A modified Adaptive Wavelet PID Control Based on Reinforcement Learning for Wind Energy Conversion System Control

Advances in Electrical and Computer Engineering

关键词: control;    reinforcement;    neural network;    wavelet;    wind energy;   

DOI  :  10.4316/AECE.2010.02027

来源: DOAJ

【 摘 要 】

Nonlinear characteristics of wind turbines and electric generators necessitate complicated and nonlinearcontrol of grid connected Wind Energy Conversion Systems (WECS). This paper proposes a modified self-tuningPID control strategy, using reinforcement learning for WECS control. The controller employs Actor-Criticlearning in order to tune PID parameters adaptively. These Actor-Critic learning is a special kindof reinforcement learning that uses a single wavelet neural network to approximate the policy functionof the Actor and the value function of the Critic simultaneously. These controllers are used to controla typical WECS in noiseless and noisy condition and results are compared with an adaptive Radial BasisFunction (RBF) PID control based on reinforcement learning and conventional PID control. Practicalemulated results prove the capability and the robustness of the suggested controller versus theother PID controllers to control of the WECS. The ability of presented controller is tested by experimental setup.

【 授权许可】

Unknown