Research On Predictive Power Control Of VSCF Doubly Fed Wind Power Generation System

With the growing energy crisis and environmental problems, wind power generation technology as a renewable power generation technology has been widely studied. Doubly Fed Induction Generator (DFIG) has become the mainstream model of VSCF wind power generation with its small inverter capacity and reliable performance and direct Power Control (DPC) has become a research focus of DFIG inverter’s control due to its fast dynamic performance and simple control structures. In recent years, predictive control based on the computer discrete control has attracted more and more attention. This thesis proposed predictive power control of DFIG system combining predictive control and direct power control for doubly fed wind power generation system under ideal grid and unbalanced grid condition and conducted theoretical analysis, simulation modeling and experimental studies.This thesis researched grid side converter and rotor-side converter respectively. First, a summary of several existing DFIG control strategy and an introduction of several control method of DFIG under unbalanced grid condition was given. Secondly, under the ideal grid conditions, a mathematical model of the grid side converter in the stationary coordinate and the mathematical model of the rotor-side converter in the stator stationary coordinate and the rotor stationary coordinate was established, and the power equation of the grid side converter and rotor side converter in each coordinate system was derived, providing a theoretical basis for predictive control afterwards. Subsequently, for the grid side converter and the rotor side converter, looking-up table based direct power control (LUT-DPC) was introduced briefly, for reducing large steady-state power fluctuations of LUT-DPC, A variable switching frequency predictive optimal power control based on the finite control set (FS-PPOC) was proposed. The principal of FS-PPOC and control delay compensation was described in detail. Simulation results show that FS-PPOC was comparable to the LUT-DPC in the dynamic performance, and significantly reduced the power fluctuations and static error in the steady state performance due to the optimized switching vector selection. Because LUT-DPC’s and FS-PPOC’s switching frequency is not constant, two fixed switching frequency direct power control strategy, deadbeat principle-based direct power control (DB-DPC) and predictive direct power control based on three vectors (PDPC) was proposed. The basic principle and delay compensation strategy was explained in detail for the DB-DPC. Simulation results show that DB-DPC and PDPC can get a smaller power fluctuations and constant switching frequency and delay compensation can improve steady-state and dynamic performance of the DB-DPC For grid side converter’s PDPC, this thesis focused on the problem of negative vector’s action time. Two methods, including re-selecting vector and determining sector by the voltage vector of the inverter side were proposed to ensure the action time of vectors is optimal and positive, eliminating power fluctuations caused by the negative action time. For rotor side converter’s PDPC, a single switching-table PDPC (ST-PDPC), eliminating the need for hysteresis and simplifying switching table was proposed. In generating reactive power conditions, an action time processing method was introduced to optimize the control performance of the ST-PDPC. The simulation results proved the correctness and effectiveness of improved PDPC. Under unbalanced grid condition, a improved FS-PPOC adding power command compensation terms was proposed and verified by experiment to achieve three control objectives for ride-through of DFIG.Finally, the hardware structure of experiment platform and software flow chart of the control strategy was introduced briefly. The experimental results of LUT-DPC, FS-PPOC with and without compensation, DB-DPC with and without compensation, PDPC, improved FS-PPOC under unbalanced grid and action time processing of PDPC were given and quantitatively analyzed. Experimental results proved the correctness and effectiveness of these proposed control strategies.