RBF neural network based self-tuning PID pitch control strategy for wind power generation system

For the problem of wind power generation system (WPGS) with much uncertainty and serious nonlinearity, precise mathematical model based traditional PID controller cannot meet the requirements of pitch control. In order to improve dynamic performances of WPGS in constant power output running area, a radial basis function neural network (RBFNN) based self-tuning PID pitch control strategy is resented in this paper. The error between measurement and given value of generator power is given as the input of the controller. RBFNN is used to be the identifier, giving an identification to pitch system and obtaining the identification information of the system. According to the identification information, the gradient descent method is used to updating the PID Parameters, realizing PID parameters self-tuning. Simulation Model of 1MW variable pitch WPGS is built in MATLB/SIMULINK toolbox and simulation experiments of both the proposed controller and traditional PID controller are carried out under a random wind speed. The results show that RBFNN based self-tuning PID pitch controller has a good dynamic performance, possessing the advantages of response fast, small overshoot and high control precision.