DSP and FPGA based system to control a wind turbine generator implementing a variable speed vectorial control method

DSP and FPGA Based System to Control a Wind Turbine Generator Implementing a Variable Speed Vectorial Control Method M. Perales, F. Barrero, J.L. Mora, E. Galv´ an, J.M. Carrasco and L.G. Franquelo Dpto de Ingenier´ıa Electr´ onica, Univ. de Sevilla, Avda Reina Mercedes s/n, 41012-Sevilla (SPAIN) phone: +34-5-4556873; fax: +34-5-4556849;e-mail: [email protected] Abstract— Thepurpose of this paperisto describeaDSP and FPGAcontrol systemtoimplementavariablespeedvectorial control . Two semi-systems, both of them consisting on a DSP, a FPGA and A/D, D/A & digital I/O’s are used. Each one will control an inverter: The first inverter implements a variable speed vector control of the induction generator and thesecondonehandlethepowerinjectedintotheutilitygrid. Experimentalresultswillbeshown to confirm thevalidity of the proposed controller. I. INTRODUCTION Many horizontal axes, grid-connected, medium to large scale wind turbines are regulated by pitch control andmostofwindturbinessofarbuilthave practically constant speed, since they use an AC generator, directly connected to the distribution grid, which determines its speed of rotation. In the last years, variable speed control is been added topitch–angle controlled design ([1]–[4]) in order to improve the performance of the system. Variable speed operation of a wind turbine has a number of advantages: the reduction of electric power fluctuationsbychanges in kinetic energy of the rotor, the potentialreduction of stressloads on the blades and the mechanical transmissions and the possibility to tune the turbine to local conditions by adjusting the control parameters. The objectives for variable speed control system are summarized by the following general goals: 1. To regulate and smooththe power generated. 2. To maximize the energy capture. 3. To alleviatethetransientloadsthroughoutthe wind turbine. 4. Unity power factor in the line side with no harmonics current injection. 5. To reduce the machine rotor flux at light load reducing core losses. Objectives for the pitch–angle control are: 1. To permit the starting blades angle to differ fromtheoperationbladepitchangle,henceallowing easier starting and optimum running. 2. Overpower and overspeed can be dissipated through rotor pitch regulation. The paper describes a digital controller based on a DSP and a FPGA to implement these control policies. The controller has been evaluated using a the test rig including the power circuits and the control strategies, that will be described later. The paperhasbeenorganizedinthefollowingsections: section two describes test rig implementation and control algorithms applied tothe induction generator. Next sectiondescribes theimplementationof the proposed controller. Section four shows some experimentalresultconfirmingtheperformanceof the prototype and finally, will be given some conclusions. II. TEST-RIG DESCRIPTION A. Power Circuits Figure 1 shows the power and control systems of the test-rig. In that figure, two subsystems can be found: the DC motor and the AC generator control. The DC motor is controlled by a commercial thyristor full bridge rectifier with its control systembased ona microcomputer. In order tocontrol 1 theinductiongenerator, twoinverterconnectedby a DC–Link capacitorhave been used. The first one works as a controlled rectifier and the second one handles the power injected into the utility grid. The overall systems is controlled using a personal computerwerethewindturbinemodelwasimplemented by software. In figure 2 is shownthe Test– Rig general view. Current controlled DC motor 000 00 0 00 0 111 11 1 11 1 +DC AC AC DC AC DC Microcomputer Control Drive Signal and Pulse Drive Conditioning RS232 PUBLIC GRID AC GENERATOR DC MOTOR ENCODER DSP Control Board Personal Computer LOAD CELL TORQUE SIGNAL Figure 1: Power and Control Diagrams of the Test-Rig Variable Speed Wind Turbine Figure 2: Photograph of the test–Rig. is used to implement a Variable Speed and Pitchangle Wind Turbine. In the wind turbine model the reference torque is compounded of three components: aerodynamic, oscillatory and dynamic torque which are calculated on the PC. The referencetorqueissenttotheDCmotormicrocontroller from the PC by a RS232 serial communication. B. Control Strategies The control of the induction generator can be divided inthreeparts: Powercontrol,inductiongenerator control and power injected into the utility grid control. Power Control Block: The power control block implements different control policies depending on the wind speed and pitch–angle position. The control policies are torque, rotor speed or power constant reference. Induction Generator Control Block:This control is based on an indirect vector method ([5]). The well known block diagram of this control system is shown in figure 3. The rotor speed is added to slip speed to obtain the synchronous speed of the reference frames: ! e = p 2 ! r + KCV i q ( 1 ) ! e is the synchronous frequency. p is the pole number. ! r is the mechanical frequency. KCV is a vector control constant. In thatreference frame, thetorque and flux control is performed. A Space Vector technique is used to control the induction generator current. This current control has been implemented in a synchronous reference frame. Fuzzylogic techniques ([6])are beenimplemented in order to improve speed control. In a future, fuzzy selftuningcontrol systemfor Induction generator will be used ([7]). Control of the Power Injected to the Public Grid: The DC–linkcapacitorsvoltagemustbecontrolled to maintain a constant reference voltage. This is performed, injecting the active power delivered of the induction generator. The control block diagram is represented in figure 4. A space vector θ ENCODER GENERATOR + KCV ωω ωr cos sen θ θ iq d i S 1 SENSE CURR. MEASUR. sl e CURRENT S. V. CONTROL SPEED Figure 3: Indirect Vector Control Method Block Diagram. technique is used to control the current injected into the utility grid. In this case, current control has been implemented in a stationary reference frame. 2 θ + SENSE - + + + V r P I icapac. UTILITY GRID POWER CALC. POWER CALC. S. V. CURRENT CONTROL CURR. SENSE VOLT. SENSE VOLT. Figure 4: Public Grid current injectionBlock Diagram. III. CONTROLLER IMPLEMENTATION In this section it will describe the control system physically. The system itself is composed of two semi-systems, each one controlling a three-phase full-bridge inverter, linked together thru a DualPort Ram. in the Figure 5 we can see a global description of this system: In this figure, the main Control LinesControl Lines trigger trigger UTILITY GRID (SLAVE)GENERATOR ( MASTER) RS232 DSP1 DPRAM F.P.G.A. P.P.I. 32 bit ROM 32 bit RAM U.A.R.T. 12bit DAC. Conv. MUX 4:1 12 bit A/ D MUX 4:1 12 bit A/ DA/ D 12 bit MUX 4:1 DSP2 TMS320C30 DPRAM F.P.G.A. P.P.I. 32 bit ROM 32 bit RAM U.A.R.T. 12bit DAC. Conv. MUX 4:1 12 bit A/ D MUX 4:1 12 bit A/ DA/ D 12 bit MUX 4:1 TMS320C30 WATCH DOG DOG WATCH Figure 5: Overview of the control system implemented. components are shown. They are: a TMS320C30, 40 MHz DSP, a FPGA as co-processor, a 8255 PPI, a 16552DUART, a WatchDog module, a Dual Port RAM to provide communication between subsystems,and 32 bit RAM and EPROM. Due to real time process requirements, and control two complex systems (two inverters) a floating point DSP was chosen. The FPGA used in the design was the TPC1020BFN-068C, and perform the following tasks: interrupt handling, address generation, speed encoder read, and managing triggers of the inverter. IV. EXPERIMENTAL RESULTS In thissection experimentalresultsobtainedof the test-rigprototypeare analyzedtoevaluatetheperformance of the proposed controller based on a DSP and FPGA. Fig. 6 show oscillograms of the injected current in the public grid in nominal conditions using the proposed control board, and the motor current. Fig 7 show oscillograms of the dc link capacitor voltage ripple, and the current injected in the public grid. It should be noted that the oscillogram of the dc voltage has been magnified in order to observe its ripple (600 V DC level against 15 V pp). Fig 8 show oscillograms of the generator current and the rotor speed for a step loadchange from5kWto2kWofthewindpower. The final offset, due to an unadjusted constant on the PI, is being corrected. Fig 9 show oscillograms of thegenerator current and its speed, when accelerating from 0 to 1500 rpm. Figure 6: Oscillogramsof the injectedcurrent in thepublic grid (channel 1) and the generator current (channel 2) in nominal conditions (1500 rpm 380 V). Channel 1 10A/div, channel 2 5A/div and 10ms/div. Figure 7: Oscillograms of the dc link capacitor voltage ripple (channel 1),and thecurrent injectedinthe publicgrid (channel 2) innominal conditions (1500rpm 380V). Channel1 10V/div, channel 210A/div and 20ms/div. 3 Figure 8: Oscillograms of the generator current and the rotor speed for a step load change from 5 kW to 2 kW of the wind power. Channel110A/div,channel2500rpm/divand1s/div. Figure 9: Oscillogramsof the generator current (channel 1)and its speed (channel 2), when accelerating from 0 to 1500 rpm. Channel 1 20A/div, channel 2 250rpm/div and .2 s/div. V. CONCLUSIONS A control board based on a DSP and FPGA is proposed in this paper in order to evaluate some control policies for wind-energy conversion systems. To confirm the validity of the propose control board and the control strategies a test rig prototype has been built in the laboratory. A current controlled DC motor has been used to implement a model of variable speed and pitch-angle wind turbine. The DC motor has been used to drive an induction generator. The control of the generator has been implemented using vector control method and space vector current control. Preliminary experimental results confirming the validity ofthecontrolmethodproposedinthetest–righave been shown. ACKNOWLEDGMENT This work has been supported by the First User Program (FUSE) included in the ESPRIT project, entitled “DSP-Based control for variable speed wind turbine” experiment number 2063. REFERENCES [1] R. DavidRichardson andGerald M. McNerney, “Wind Energy Systems”, Proceeding ofIEEE, VOL. 81, NO.3, March 1993. [2] M. Godoy Simoes, Bimal K. Bose, “Fuzzy Logic Based Intelligent control of a Variable Speed Cage Machine Wind Generation System”. IEEE, 1995. [3] Torbjonrn Thiringer and Jan Linders, “Control by Rotor Speed of a Foxed–Pitch Wind Turbine Operating ina Wide Speed Range”. IEEE/PES Summer Meeting, Seattle, 1992. [4] A.D.Simmons, L.L.Freris andJ.A.M.Bleijs,“Comparison of Energy Captureand StructuralImplementations of Various Policiesof Controlling Wind Turbines”, Wind Energy: Technology and Implementation (Amsterdam EWEC’91). [5] B. K. Bose, Power Electronic and AC Drives, Prentice-Hall Englewood Cliffs, N. J., 1987. [6] E. Galv´ an, F. Barrero, M. A. Aguirre, A. Torralba, L. G. Franquelo,“A Robust Speed Control of AC Motor Drives based on Fuzzy Reasoning”, IAS–93 Annual Meeting,pp 2055–2058,Toronto 1993. [7] F. Barrero, E. Galv´ an, A. Torralba, L. G. Franquelo, “Fuzzy Selftuning System for Induction Motor controllers”, European Power Electronics EPE’95, Seville, Spain, September 1995. 4