Compensation Control Studies for Variable-Speed Wind Power Systems Based on Neural Network

摘要

Wind power systems have strong nonlinear characteristic, when wind speed is below the rated value. A novel intelligent hybrid maximum power point tracking (MPPT) controller for variable-speed wind power systems is designed, linear parameter-varying (LPV) gain-scheduling controller is designed to regulate output of fast-dynamics generator torque, neural network compensator is used to eliminate the interference caused by unknown parameters, WPSO algorithm is used to train neural network. Hardware-in-loop simulation model is built up based on Matlab/Xilinx, results prove that the method can effectively improve wind energy utilization efficiency, mechanical vibration of wind turbine is reduced, the maximum wind energy is captured, a better idea is provide for application of FPGA in wind power field.%Zaproponowano nowy system sterowania systeme elektrowni wiatrowych uwzględniający zmienność szybkości wiatru. System wykorzystuje sztuczne sieci neuronowe co szybkiego sterowania dynamiką momentu.