The Intelligent Control Of The Nonlinear PID Nueral Network Of The BLDC

Since the brushless DC motor (BLDC) has the characters such as simple structure,convenient maintenance and so on, so it has been widely used in the fields of nationaldefence, aviation and industrial process control. In nowadays, the researches for thedesign of electrical motor and strategy of advanced control have a great profit botheconomically and socially.Brushless DC motor speed control system is nonlinear, multivariable and closecoupling, while neural network has the advantages such as selflearning function,strong adaptability and simple structure. These advantages have a fine effect on thetarget which is nonlinear and multivariable. And the traditional PID regulator also hasthe characters of simple and convenient.To solve these problems, on the basis of Gauss function neural network, thisthesis puts forward a nonlinear PID control algorithm. At first, it used the Gauss formto instead the curves which are vary with the errors of the traditional PID’s proportion,integration and differential coefficients. Thus, to create a nonlinear PID control lawthat is depend on the Gauss function neural network. Then, training the three weightcoefficients of the PID online through the neural network. In this way, the nonlinearPID intelligent control of the BLDC can be realized.At last, the thesis established a model of BLDC’s double closed control systemon the basis of Matlab/Simulink. The traditional PID control strategy and thenonlinear neural network PID control strategy，which the thesis puts forward，haveused on the control speed of the outer race. The simulation result shows that thecontrol algorithm which the thesis puts forward has a clear advantage than thetraditional PID control.