energy Volume 32 Issue 13 October 2007 Pages 2268 2290 ISSN 0960 1481 DOI 10 1016 j renene 2006 04 011 DNV Ris 2002 Guidelines for the Design of Wind Turbines Second Edition Ris National Laboratory ISBN 87 550 2870 5 Dolan D S L Lehn P W 2005 Real Time Wind Turbine Emulator Suitable for Power Quality and Dynamic Control Studies International Conference on Power 122 Systems Transients IPST 05 Montreal Canada June 19 23 2005 Paper No IPST05 074 Dolan D S L Lehn P W 2006 Simulation model of wind turbine 3p torque oscillations due to wind shear and tower shadow IEEE transactions on energy conversion 21 3 Fleming F Edrington C S Steurer M Vodyakho O 2009 Development and implementation of a 25 kW virtual induction machine test bed utilizing the power hardware in the loop concept Electric Machines and Drives Conference 2009 IEMDC 09 IEEE International vol no pp 1161 1166 3 6 May 2009 doi 10 1109 IEMDC 2009 5075350 URL http ieeexplore ieee org sta
74. erator inertia Qrotor Rotor torque Qgenerator Generator torque Wgenerator Rotational speed of the generator Figure 8 2 mass model of turbine drive train AJ 2 2 5 Torque oscillations The majority of large turbines employ a design in which the rotor blades are oriented into the wind in front of the nacelle and tower by an electric motor known as the yaw drive The alternative arrangement is to place the rotor downwind of the nacelle so that the wind itself maintains the rotor in the ideal position This passive yaw control makes the yaw drive redundant The simplicity of the second option means reduced manufacturing cost reduced maintenance and lower power consumption than the first however the downwind arrangement is more susceptible to the phenomenon of tower shadow The tower presents an obstacle to the flow of air both upwind and downwind and as a rotor blade passes the tower it experiences a momentary drop in air flow The effect is much weaker in the upwind direction and can generally be ignored in electrical calculations Chan et al 1984 however it must still be taken into consideration in relation to blade loading Burton et al 2001 p 234 In the downwind direction the drop in velocity of the air flow due to tower shadow can be of the order of 20 and is therefore significant in both cases Chan et al 1984 Tower shadow can be reduced by placing the rotor further from the nacelle although the effect can still be si
75. eteet eet See et ee eege eg ee eet eer eet aere eegen 105 SS Men tem Siomin ro ao se iacnces ores sues aO Ren pees pietegiaseseen serene 105 AM ONGEN EE EE 106 BD NANGATON ageet edd 107 te eebe 108 4 4 XML netomat 109 bh Manual mode 111 Ball NOTES sikh tet She hee he seed dan See eebe 111 6 Test management module sts vised Seege Eer 111 7 inverter communication 112 TA ET EE 112 7 2 JInttalteaton ENE ecceeccecceceecceccueececceceaeeceecueeaeeceecaeeceecueeaeeceecaeeneeseeens 114 Z3 Start ACION 2 cece e cece cecceeccecceceeececceeeceeceecaueceecueeaueceecaeeseeceecaeeceeseeeaeeeeeeas 114 7 4 Acceleration ooo eee eec cece cecceeccecceccecceeceecceccueeaeecuecaeecuecuecaeecuececaeecsececseeaueeas 114 7 5 Deceleraion 115 7 6 Stop ACHOM eet eet ee eet 115 Leb Stat s lte DEE 116 7 8 Communications PIOlOCOlas c apo wae ee 116 8 Turbine controller COMMUNICATIONS eee cece eeeeecceeceeececeeeceeceeeeeneeeeeeeenneeness 116 9 RAUGIMONAl TEAUISS EE EELER 117 SCT Graphical OU MOU cscs eeh EE 117 9 2 Excel export beet Eeegee 117 Appendix L Communications regarding data collechon 118 1 Introduction 1 1 General Wind turbine design and development is currently a very dynamic industry With the increasing costs associated with traditional fuels the extraction of energy from a renewable resource like the wind is extremely attractive To maximise the value of such extraction the efficiency of turbine designs from the aero
76. ewable Energy Laboratory Personal computer Windows Power hardware in the loop Proportional integral Proportional integral derivative Programmable logic controller Permanent magnet synchronous machine Pulse width modulation Squirrel cage induction machine Shielded twisted pair Universal Serial Bus Voltage frequency Variable frequency drive Voltage Vector Weg Weg ladder programming Wind turbine emulator Ampere current kilowatt hour energy metre distance Newton force Rotations per minute second time Volt potential difference Watt power Contents AY UMMRO GU COM iettsestsesercctecdeteatseocersuetrectcestotersnetsecboeseonecanstserboeseeeneas ese band iaiia Eein 10 TT Ie ee CR Re Ree ee eae nee DATE eae eee Te ea eee eee eae ea 10 1 2 Aims ANG ODICCIVES E 10 1 3 Structure of report EE 11 2x EEN 12 Ca E le Re eu EE 12 2 2 Turbine design Ssues o 2 05 cdcacdededenieedadeatdededeatbeiadeaibededshbededeideiedeatbeeedenitee 14 2 2 1 Conceptual desen se ccd nectadende cc centec nated db cdendecnedandeanisenducnteendetes 14 Bree BIAOS COSI e ME 17 22 ENEE 19 2 2 4 Mechanical construction ccccccccceccceccceeeeeeeceeeeeeeceeeeeeeeeeeeeeeeeeeeeess 21 2220 orgue OSCIALONS ica tare ny tr as en S 24 2 26 Electrical SY StS EE 27 er EE EN 2 3 Turbine emulation aere ee 34 23 1 Drive SClECHON EE 37 2 0 2 Ee EE 39 2 3 3 Instrumentation and Control 40 29 4 EEGEN 42 2 4 Conelusion ee 4
77. f other safety systems fail If deployed the tip brakes need to be reset by an engineer The downwind design provides passive yaw control which eliminates the need for a yaw drive however that means that tower shadow is significant Gaia Wind Generator Vibration sensor speed Brake sensor Rotor hidden from view sensor brake is shown speed disconnected sensor Figure 25 Gaia Wind nacelle detail 57 The turbine is equipped with speed sensors for both the rotor and the generator as shown in Figure 25 A brake sensor and a vibration sensor provide information about the state of the machine primarily for safety control Not shown in Figure 25 but visible in Figure 22 is an anemometer which is positioned upwind of the nacelle to measure the instantaneous wind speed Measurement equipment in the control box whose interior is shown in Figure 26 provides instantaneous voltage and current readings Controller Operator RS 232 port keypad Communications unit for remote monitoring Current and voltage measurement Soft starter Capacitor bank Figure 26 Gaia Wind control box deiail Under normal conditions control of the turbine is provided by an 1C1000 processor unit supplied by Danish company Mita Teknik The controller takes input from the various sensors and generates control signals to the brake grid connection relays and soft starter In particular the controller is responsible for initiating the connectio
78. f the status polling message is used to update the feedback panel in the HMI The turbine controller uses the proprietary M NET protocol from Mita Teknik see Appendix I Messages are prepared in a similar way to those for the inverter A main difference between the two protocols is that M NET has specific start and end markers that encapsulate the entire message including the two bytes of cyclic redundancy check CRC The communications modules were based on existing applications developed by Gaia Wind These were largely re worked but provided useful models In particular they provided ready made solutions to the CRC calculations for both protocols During the development of the communications modules two tools in particular proved useful ModbusMat is a free application for managing and monitoring Modbus communications and Free Serial Port Monitor displays the detail of outgoing and incoming messages over any serial port A major issue was discovered when trying to run the application with both inverter and turbine connected The issue was that the RS 232 communications to both the inverter and the turbine controller were disrupted with the inverter running at low speeds The turbine communications were also disrupted with the inverter running at higher speeds to the extent that status polling was ineffective The problem was assumed to be the result either of electromagnetic interference EMI produced by r http Awww modbus org s
79. for a turbine with a fixed blade pitch see Figure 2 indicates that maximum efficiency is only achieved at a particular tip speed ratio By adjusting the pitch angle however a turbine can be made to maintain its rated power output at wind speeds higher than rated Burton et al 2001 p 181 Active pitch regulation requires that additional mechanical features to allow each blade to rotate about its own longitudinal axis This necessarily includes a shaft and bearings for each blade but the selection of individual or central actuator is a function of detailed turbine design Burton et al 2001 p 351 21 Rotor Disk brake Speed multiplier Electric generator Low speed shaft High speed shaft Hub Coupling Yaw bearing Junctions Blade Tower Mainframe D HAWT Figure 7 Mechanical construction of horizontal axis wind turbine Munteanu et al 2008 A teeter hinge can also be added to the turbine design in order to limit undesirable loads on blades low speed shaft nacelle and yaw mechanism Burton et al 2001 p 347 Most often used on two bladed turbines a teeter hinge allows the rotor to pivot by a small amount at the hub This allows small out of plane moments periodic forces arising from the rotation of the blades and random forces due to turbulent air flow to be absorbed rather than being transferred to the main structural components Despite the wide range of mechanical considerations illustrated abov
80. for example require additional electric motors These are typically fed directly from the distribution network and do not therefore affect the operation of the drive train components 2 2 7 Control The main purpose of control is to ensure that a complex technical process responds in a desired manner to operational changes and external disturbances both to protect the plant and to optimise its operation Schleicher and Blasinger 2003 Munteanu et al 2008 p Wind turbine control can be subdivided into three separate categories Burton et al 2001 p 472 The first is supervisory control in which an operator brings the turbine into a particular operating state The second is closed loop control in which the adjustment of operating parameters is controlled automatically by a digital controller in response to sensor readings and the third concerns mechanisms for dealing with potentially dangerous or fault conditions such as excessive tower vibration Although safety control also depends on sensor readings it differs from closed loop control in that it typically brings the machine into an exceptional fail safe state rather than one that is associated with normal operation 31 In the context of wind turbine design it is the external changes in airflow which drive the majority of control actions Munteanu et al 2008 p 5 identify three distinct types of control system in wind turbines e Aerodynamic power control e Generator cont
81. g and resolving future problems has been established with the company A good deal of subjective evaluation of the work of the project has been presented at various stages of the report The objective of evaluation of results can therefore be considered partially successful 6 1 5 Definition of programme of future work 0 sets out three types of project that could follow from this one with several examples under each heading Clearly the most important of these is to continue the development to the planned second stage with the inclusion of a turbine model Even this however can be broken down into smaller steps providing the basis for a number of development projects The other two categories of project involve the development of comparable systems using other platforms and tools and the development of facilities to support the future developments themselves 6 2 Personal reflection To attempt an MSc dissertation project of this type over a single semester was an ambitious undertaking even given the reduced scope of this initial project There were many unknowns at the outset which included not only the details of the turbine subsystems but also the NET development environment The knowledge gaps were largely filled simply by putting in long hours but this approach has been reasonably successful Setting the scope of the project to its current definition was a deciding factor in the success of the project and the author was gu
82. g them The complete list can be found in the software system specification in Appendix K and Table 7 provides some 61 selected examples It should be noted that the value of P0010 is the calculated output from the VFD Parameter Description Purpose P0402 Motor rated speed One of a set of static parameters used to characterise the motor P0312 Serial protocol One of a set of static parameters used to configure the serial communications channel P0683 Serial speed ref Dynamic parameter used to set the required drive speed P0100 Acceleration time Dynamic parameter used to ramp the speed over a period of time P0010 Output power Read only parameter Table 7 Example inverter parameters 4 1 5 VFD control The CFW 11 unit offers four strategies for the control of a VFD the simplest of which is known as V f control In this mode the inverter maintains a constant ratio of voltage to frequency The developed torque is directly proportional to this ratio and therefore a steady torque is delivered over a wide range of motor speeds Parekh 2004 The resulting torque speed characteristic shown in Figure 28 is somewhat different from that of the directly supplied induction machine in Figure 10 It is clear from the figure that V f control is not ideal for a WTE whose purpose is specifically to control the developed torque to replicate a wind regime however V f control is appropriate when the requirement is to
83. gnificant at a distance equivalent to four tower diameters Burton et al 2001 p 373 A lattice tower can be used rather than a tubular one to reduce the effect Burton et al 2001 p 233 although this approach can make the remaining effect dependent on wind direction Burley et al 1979 cited by Chan et al 1984 A number of studies exist that deal with tower shadow in upwind three bladed turbines eg S rensen et al 2002 Dolan and Lehn 2006 and several mathematical models are available of varying complexity however the potential flow model used in Garrad Hassan s modelling application GL BLADED can equally well be applied to two bladed machines Bossanyi and Quarton 2003 In this model a 120 sector of the rotor swept area is deemed to be influenced by tower shadow as shown in Figure 9 The wind speed at a point within that zone of influence is given by Eq 5 24 v u 1 5 G3 where Uis the uninterrupted wind speed T is the tower diameter x is the lateral displacement of the point from the tower centre line Zis the minimum distance between the plane of the rotor and the centre of the tower Bossanyi and Quarton 2003 Figure 9 Tower shadow When a blade is within 60 of top dead centre it is assumed that V U For all other rotor positions the factor in brackets in Eq 5 is modified to F 0 5 cos 0 5 cos where F represents the bracketed term in Eq 5 and is the blade azimuthal angle
84. he nominal rotor speed which corresponds to the speed required to produce 11kW at rated wind speed This can be verified by applying Eq 2 using the design tip speed ratio of 4 the rated wind speed of 9 5 m s and the rotor radius of 6 5 m aa OR U oa R g x 9 5 65 Q 5 846 rad s Converting to rpm 5 846 x 60 Q _ 21 Q 55 83 rpm 95 The synchronous generator speed of 1200 rpm on a 60 Hz grid corresponds to a rotor angular velocity of 55 55 given the gear ratio of 1 21 6 The rotor must reach this speed in order to generate power rather than draw it from the grid From the NREL data this occurs at a wind speed between cut in at 3 5 m s and 4 m s Thus the rotor angular velocity is assumed to increase steadily from 55 55 rom to 55 83 rpm between cut in speed and rated speed Thereafter up to cut out speed the angular velocity is assumed to increase steadily to 57 rom at a wind speed of 20 m s This is lower than the maximum rotor speeds suggested by the NREL reports however it was noted that the highest wind speed in the NREL dataset was 20 m s whereas the cut out speed is quoted as 25 m s This estimated high bound for rotor speed was checked with the Gaia Wind engineer The interpolation was performed using Microsoft Excel The result which is tabulated in Appendix G yields the graph in the figure below when the power coefficient is plotted against high speed shaft HSS angular velocity Power cur
85. hnology Inc Available online at http ww1 microchip com downloads en AppNotes 00887a pdf accessed 30 July 2011 Parekh R 2004 VF Control of 3 Phase Induction Motors Using PIC16F7X7 Microcontroller Application note AN889 Microchip Technology Inc Available online at http Aww1 microchip com downloads en AppNotes 00889b paf Pena R Cardenas R Blasco R Asher G Clare J 2001 A cage induction generator using back to back PWM converters for variable speed grid connected wind energy system In Industrial Electronics Society 2001 IECON OI The 27th Annual Conference of the IEEE Vol 2 2001 pp 1376 1381 Pillay P And Krishnan R 1988 Modeling of Permanent Magnet Motor Drives IEEE Transactions on industrial electronics 35 4 Powles S R J 1983 The effects of tower shadow on the dynamics of a HAWT Wind engineering 7 1 Rabelo B Hofmann W Gl ck M 2004 Emulation of the static and dynamic behavior of a wind turbine with a DC machine drive PESC 04 Aachen Germany 2004 pp 2107 21 RenewableUK 2011 International comparisons turbine densities and capacity factors Available online at http www bwea com pdf publications RenewableUK Turbine Density Study pdf Accessed 23 July 2011 RenewableUK 2011b Small wind systems UK market report Available online at http www bwea com pdf publications 8942 Report WEB pdf Accessed 23 July 2011 Schleicher M and Blasinger F 2003 Control engineering
86. ich helps to provide rigidity and resistance to buckling Variations on this general design include a larger number of internal webs and variations in shell materials between the leading and trailing edges For other turbine components the choice of materials is more limited The supporting tower for example could be made of steel or concrete but steel is almost invariably preferred because of the logistical associated with transporting large concrete members Burton et al 2001 p 453 20 EE EE W An _ Glass Epoxy pecht Ce SE Ee E 25mm Wood veneers S LS Epoxy glue LS SS Aluminium screen for lightening protection N Glass Epoxy unusual Polyurethane point Figure 6 Turbine blade construction Burton et al 2001 p 379 2 2 4 Mechanical construction The main mechanical components of a turbine are similar despite the wide range of detailed designs and are illustrated schematically in Figure 7 The general function of these elements is largely self explanatory The components in Figure 7 make up the turbine drive train which is responsible for transferring the torque produced by the blades to the rotor of an electrical generator Major design decisions such as the position of the rotor and the speed and control strategies significantly affect the actual arrangement of the elements of the drive train and the need for other mechanical components The characteristic shape of the Cp A curve
87. ided strongly in this by the Gaia Wind engineer The immediate requirements of the company were relatively straightforward and at the basic level constituted a fairly standard software development project Some of the material covered in the literature review is not directly relevant to this and from that point of view less effort could have been expended in that area However a deep investigation into previous WTE work has 82 eventually delivered a much richer background to the current project This is valuable not only to the author personally but also to potential follow up projects There was an element of risk involved with the chosen project selection process which was dependent on receiving replies for the companies that were approached A back up plan was prepared however which meant that the risk was adequately managed Ideally the project should have started earlier than it actually did but the negotiation with possible partner organisations took longer than expected This was another factor that contributed to the intensity required during the development period With this one drawback the approach was very successful and has led to the establishment of contacts with a series of companies that may lead to other student projects in the future The weakest part of the project is the evaluation which had to be curtailed for a number of reasons The author struggled for some time to formulate a reasonable evaluation method for the
88. ing an opto isolator on the inverter serial cable however did appear to solve the problem but only when using a PC with a native serial port When using an RS 232 to USB adapter communications were improved but still not completely reliable This suggests that the problem was due to a ground loop or similar effect but also that the USB adapters might be part of the problem 69 Solution Opto isolated Notes Using an opto isolator would break the electrical circuit between the PC RS 232 and the other equipment This would guard against ground loops but probably not EMI which may be inducing spikes in the cable directly NMEA Specifically designed for noisy environments the National Marine Electronics Association communication standard uses special equipment which appears very robust but is also very expensive RS232 over fibre Rules out induced or conducted interference completely however equipment is expensive RS232 current loop Uses current pulses rather than voltage modulation to carry the signal and therefore less susceptible to spikes Not as expensive as the previous two options but would require some work to make up the required cables RS485 Very feasible in the long term for the inverter but would require a new interface module which would increase the expense Communications with the turbine controller would require RS 232 to RS 485 adapters RS 232 over STP Shielded twisted pai
89. ions which remain to be investigated It is likely that these will rely on code already developed by Gaia Wind 102 1 2 Manual and automatic operation The design provides a tabbed layout with one tab for manual operation and another for automatic The manual tab allows the user to control the inverter directly by setting an acceleration and deceleration ramp time and a speed Two buttons enable and disable the converter If a non zero speed is set before the inverter is enabled the acceleration ramp time is used Likewise if the speed is non zero when the disable button is clicked the deceleration ramp time is used If the drive is already running and the speed value is changed the instruction is sent directly to the inverter Although the ramp settings are still active their effect in this case will be negligible On the automatic tab the user is able to enter a sequence of instructions directly into the interface and run Instruction lists can be as long or as short as the user requires which allows the user to build up a sequence of instructions to suit immediate purposes The current set of instructions can then optionally be saved to the internal database for later re use or exported to an external file in XML format The use of an internal database is intended to simplify the user interface development and to provide better management for the tests themselves and for any logfiles that are generated Specifically there may be many
90. itney Bowses Management Services Denmark A S Denmark ISBN 87 550 3198 6 Helsen J Vanhollebeke F De Coninck F Vandepitte D Desmet W 2011 Insights in wind turbine drive train dynamics gathered by validating advanced models on a newly developed 13 2 MW dynamically controlled test rig Mechatronics Volume 21 Issue 4 pp 737 752 DOI 10 1016 j mechatronics 2010 11 005 Hsu Wen Ko 2010 Measurements on a wind turbine condition monitoring test rig University of Durham MSc dissertation Huskey A Bowen A Jager D 2009 Wind turbine generator system power performance test report for the Gaia Wind 11 kW wind turbine Technical Report NREL TP 500 46151 National Renewable Energy Laboratory Available online at http www google co uk url sa t amp source web amp cd 1 amp ved 0C DsQFjAA amp url http 3 A 2F 2Fwww nrel gov 2Fwind 2Fsmallwind 2Fpdfs 2Fgaia_ power performa nce_test_report pdf amp ei Qaw5 ToaAOlewhAeb2N20Ag amp usg AFQICNFV4hxMy CWj6qyBbDID6pWLjUoyA amp ssig2 wY U6MKVbShIgXD1icT Qlg Accessed 3 August 2011 Johnson K E Fleming P A 2011 Development implementation and testing of fault detection strategies on the National Wind Technology Center s controls advanced research turbines Mechatronics Volume 21 Issue 4 June 2011 Pages 728 736 ISSN 0957 4158 DOI 10 1016 j mechatronics 2010 11 010 Kojabadi H M Chang L Boutot T 2004 Development of a Novel Wind Turbine Simulator for Wind Energy Conversio
91. le with the protocol definitions to ensure that they were correctly formatted This approach proved effective with very few alterations being required once the system was connected to the hardware 4 2 2 HMI and general system features All of the main functions identified in the design were implemented within the three week development period Those minor issues that were not completed are described in 4 2 5 along with additional work that has been requested by the company in order to make the current implementation effective At the highest level of description the HMI provides two modes of operation In the automatic mode shown in the screen shot in Figure 29 the operator can define a series of steps for the system to follow labelled 1 in the figure Apart from the obligatory start and stop steps each has a target speed and duration The scripted test can be run from the toolbar or the menu 64 The HMI layout for manual mode is shown in Figure 30 This illustrates how the operator may directly control the VFD by altering the speed setting directly 1 The HMI has clearly visible start and stop controls 2 to minimise error and the animated indicator and summary calculations are the same as in the automatic mode Config Help Total KWh Wind speed m s Rotor speed pm Power kW Generator speed pm Mode State Errors Fae P0134 Speed pm Torque Power kW P0220 P0680 Pei P0682
92. ll the software commands to enable disable start and stop the inverter and to set the speed reference have any effect The LOCAL REMOTE setting determines where the inverter looks for its speed reference In the LOCAL condition it will used the keypad setting P0121 and in the REMOTE condition it will use the serial setting P0683 113 7 2 Initialisation Further communications parameters are set Description Value Meaning Defaul t Param P0133 Minimum speed ref minRPM Read from config table P0134 Maximum speed ref maxRPM Read from config table P0402 Motor rated speed 1000 Rated rpm 7 3 Start action The inverter is enabled but with minimum RPM This action also set any other required parameters for the specific test Param Description Value Meaning P0682 Serial control word 0x0000000000000011 Start and enable P0683 Serial speed ref minRPM Minimum rotation 7 4 Acceleration For acceleration actions the user specifies a duration value The inverter stores a ramp parameter which can be used to control the rate of acceleration The parameter value stored by the inverter is the length of time to accelerate from 0 to maximum rpm The duration of an acceleration event is therefore a function of current rpm target rpm maximum rpm and ramp The task is to set the value of ramp R that will produce the required time interval At for a particular speed interval Arpm and a given maximum speed rpMmax This is calcula
93. mp stamp jsp tp amp arnumber 5075350 amp isnumber 5075166 Gabriel R Leonhard W Nordby C J 1980 Field oriented control of a standard AC motor using microprocessors IEEE transactions on industry applications 1A 16 2 Gaia Wind 2008 User manual Gaia Wind 11kW turbine Available online at http www dcpower systems com documents Gaia WindUserManual pdf Accessed 14 August 2011 Gaia Wind 2009 Gaia Wind 133 11kW Data Sheet Available online at http www qgaia wind com index php download file 91 78 Accessed 12 August 2011 Gaia Wind 2009b Small wind turbine general information Available online at http sw ea co uk en images Gaia 20133 Covered 20Enquiry 20Supplementary 20Infor mation pdf Accessed 12 August 2011 Gamesa 2010 Press release available online at http www gamesa es en communication news eleven spanish companies join forces on the azimut project to develop a 15 mw offshore wind turbine using 100 spanish technology html idCategoria 0 amp fechaDesde amp especifica 0 amp texto amp fechaHasta Accessed 23 July 2011 123 Gregg J R 2011 Design and experimental testing of small scale wind turbines Baylor University MSc thesis Available online at http hdl handle net 2104 8142 Gross C A 2007 Electric machines CRC Press Boca Raton USA Hansen A D Jauch C S rensen P lov F Blaabjerg F 2003 Dynamic wind turbine models in power system simulation tool DIgSILENT Ris R 1400 P
94. n to the grid when the wind speed rises above the cut in speed of 3 5 m s At this time the controller briefly activates the soft starter which puts the SCIM into motoring mode to ensure that the rotor is rotating faster than the network synchronous speed In the WTE the HUT is connected to a standard controller so all usual sensor measurements are available with the exception of wind speed 58 The controller unit offers an RS 232 output port for connection to a PC The instantaneous sensor readings can be polled using the proprietary M NET protocol whose structure is summarised in Appendix Two versions of the M NET protocol are in use The older normal version only provides wind speed and output power The newer extended version however provides the following details e Wind speed m s e Rotor speed rpm e Generator speed rpm e Electrical output power kW e Cumulative energy production kWh e Status codes 4 1 3 SCIM The squirrel cage induction machine SCIM that performs the function of generator in the turbine and motor in the variable frequency drive VFD of the WTE is a Weg Industrias IE2 high efficiency unit whose datasheet can be found in Appendix E The slip values quoted on the datasheet are positive showing that the unit was designed to operate in motoring mode The question of whether its performance will be the same when operating as a generator can only be settled by experimentation which is
95. n Systems Using an Inverter Controlled Induction Motor IEEE IEEE Transactions on Energy Conversion Vol 19 No 3 September 2004 Kojabadi H M and Chang L 2011 Wind turbine simulators In Al Bahadly Ed Wind turbines InTech ISBN 978 953 307 221 0 124 Lee J Pillay P Harley R G 1984 D Q Reference Frames for the Simulation of Induction Motors Department of Electrical Engineering University of Natal Electric Power Systems Research Regular Papers 8 1984 85 Lopes L A C Lhuilier J Mukherjee A Khokhar ME 2005 A Wind Turbine Emulator that Represents the Dynamics of the Wind Turbine Rotor and Drive Train Power Electronics Specialists Conference 2005 PESC 05 IEEE 36th vol no pp 2092 2097 16 16 June 2005 doi 10 1109 PESC 2005 1581921 URL http ieeexplore ieee org stamp stamp jsp tp amp arnumber 1581921 amp isnumber 33408 Martinez F de Pablo S Herrero L C 2009 Fixed Pitch Wind Turbine Emulator using a DC Motor and a Series Resistor 13th European Conference on Power Electronics and Applications EPE 2009 Barcelona Mauri M Dezza F C Marchegiani G 2008 Hardware in the Loop HIL test bench for small scale Distributed Generation systems ndustrial Electronics 2008 ISIE 2008 IEEE International Symposium on vol no pp 2177 2182 June 30 2008 July 2 2008 doi 10 1109 ISIE 2008 4677136 URL http ieeexplore ieee org stamp stamp jsp tp amp arnumber 4677136 am
96. ncluded in the literature review The University library catalogue was used to identify appropriate textbooks for this purpose hence the reliance on Burton et al 2001 for much of the background on wind turbine technology This particular choice was justified by the frequent references to Burton et al in the research literature 45 Further reference material was identified during the literature search and review process by following up references in the papers themselves This led to the identification of several authors who filled important gaps in the background Two main approaches were taken to organise the identified references The first was to make paper copies of those papers relating to WTE studies and to highlight important passages These were later copied into an early draft of the literature review so that they could be incorporated into the text at the appropriate points The second approach was to collect electronic copies of all available references onto a PC This facilitated fast electronic search within the documents When setting out the structure of the literature review the outline view in Microsoft Word was used to develop short notes which were developed into complete sections as a second stage 3 4 Project scope The overall aim of the development was to create a general purpose WTE which could be used to recreate arbitrary wind conditions to test the response of the Gaia Wind turbine controller However it wa
97. nclusion of this feature makes the generation of test data a simple process The software automatically logs the following data to an Excel file at a location specified by the user which can then be used in the evaluation as explained in the next section e Timestamp e Drive rpm e Drive power e Generator rom e Generator power e Efficiency e Slip The full software system specification can be found in Appendix K 3 9 Evaluation of results The correct operation of the software compared to the specification can be verified through observation As with any piece of software it is to be expected that unforeseen errors will be identified at various stages through the development and after delivery However these are of less importance to the current project than the physical operation of the WTE itself and the accuracy of the information provided by the HMI The goals of the evaluation are therefore to answer the following questions e ls the behaviour of the HUT comparable to that of a real installation e Can the efficiency of the HUT be disaggregated from that of the other WTE components e What variation is evident in the performance of the WTE 3 9 1 Comparison with real installation In 2009 and 2010 NREL carried out a series of tests on the Gaia Wind turbine according to the International Electrotechnical Commission IEC standards The reports are now published and constitute a reliable description of the operation
98. nd premises is essential for any further development This bottleneck could be eased by the development of simulations of the major hardware components 78 5 4 1 Weg inverter simulation The current application defines a limited interface with the CFW 11 inverter through the manipulation of a subset of its available parameters The behaviour of the unit could however be simulated using some of the same methods as in the current project This would make it possible to develop the WTE software further in the laboratory with field testing on the Gaia hardware at a later stage An inverter simulation could take the form of a C application on a separate PC replicating the appropriate subset of inverter behaviour over a serial communications link 5 4 2 Other simulations Gaia Wind is currently investigating the use of a GenDrive back to back inverter to decouple the turbine generator from the distribution network in a departure from the Danish model A future simulated version of the inverter would allow parallel development of the turbine control system This raises the possibility of developing a library of equipment simulations along the lines described in 5 4 1 each of which could be the subject of a short project 13 http Awww gendrive co uk 79 6 Conclusions 6 1 Comparison with aims and objectives The overall aim of this project was achieved in full with the inclusion of all the major functions in the software system tha
99. nd speed Burton et al 2001 p 176 Thus it does not give any more information than the power coefficient but can be useful in determining instantaneous torque Q using Eq 3 1 3772 Teodorescu et al 2003 0 5 T 0 4 aal One blade Two blades Cp 8 Three blades Four blades 02 O Five blades 0 1 0 5 10 l 15 Figure 2 Effect of blade number on performance Burton et at p 175 A limitation of Burton s simple relationship between Cp and Ca is that it does not account for the existence of a non zero torque at standstill Teodorescu et al 2003 An alternative presented in the literature is produce a Cq A curve from test data 16 Lopes et al 2005 Weiwei Li et al 2007 Munteanu et a 2008 which leads to a polynomial expression for Co as a function of A In their derivation of Ca Yaogin Jia et al 2007 assume a value for startup torque as a given parameter while Weiwei Li et al 2007 also suggest using a lookup table which amounts to the same thing 2 2 2 Blade design It should be acknowledged from the outset that any brief consideration of the design of turbine blades requires many simplifying assumptions Issues such as the non uniformity of air flow in real situations the smaller aerodynamic effects that occur at specific points on a blade and the variation in blade geometry along its length are all worthy of exploration For a detailed look
100. ng W Tavner P J Crabtree C J and Wilkinson M 2010 Cost effective condition monitoring for wind turbines EEE transactions on industrial electronics 57 1 pp 263 271 Yaoqin Jia Zhaoan Wang Zhongqing Yang 2007 Experimental Study of Control Strategy for Wind Generation System Power Electronics Specialists Conference 2007 PESC 2007 IEEE vol no pp 1202 1207 17 21 June 2007 doi 10 1109 PESC 2007 4342164 URL http ieeexplore ieee org stamp stamp jsp tp amp arnumber 4342164 amp isnumber 4341940 Zimmermann S 2009 A simple C library for graph plotting Available online at http www codeproject com KB miscctrl GraphPlotting aspx Accessed 14 August 2011 129
101. nication needs to be selected and the communication management software needs to be written Part of the brief from the company is that this should take the form of an identifiable module within the overall software development 3 4 3 Establish communications with turbine controller The existing turbine controller is supplied by Danish company Mita Teknik and provides information on a range of operational parameters The achievement of this objective will depend on developing a familiarity with the communication protocol implemented by the controller and writing the communication management software Again the company brief specifies that this should be an identifiable module within the overall software application 3 4 4 Construct HMI The company brief specifies that the HMI should take the form of a PC application written in Microsoft Visual C The basic requirement is therefore to develop an interface which provides a means to control the hardware as specified and to log performance data during a test run which includes generator efficiency and generator slip This implies integration with both communications modules and some basic calculations based on the data that they supply The company brief can be found in Appendix B and the final functional specification in Appendix K The appropriateness of the selected development tools is also considered 3 4 5 Evaluate results The initial development is quite simple in comparison to the
102. ns with this approach e Because the current project does not include a turbine model the overall performance of the two cases cannot be directly compared e The American grid operates at 60 Hz and therefore a different gear ratio and model of generator are used Ideally the torque speed characteristics of the two generators could be compared to identify any major differences in behaviour Because the grid frequencies are different however the synchronous speed will be different in each case and two completely different models of generator have to be used The significant divergence between the two cases means that a direct comparison will have little validity A closer examination of the NREL data does however reveal some interesting aspects The NREL dataset provides power output and a power coefficient value for a range of wind speeds from 0 49 m s to 19 92 m s at intervals of roughly 0 5 m s The starting point for a meaningful comparison would be to generate a dataset from the WTE for this same wind speed range Because the WTE does not yet include a turbine model this cannot be done directly and the corresponding high speed shaft angular velocity must first be derived for each data point in the NREL data Thus the comparison would be limited to the performance of the generators A speed for the high speed shaft can be derived from the rotor speed Three NREL reports provide a range for the rotational speed for the turbine rotor unfor
103. of a real turbine The most relevant report is the result of the power performance test Huskey et al 2009 which includes binned wind speed data with corresponding 51 instantaneous power and power coefficient Unfortunately there are significant differences between the NREL installation and the WTE The primary reason is that the American distribution grid runs at 60 Hz while in Europe the grid frequency is 50 Hz This means that the synchronous speed of a 6 pole SCIM is different and therefore a different model of generator is used in the American version A further consequence of the different rotational speeds is that a different gear ratio is required Because the current version of the Gaia WTE does not include a turbine model a direct comparison with the NREL data is not possible however had the grid frequency not been different the torque speed characteristics of the generators in each case could have been compared to identify any divergences in behaviour Unfortunately this exercise would have little validity given the differences in the two cases Using the NREL data for comparison was therefore abandoned however in exploring ways of using that data some interesting observations were made and these are provided in Appendix H No other dataset for the European version of the turbine was available at the time of writing and therefore a comparison with a real installation was not feasible 3 9 2 WTE efficiency The HMI pr
104. olling The status of the inverter is polled at the rate specified in the sampling period field in the user interface Some parameter values need to be divided by 10 to get the actual value eg P0003 Any or all of the values in the table below can be retrieved Param Description P0009 Motor torque P0010 Output power Poo50 g9 Fault stack P0090 95 Last fault details P0316 Serial interface status P0680 Logic status P0681 Motor speed 7 8 Communications protocol Weg uses the industry standard Modbus protocol for communicating with devices Modbus is a general protocol which provides a wide range of communications features only some of which are implemented by Weg An even smaller range of messages are needed for this project 8 Turbine controller communications The Gaia Wind turbine provides a standard RS 232 communications channel over which a number of parameters can be monitored The communications are provided 116 by an 1C1000 controller system supplied by Mita Teknik who also defines a proprietary communications protocol called M NET This protocol has two versions normal and extended and older Gaia turbines make used of the normal version The system therefore needs to be able to handle both The major difference is that the extended version allows the polling of more specific information and the transmission of three commands start stop and reset whereas the normal version is entirely passive 9 Additional
105. on the current work 10 1 3 Structure of report Chapter 2 of this report summarises the main relevant aspects of wind turbine design and operation and ends with a review of previous WTE studies Chapter 3 discusses the methods used during the project and summarises some of the organisational processes Chapter 4 presents the results of the project work while chapter 5 outlines a series of potential follow on projects based on this work Chapter 6 summarises the conclusions from the project and includes a personal reflection on the project experience Appendix material follows and the list of references is positioned at the end for easy access 11 2 Literature review This section explores a range of aspects of wind turbine design that need to be taken into account when constructing a WTE The section ends with a review of previous work on WTEs 2 1 Wind energy context Although wind energy has been used for centuries to do useful work such as driving ships and grinding grain into flour it is only since the oil crisis in the 1970s that it has been considered as a serious candidate for electricity generation Gross 2007 Today energy security continues to be a major driver for the development of wind energy devices along with the abatement of airborne pollution from fossil fuel power stations and combating climate change Because of these pressures the market for wind turbines has shown over 20 growth every year since 1998 and total
106. ovides a calculated efficiency value n based on the relationship between the power output by the generator and the power delivered by the VFD according to Eq 12 12 Some power loss is associated with each energy conversion step in the WTE which means that the overall efficiency is the product of the individual efficiencies of the components This is expressed in Eq 13 N Na Nagb Nggb Ng Nu 13 52 where Na drive efficiency Nago drive gearbox efficiency Ng generator gearbox efficiency Ng generator efficiency Nu factor for unknown losses The datasheet for the SCIM Appendix E provides values for efficiency at different percentage loads Manually controlling the load to match the values provided reduces the theoretical uncertainty in the value of the SCIM efficiencies and any remaining inefficiency can be attributed to the other components Knowledge of the losses attributable to the WTE setup itself can be useful in interpreting the performance of the HUT The procedure for this test is to set the percentage load on the generator manually to 50 75 and 100 of rated using the power value from the turbine controller The WTE is run for 10s at each speed and the calculated efficiency values are logged every second The efficiency of the balance of system components gearboxes drive and unknowns is found by dividing by the quoted efficiency for the generator at these standard loads 3 9 3 Duration test
107. p isnumber 4676877 Ming Qiao Fei Lin Ruixiang Hao Xiaojie You Zheng T Q 2007 The Research and Development Platform for Wind Energy System Used Induction Motor Replacing Wind Turbine Industrial Electronics and Applications 2007 ICIEA 2007 2nd IEEE Conference on vol no pp 2579 2582 23 25 May 2007 doi 10 1109 ICIEA 2007 4318879 URL http ieeexplore ieee org stamp stamp jsp tp amp arnumber 4318879 amp isnumber 4318335 Monfared M Kojabadi H M Rastegar H Static and dynamic wind turbine simulator using a converter controlled dc motor Renewable Energy 33 2008 906 913 Moore l Ekanayake J 2010 Design and development of a hardware based wind turbine simulator Universities Power Engineering Conference UPEC 2010 45th International vol no pp 1 5 Aug 31 2010 Sept 3 2010 125 URL http ieeexplore ieee org stamp stamp jsp tp amp arnumber 5649 1 86 amp isnumber 5648789 M ller S Deicke M and De Doncker R W 2002 Doubly fed induction generator systems for wind turbines IEEE Industry Applications Magazine pp 26 33 May June 2002 Munteanu l Bratcu A I Andreica M Bacha S Roye D Guiraud J 2010 A new method of real time physical simulation of prime movers used in energy conversion chains Simulation Modelling Practice and Theory Volume 18 Issue 9 October 2010 Pages 1342 1354 ISSN 1569 190X DOI 10 1016 j simpat 2010 05 007 Munteanu l Bratcu A l Bacha
108. pecs php R http www ataytugal com ModbusMat htm 3 http www serial port monitor com index html 68 the inverter or the existence of ground loops when both devices were connected to the same PC Both of these conditions are notoriously difficult to diagnose RS 232 is more susceptible to EMI than some other protocols because it relies on a so called single ended signal Binary 1s are represented by voltage peaks on a single wire while zeros are represented by voltage troughs Because the signal relies on a single electrical channel voltage spikes caused by EMI are easily mistaken for signals RS 485 on the other hand uses two wires for carrying a binary signal in which 1s are represented by a high signal on one wire and a simultaneous low signal on the second Zeros are represented by the opposite signals on both wires The advantage of a differential signal is that an unintended voltage spike will occur simultaneously on both wires and will therefore be ignored A number of trial and error solutions were attempted by the Gaia Wind engineer to rule out ground loops and improve the signal quality however none completely resolved the issue Table 9 summarises several alternative solutions that involved the purchase of additional equipment which were therefore also considered In the end opto isolation and RS 232 over USB were both attempted The USB approach did not appear to make any difference despite its differential signalling Us
109. r STP cables are cheap and come with RJ 45 connectors so this solution would also require the use of adapters on both channels RS 232 over USB Universal Serial Bus USB communications are the modern PC communications standard and like RS 485 use a differential signal The original arrangement already made use of RS 232 to USB adapters to connect the equipment to the PC The solution in this case would simply be to eliminate the standard serial cable by placing the adapter at the equipment end of the connection rather than the PC end Table 9 Possible communications solutions 70 4 2 5 Residual issues Very few items were originally planned but not implemented One of these was the facility for the operator to specify a standard directory in which to store logfiles However the operating system stores the location last used in the Save File dialog and this is sufficient in most cases A further logfile item that was not implemented was the ability to open a logfile from within the application However since the files are in Excel format it is a simple matter to open them using the standard operating system tools The help documentation for the system has not yet been implemented This will be completed for the company after the end of the academic project There is one known bug in the system which related to the graphical speed and power displays The open source project that has been used assumes that the
110. rbine behaviour which reflects the response lag due to inertia in the rotor b To construct a control program for the Weg inverter which implements the model and also compensates for losses in the test rig itself c To evaluate the results of the control algorithm against the original logged wind and turbine data The first part of the project will be to produce a review of current literature on turbine testing and wind regime modelling The practical phase of the project will include the following activities a Collection of wind and turbine data from at least one site b Familiarisation with the Weg inverter and the selection of appropriate control methods c Analysis of the wind and turbine data to identify the nature of the turbine response to changes in wind speed and direction d Implementation of the derived model in software for control of the inverter e Iterative refinement of the program to compensate for physical losses and inertia in the test rig The evaluation phase will involve the correlation of the original logged turbine data with that observed in the test rig to determine the effectiveness of the model and the software implementation Consider the academic challenges that you will face and what solutions you proposed to deliver as part of the project and consider what will make the project a success There should be no more than 5 N B Approved deliverables will be compared with your final report Literature review Mathemati
111. real time operating systems for the efficient processing of signals to and from the WTE hardware Although no further details are provided this is probably a reference to a Unix variant that is optimised for real time processing An equivalent development could be produced using Ubuntu for example with the interface code built in Java The likelinood of this approach delivering a more efficient overall system is not as strong given that Munteanu was using a DC drive which requires simpler control The use of a SCIM and VFD may well deliver equivalent 1 https wiki ubuntu com RealTime 76 performance and control however the Java Ubuntu alternative would provide the opportunity to verify this 5 2 4 MATLAB The majority of previous WTE studies make use of the MATLAB development environment Specifically designed for scientific computing application MATLAB provides many features for developing algorithms for performing numerical computations efficiently and for visualising data MATLAB would provide similar benefits to using Excel in that less effort would be required to develop the HMI In addition MATLAB is specially designed to accommodate sensor input 5 2 5 Increased instrumentation The use of a shaft encoder on the VFD has already been mentioned in the context of closed loop control Other sensors could be added to the WTE hardware to provide verification of the inverter and turbine controller feedback A torque transducer on
112. red torque in the second stage project 4 1 4 Inverter Together with the SCIM the second component of the VFD is the CFW 11 IGBT inverter also manufactured by Weg Industrias The inverter is a wall mounted device with the general structure shown in Figure 27 As part of this project an additional RS 232 communications module had to be installed in the position labelled F in the figure A Mounting supports for through the wall mounting B Heatsink C Top cover D Fan with mounting support E COMM 2 module anybus F Option board accessory module G FLASH memory module H Front cover l Keypad KR Figure 27 Weg inverter components Weg 2008 60 The inverter whose block diagram can be found in Appendix J is built around a 32 bit central processing unit CPU which provides considerable onboard processing capability The unit has 16 digital and two analogue sensor inputs and offers the facility to upload control programs to the internal memory which can be written in one of the standard languages defined in the IEC 61131 3 standard including ladder logic Weg 2007 The discussion in 2 3 3 highlights the need for closed loop control and this could be provided using one of the digital input channels and the standard speed sensor on the drive SCIM however this was not in place in the Gaia WTE The control of the VFD is discussed in 4 1 5 below in light of this limitation Constru
113. rol e Grid interface control and power conditioning Intuitively the more power a turbine delivers the better In practice however a machine that operates beyond its rated value risks damage to components such as the generator It is therefore important to limit the power developed at high wind speeds so that safe limits are not exceeded A simple way to enforce power control is to select a blade geometry which induces a stall condition above a particular wind speed Referring to the Cp A curves in Figure 2 this ensures that the power generated by the turbine declines after a predictable point This arrangement is known as passive stall control since it is simply a feature of blade shape Burton et al 2001 p 350 and is necessarily only concerned with system protection Munteanu et al 2008 p 26 In contrast turbines which include the facility to control the pitch angle of the blades can not only avoid potentially dangerous conditions but can also actively optimise the angle of attack for a given prevailing wind speed This can be done by increasing the angle of attack to induce stall or by reducing the angle of attack to reduce the aerodynamic lift force a process known as blade feathering The additional energy extracted by active pitch control is only about 4 Burton et al 2001 p 351 but allows the turbine to better track the ideal power curve as shown in Figure 13 32 600 See Pitch regulated machine with TR blades
114. rupted wind as given by Eq 2 _ Output power p nid where o is the density of air 1 225 kg m U is the wind velocity m s Ais the circular area swept by the turbine blades m Burton et al 2001 p 44 where Q is the angular velocity in radians per second R is the rotor radius in meters Burton et al 2001 p 49 The power coefficient can be plotted against tip speed ratio for a given turbine design to give a characteristic performance curve Figure 2 shows this curve for turbines with different numbers of blades from which several observations can be made Burton et al 2001 p 175 e Fewer blades produce a broad flat curve where Cp remains roughly constant over a wide range of A but the maximum value of Cp is low e More blades give a higher maximum value for Cp but the curve has a narrow peak making the design sensitive to changes in A e A design with three blades produces the highest maximum Cp but a two bladed design gives more consistent performance over a wider range of A It should be noted that the curves in Figure 2 assume that the blades themselves are of similar dimensions in each case 15 A further performance measure which is commonly used is the torque coefficient Cg Burton et al 2001 p 64 provide a derivation of turbine torque from first principles and also state that the torque coefficient can also be calculated by dividing the power coefficient by the tip speed ratio for a given wi
115. s agreed that this would not be possible within the time constraints on the project The WTE development was therefore split into two phases of which this project is the first with the general aim of establishing communications with the VFD and turbine controller via a PC based human machine interface HMI Thus the initial system will not include a turbine model Instead of wind speed as input a target speed of rotation will be specified by the user This is sufficient to satisfy the immediate requirement of the company for a means for performing benchmark tests The introduction of the turbine model the provision of wind speed data as input and closed loop torque control are therefore deferred to the second phase The following sections discuss the specific objectives related to the aim of the first phase project 3 4 1 Evaluate hardware setup Gaia Wind has already put in place the hardware for a WTE The adequacy of this setup needs to be evaluated with respect to the requirements discussed in the literature review This evaluation will set the limits on what is possible with the current configuration and suggest ways that the hardware setup could be improved 46 3 4 2 Establish communication with VFD The VFD of which a Weg CFW 11 IGBT inverter is the main component provides a range of control opportunities via its programmable parameters To achieve this objective those opportunities need to be identified a suitable method of commu
116. s stage Logged data from a European 75 version of the turbine would be sufficient but data from controlled tests would provide a more reliable benchmark 5 2 Alternative implementations One category of possible future project explores the possibilities offered by using different platforms and development tools C was used as a development tool on request however 4 2 6 outlines other ways in which the same functionality might have been delivered Alternative implementations would allow direct comparisons between different toolsets with respect to efficiency of operation and ease of use and maintenance Some possibilities are discussed below 5 2 1 Excel Rebuilding the current project as an extension to Microsoft Excel using either C or Visual Basic would require less new work to develop the user interface and would also perhaps be easier to use due to the users familiarity with Excel With less code devoted to the user interface maintenance may also be easier 5 2 2 Weg Ladder Programming Although a lower level programming environment WLP may offer greater efficiencies due to its independence from an operating system In a second stage project aimed at providing closed loop torque control it is likely that ready made components would be available for implementing proportional integral and derivative control 5 2 3 Real time operating systems Some previous studies such as Munteanu et al 2010 have mentioned the use of
117. s with DFIG generators which provide many opportunities for adjusting operating conditions For simpler turbine designs with fewer controllable parameters generator control is still important in order to isolate the turbine from the grid at lower speeds to avoid drawing power 33 For smaller turbines the need for control over their interface with the distribution network is minimal For larger turbines however the quality of their output can affect other devices attached to the network through phenomena such as voltage transients due to wind speed variations harmonics caused by frequency conversion equipment and periodic fluctuations known as flicker Burton et al 2001 p 580 The requirements for power quality control can be summarised as follows S rensen et al 2005 cited by Munteanu et al 2008 p 101 1 Power frequency control ability with focus on a primary control fast automatic adjustment of power to frequency b secondary control slower automatic or manual regulation of the power to the power reference imposed by the system operator at any time 2 Voltage control ability with focus on voltage regulation and reactive power capability 3 Dynamic stability with focus on the ability of wind turbines to remain connected to the grid during some specific grid faults 2 3 Turbine emulation From the foregoing discussion it is clear that wind turbines can be very complex pieces of equipment that require complex an
118. sition between states in the same way that the current application does for speed inputs 5 3 3 Variable torque control This advanced version of the system would implement a full turbine model including torque oscillation It would also allow the user to provide input in the form of a variable wind speed reference either from a wind signal generator like those described by Diop et al 2007 or from logged data The implementation would require the use of standard control techniques such as proportion integral PI control as used by Fleming et al 2009 or proportional integral derivative PID control as used by Monfared et al 2007 5 3 4 Advanced control investigation Once the WTE is capable of replicating a variable wind profile the development focus could change to examine the control algorithms themselves A question for a project at this level would be whether the standard control algorithms are sufficiently dynamic to replicate the profile accurately One of the characteristics of a realistic wind profile is that there is a significant stochastic element Control algorithms designed to track a steady reference are not necessarily best suited to such a signal and work would focus on more experimental control approaches 5 4 Laboratory based simulation A further category of further projects involves the construction of tools to aid the development process itself At present access to the hardware configuration at the Gaia Wi
119. splays are intended to show instantaneous values and a continuous trace against time Two sets of values are shown in each panel one for the drive and the other for the generator E This box is used to set the sampling period for the speed and power traces It also determines the period of the logfile entries 4 2 Validation There are a number of obvious validation rules that can be applied to a sequence of instructions Clicking the Validate button causes the system to apply the rules to the sequence of steps currently displayed and to report any errors The table below lists the validation rules that correspond to specific actions In addition to these all tests must finish with a Stop instruction 107 Action Validation rule Meaning Start Inverter state off Inverter state is off by default Start n 1 Can only occur as first instruction Acc Inverter state on Test must start with a Start instruction Acc 0 lt Duration lt Duration cannot be negative Duration max RPM 1 gt RPM Target speed must be greater than current REC speed Acc DM lt RPMmax Target speed must not exceed limit Dec Inverter state on Test must start with a Start instruction Dec 0 lt Duration lt Duration cannot be negative Durationmax Dac DM lt RPM Target speed must be less than current speed Dec RPM 1 gt 0 Target speed may not be negative n Inverter state on Test must start with a Start instruction on c c
120. t al 2008 Yang et al 2010 Crabtree 2011 Johnson and Fleming 2011 Power quality and network transient analysis Seman et al 2005 Mauri et al 2008 Table 3 Motivations for the use of wind turbine emulators The basic requirement for a wind turbine emulator WTE is that it should reproduce as faithfully as possible the torque generated by a turbine for a particular wind speed Kojabadi and Chang 2011 Monfared et al 2008 summarise required elements of a WTE model as e A variable wind speed e Turbine inertia e Wind shear and tower shadow e Steady state characteristics The specific purpose of the WTE can influence its actual form For example because Helsen et al 2010 are primarily concerned with the performance of the gearbox it is important that their model isolates that particular component and they therefore take a multi body modelling approach which leads to significant complexity and calculation times For other purposes however the effective two mass model described by Anaya Lara et al 2009 is usually sufficient With this in mind the basic structure of a WTE can be derived from the schematic representation of a wind 35 turbine in Figure 14 where the components inside the dashed boundary are replaced by an alternative source of torque typically an electrical drive A computer model of the rotor characteristics provides a reference torque value Lopes et al 2005 anda control algorithm
121. t was developed It was also possible to introduce a few unplanned features such as the possibility of resetting the turbine controller via the HMI It quickly became clear during initial discussions with the company that there would not be sufficient time or resources available to complete the development of a fully featured WTE in one stage The scope of the project was therefore set at a lower level and this decision has been shown to be appropriate The current system addresses the immediate needs of the company for a simple interface for performing benchmark testing on new turbines before shipping However it does not include the more involved modelling of turbine behaviour seen in the previous studies reviewed in 2 3 Neither does it fully exploit the control capabilities of the CFW 11 inverter and in hindsight it would probably have been possible to use sensorless vector control rather than the much simpler V f approach from the outset This only became clear during the project however and to change the approach mid way would have introduced a much greater degree of complexity This would have put the overall completion of the project within the agreed timescale at risk 6 1 1 Review of current relevant literature A substantial range of previous WTE studies was identified during the literature review and a consistent picture of WTE requirements duly emerged An early focus on WTE studies also identified the relevant aspects of turbin
122. t would have taken longer In terms of the current project aim this objective is successfully delivered and has also revealed useful directions in which the WTE can be developed in the future 6 1 3 Software development The delivered software performs according to the initial requirements and can therefore be considered successful A major problem late on in the development was the disruption of the RS 232 communication signals that prevented a full evaluation of the results This was partly the result of a lack of access to the WTE hardware in the first week of development which led to a review of the schedule Had the original schedule been adhered to the inverter communications module would have been tackled in the first week This would probably have led to the discovery of the communications issues earlier and this might have made a solution possible in time to carry out the evaluations fully This is speculation however since efforts to resolve the communications problems might have reduced the time spent on other features 6 1 4 Evaluation of the project outcomes As alluded to in the previous section the evaluation of the project outcomes is the weakest part of the project since time and other resources were very short towards 81 the end of the project period Apart from the incorrect recording of values in the log file the software has been shown to work correctly according to the specification and a procedure for reportin
123. tant air flow which is relevant when calculating the forces on the blades of a rotating turbine and Figure 4 shows its derivation Also shown in Figure 4 is the blade pitch angle B The pitch angle is significant because it is a controllable design parameter whereas the angle of attack varies with local conditions Induced air flow due to Wind direction Figure 4 Resultant airflow on rotating blade Adapted from Burton et al 2001 p 61 Because we are interested in the torque applied to the rotor shaft a final decomposition of the lift and drag forces is required to determine the component that acts in the direction of rotation By simple vector decomposition this turns out to be L sin D cos where L and D are the lift and drag forces and is the angle between the airflow experienced by the blade and the plane of rotation ie a B This decomposition is shown in Figure 5 which also shows the considerably larger force acting perpendicularly to the plane of rotation Burton et al 2001 p 61 18 L cos D L sind D Figure 5 In plane and out of plane forces on turbine blade Adapted from Burton et al 2001 p 61 The actual aerodynamic behaviour of the rotor blades depends on their detailed design features Some of the parameters that can be modified are the rotor radius the width of the blades at different points along their length the variation in the pitch angle along the blade length and the d
124. ted by At R rMPmax Arpm A further calculation is required to give the value of P0683 the serial speed reference This is a 13 bit value which is calculated as a ratio of the required speed to the synchronous speed of the motor times a constant of 2000h 8192 decimal 114 The synchronous speed of the motor is stored in P0402 The value of P0683 is given by rpm 8192 A where rom is the required speed and S is the synchronous speed of the motor Param Description Meaning P0100 Acceleration time Based on duration P0683 Speed reference Speed read from Ul 7 5 Deceleration The duration of a deceleration action is controlled in the same way as that of an acceleration action Param Description Meaning P0101 Deceleration time Based on duration P0683 Speed reference Speed read from Ul 7 6 Stop action The same inverter instruction is required for a normal stop action which follows a deceleration to 0 rpm and for an interruption to the test run If a test run is interrupted prematurely the drive may still be rotating at high speed To avoid tripping the inverter it is therefore important to bring the drive to a stop over an extended period of time The deceleration ramp is therefore set to a value which will stop the drive over 20s Param Description Value Meaning P0101 Deceleration time calc Based on duration P0682 Serial control word 0x0000000000000000 Stop and disable a E C 7 7 Status p
125. ting but not for anything more realistic In particular it is not possible to match behaviour of the WTE with wind speeds The current software implementation uses the V f control mode of the inverter This does not allow direct control over the generated torque which would be a necessary step on the way to a fully featured WTE The current hardware does not support closed loop control of the VFD Although the inverter manual suggests that sensorless vector control provides speed control precision of 0 5 or rated speed it does not make any claims about the precision of torque control Adding a shaft encoder to the drive would theoretically increase speed control precision to 0 01 of rated speed and is recommended for dynamic performance and torque control Weg 2010 The behaviour of the SCIM used as both motor and generator in the American version of the turbine does not seem to correspond to that expected on the basis of its datasheet Appendix F judging by the examination of the NREL data Appendix G Independent testing of the unit in the WTE would help to determine its operating characteristics with greater reliability Direct measurement of the resistance and reactance of the stator windings using standard tests would verify the value provided on the datasheet and also allow the direct calculation of generated torque The absence of appropriate comparator data makes the verification of the measurements from the WTE impossible at thi
126. tions Communications with the Weg inverter will use the Modbus protocol over an RS 232 serial connection The inverter will be modelled as a C class whose properties correspond to the inverter parameters that need to be read or updated during a speed test 7 1 Setup Those parameters that configure the inverter with the basic communications selections will need to be set manually using the front panel on the physical unit These are shown in the table below L R Param Description Value Meaning Default P0202 Type of control 1 VI 50 Hz LOCAL REMOTE 2 L R key selection Speed ref LOCAL Keypad P0220 P0221 P0222 Speed ref REMOTE Serial USB FORWARD REVERSE Always REV LOCAL LOCAL run stop 0 I O keys P0223 P0224 P0225 JOG selection LOCAL 1 JOG key FORWARD REVERSE 1 Always REV REMOTE REMOTE run stop 2 Serial USB P0226 P0227 P0228 JOG selection REMOTE 2 DIx Y Y P0308 Serial address 1 Device id P0310 Baud rate 0 9600 Y P0311 Serial byte config 0 8 bits no parity 1 stop P0312 Serial protocol Modbus RTU The setup configuration above assumes that the keypad on the front of the unit is the LOCAL condition and the serial connection is the REMOTE condition Switching between the two is always done from the keypad This provides a failsafe method for stopping the drive when under software control To begin a software controlled session the inverter must be set to REMOTE using the keypad Only then wi
127. to make the turbine begin to try and slow then finally stage 14 brings the turbine to a stop I m keeping a track of bugs and thoughts on improvements The one bug I ve found is that the Turbine power is displayed and logged as unsigned It needs to be signed as the power can go negative You ll see this in the attached log Also drive power is being displayed correctly but not logged correctly Derek Sent 22 August 2011 15 38 To Derek Robertson Derek Robertson gaia wind com Attachments Report v3 docx 6 MB Open as Web Page Hi Derek Here is a copy of the report It s quite long but you will probably be most interested in chapter 3 onwards If you have any comments please let me know E E You will see I ve not complete the evaluation part yet was wondering if you could run the attached test script and let me have the output To import it use the import option on the File menu Thanks Brian Sent 22 August 2011 15 39 To Derek Robertson Derek Robertson gaia wind com Attachments Rats Clicked the Send button instead of the attach button Here s the script B Sent 23 August 2011 07 41 To Davison Brian Brian I m currently testing out on site Not sure I ll get a change to run this test before Thursday d assume that this isn t going to happen on time for now If get the opertunity to test will do so and send the log ASAP II try have a read at your report today
128. tor drive n drive n2 dt where une is the torque required from the electric drive Jarive iS the inertia of the electric drive aech is Mechanical torque nis the gear ratio Jrotor IS the rotor inertia Wgenerator iS the rotational speed of the generator Monfared et al 2008 Kojabadi and Chang 2011 The torque output of the WTE drive needs to be controlled to the reference value provided by Eq 11 in order to produce a realistic result 2 3 3 Instrumentation and control Knowing the torque produced by the WTE is clearly important if it is to be controlled with reference to the turbine model There are essentially three options the first of 40 which is to estimate the torque based on known characteristics of the drive This is the approach taken for example by Martinez et al 2009 Ming Qiao et al 2007 and Monfared et al 2008 Details of the process for estimating the drive torque are not provided however Monfared et al include a component in their reference torque calculation aimed at compensating for the inertia of the drive itself This was a problem identified by Dolan et al 2005 as one of the reasons for including a torque transducer The second option is to take direct measurements from the WTE drive and calculate the torque This is easier in the case of a DC based WTE because of the direct relationship between torque and current Moore and Ekanayake 2010 and Munteanu et al 2010 for example calculate torque
129. tor of an induction EM rotates at a different speed which varies depending on the electrical and mechanical loads The difference between the speed of the stator field and the rotor is known as the slip speed and is usually expressed as a percentage of synchronous speed as defined by Eq 8 27 Ws Wr 100 Ss where Sis slip Ws is synchronous speed wr is rotor speed When slip is positive the rotor speed is less than that of the stator field and the EM is operating in motoring mode In contrast negative slip indicates that the rotor is moving more quickly than the stator field and the EM is therefore in generating mode The actual relationship between slip and torque for a typical EM is illustrated in Figure 10 S ee siamese ai Gei Braking region region Generating e gl region Figure 10 Typical EM torque speed characteristic Bakshi and Bakshi 2009 The curve is symmetrical about the synchronous speed ng for the region from 0 rom to 2n Depending on the design of the EM operating slip values typically range between 2 and 10 Gross 2007 p 146 Although the torque speed characteristic of a SCIM can be expressed as an equation it requires knowledge of certain characteristics of the machine s construction such as the resistance of a single phase stator winding which must be obtained through measurement 28 However a simpler relationship exists between the torque applied in generating mode
130. tunately they are not consistent The power performance report Huskey et al 2009 quotes 56 62 rpm the generator duration test report Huskey et al 2010 quotes 0 62 rpm and the generator function and safety test report Huskey et al 2010b quotes 60 62 94 rpm The reasons for this discrepancy are not explained which is puzzling given the continuity of the testing team for the three reports However it is possible to argue away some of the differences The duration test aims to assess among other things material degradation over time and the dynamic behaviour of the turbine Huskey et al 2010 p 1 It may therefore be important to consider the starting and stopping of the turbine as a potential source of fatigue and hence the inclusion of low rotor speeds in the range 0 56 rpm The power performance test on the other hand is concerned with the turbine s power output and therefore only rotor speeds above and immediately below the generating threshold are of interest At rotor speeds of less than 56 rpm it is assumed that the controller has isolated the turbine from the distribution network The safety and function test is more concerned with extreme conditions and explicitly refers to the dynamic behaviour of the turbine at rated and higher wind speeds Hence the higher range of speeds quoted 60 62 rpm The reports are consistent on other specification parameters The Gaia Wind turbine datasheet Gaia Wind 2009 quotes 56 rpm as t
131. user manual About Displays basic details ina dialog 4 Automatic mode File Test Log Help bn Lg 1 Es ee Na SH WS Wee LL a WE LL LL ET pad O S Sampling period Speed display i SS Power displa a ged REES VIER II TI 4 1 Notes A The test id is an internally generated number which cannot be changed Id number may be discontinuous The description is editable and required for saving a test to the database 106 B The sequence of instructions that make up a test are shown in a vertically scrolling table The details can be entered or altered manually or they can be loaded from an external XML file using the option on the File menu C Only a restricted set of actions are possible and they can be chosen by the user from a drop down list The available actions and their meanings are shown in the table below Note that Acc and Dec could have been replaced by a single action meaning Change to the following speed however using Acc and Dec indicate the intention of the user and can therefore be used for validation Action Description Duration s RPM Start Start the inverter Greyed out Greyed out A Accelerate Duration of acceleration Target value c De c Ps Decelerate Duration of deceleration Target value n Remain at constant Duration of steady state Speed to be SR speed maintained Stop Stop the inverter Greyed out Greyed out D The speed and power di
132. user to control the VFD to monitor the response of the turbine generator and to perform the summary calculations based on the information from both units VFD communications weg Test module management module Controller communications module Gaia turbine e Figure 21 High level system design External files scripts and logs The HMI allows the user to specify the speed of the VFD which is theoretically the same as the angular velocity of the generator rotor Input can be specified directly or as part of a scripted series of speeds with associated durations Among other quantities the VFD interface provides the following information about the input to the WTE e Angular velocity of the drive rotor e Electrical power e Drive torque The controller reports the electrical power output from the turbine generator and given the appropriate version of the controller firmware also report the angular velocity of the generator rotor Based on the data from the two interfaces the HMI provides calculated figures for the generator slip and overall efficiency Slip is calculated from the speed of the generator and the network synchronous speed of 1000 rpm while the efficiency figure is based on the input power and output power reported by the inverter and turbine controller respectively 50 The company brief Appendix B specifies the ability to log operational data from a test run to an external file The i
133. ve 16 An 14 4 12 10 4 Power kW O N Q o 4 4 L 1195 1200 1205 1210 1215 1220 1225 1230 1235 Generator speed rpm An interesting feature of this graph is that power output remains considerably higher than the rated value at HSS speeds greater than approximately 1205 rpm Comparing this graph to the power curve in the NREL power performance test report shown in the next figure it can be observed that the form is similar and that similar high outputs were measured at wind speeds between 9 5 and 20 m s The level of power output can be explained with reference to the test procedure in which the 96 measured data is normalised to sea level The test results represent the adjusted figures and because air is slightly denser at sea level 1 225kg m than at the test location 95kg m the power output appears raised The effect of a higher value for air density can be seen in Eq 1 The high broad curve is characteristic of a two bladed turbine and can be compared with the curves in Figure 2 Sea Le vel Air Density Normalized Power Curve forthe Gaia Wind 11 kKW Wind Turbine Norm dized Power kiy Wind Speed m s Huskey et al 2009 Torque is calculated by applying Eq 8 to the values for output power and the synchronous speed of 1200 rpm The result shown below has the correct form but raises some questions about the performance of the generator compared to the reference values pro
134. vided on the manufacturer s datasheet 97 Partial torque speed characteristic oO 7 r d o r 201195 1205 1210 1215 1220 1225 1230 1235 Torque Nm a Generator speed rpm Theoretically the main inflection in the torque speed characteristic should occur at the so called breakdown torque According to the datasheet for the 60 Hz generator which is provided in Appendix F breakdown torque is 280 of rated torque Since rated torque is given as 102 Nm breakdown torque should be 285 6 Nm however in the calculated torque speed graph above the main inflection occurs at around 120 Nm barely 117 6 of the rated value The sign is negative because the SCIM is in generating mode and so only the absolute value is needed for comparison The reason for this substantial discrepancy is not clear and suggests the need to test the behaviour of the SCIM in isolation 98 Appendix l M NET protocol Extract from the Mita Teknik IC1000 manual 1 The format of the request packet 60133 The request packet must have the following format SOH Destination Source Packet Data 2 255 type count ap O CRC 1 Joes Im ox ja The CRC is the CRC CCITT applied to Destination Source Packet type Data count and Data Below is shown an example requests to node 2 where data 1 Stop 01 02 01 FA E5 01 01 A5 81 04 2 Reset 01 02 01 FA E5 01 02 95 E2 04 4 Start 01 02 01
135. which are not needed and which therefore constitute unnecessary sources of power loss In particular the two gearboxes and low speed shaft shown in Figure 24 could be eliminated PSS SS SSS SS a SSS SSS SS SS SS Se AA SiS SS Inverter High High speed speed ane shaft shaft Grid Generator Gearbox Gearbox Figure 24 Schematic representation of Gaia Wind WTE On the other hand the structure has the advantage of being symmetrical The gearboxes are identical and so are the drive and generator In theory therefore the losses should also be symmetrical and if an overall value for system losses can be 56 determined it can be divided equally between the two halves of the system In practice the efficiency of a SCIM operating in generating mode may not be the same as when it is operating in motoring mode but the approximation may be reasonable in this case A further advantage of the Gaia hardware arrangement is that it employs the actual turbine hardware in the WTE Previous work on WTEs eg Dolan et al 2005 has typically used scaled down hardware which introduces further potential sources of uncertainty into the results 4 1 2 Instrumentation and control As a fixed pitch machine the Gaia Wind turbine relies on the aerodynamic characteristics of the rotor to induce stall at higher wind speeds to prevent over generation As an additional precaution however the rotor are fitted with aerodynamic brakes which deploy i
136. while making up only 32 of its geographical area RenewableUK 2011 Country level statistics are typically based on the deployment of multi megawatt turbines in windfarms and much effort goes into the design of turbines with larger and larger capacity in order to better exploit the resource at a particular site One of the largest turbines currently deployed for example is the 7 5 MW Enercon E126 while several larger units are currently under development and will be delivered in the next few years The Azimut consortium led by Spanish company Gamesa aims to deliver a 15 MW turbine by 2014 Gamesa 2010 At the other end of the scale however the development and deployment of smaller turbines also shows strong growth A market report by RenewableUK describes an extremely active industry in which installed capacity in the UK rose 65 during 2010 from 8 62 MW at the beginning of the year to 14 23 MW at the end RenewableUK 2011b One of the factors that contributed to this growth was the introduction in April 2010 of feed in tariffs FITS as a financial incentive for small scale renewable electricity generators DECC 2011 Under the scheme the operator receives a payment for generating electricity which currently ranges from 4 7 to 36 2 pence per kilowatt hour depending on the installed capacity If the generated electricity is also exported to the grid a second payment is made of 3 1 p kWh A rough calculation shows that for a 10 kW turbin
137. with an industrial partner e Sufficiently short timescale to allow completion within by the August 2011 deadline No one particular energy technology was deemed more attractive than another for the purposes of project selection in order to increase the range of potential opportunities The chart in Figure 20 describes the subsequent three stage process that was undertaken to arrive at a final project decision 3 Evaluate final 1 Approach potential partner companies 2 Negotiate project details options against initial criteria Figure 20 Project selection process Scottish Renewables maintains a membership directory with company contact details which is accessible on their Web ste Currently there are just over 300 companies listed During the week beginning 18 April 2011 all entries in the list were examined to filter out unlikely partner organisations such as local authorities S http Awww scottishrenewables com members membership directory 44 legal firms consultancies etc The result was a shortlist of 31 companies who were mainly small and locally based to facilitate communications during the project Some larger companies were also included if they had known interests in the Edinburgh area The list also contained a small number of research centres in addition to commercial firms All shortlist members were contacted on 22 April 2011 by email using a variation of the text in Appendix A
138. y without interfering with each other A multi threaded design is therefore implemented which allows background tasks to be completed without preventing the user interacting with the interface During the development the need for several concurrent threads raised a number of timing issues some of which were very time consuming to resolve In most cases the solution was to use NET events Access to the methods of one class from within another proved difficult to implement The eventual solution was to use the NET delegate mechanism In the initial design the relationship between the different hardware elements was not clear During the development a device class was introduced The generic features of a device could therefore be assembled in this class including the association with a serial port Both the inverter and the turbine were represented by 5 http msdn microsoft com en us library awbftdfh aspx http msdn microsoft com en us library ms173171 aspx 6 67 a subclass of device The final class diagram is available in the system specification in Appendix K 4 2 4 Communications Communications with the inverter use a subset of the industry standard Modbus protocol The required message formats are described in the Weg serial communications manual Weg 2010b Messages are assembled by the application as a series of bytes which are written to the serial port Every message generates a response which in the case o

MSc dissertation - School of Computing

Contents

Download Pdf Manuals

Related Search

Related Contents