Design of adaptive PID for pitch control of large wind turbine generator

Abstract

Pitch control system plays an important role in large wind turbine generator (LWTG) because of the direction and intensity of wind is changing every moment. Therefore, the system of pitch control is difficult to be set up the tuning parameters of controller. Conventional PID controller is not suit for all operating point. Because, tuning parameters of conventional PID controller are not consistently changed. In this paper, a self tuning PID control method using reinforcement learning is proposed the pitch control of large wind turbine generator (LWTG). Actor-Critic learning is used to tune PID parameters in an adaptive way by taking advantage of the model-free and on-line learning properties of reinforcement learning effectively. Also proposed controller includes the improved integral control action. Therefore, the robust control of LWTG is studied utilizing the new adaptive PID (NAPID) controller in this paper. This structure is composed with new type of integral control action and Actor-Critic learning part. The improved integral control has concept of error window and weight function concept. The performance of newly proposed adaptive PID controller is compared with those of conventional ones via Matlab SIMULINK simulation.