Compound control and parameter selftuning of permanent magnet synchronous motor
                                                                               HE Yunbo,CHEN Jiajun
(School of Mechanical and Electrical Engineering,Guangdong University of Technology,Guangzhou 510006,China)

Abstract: Aiming at the deficiency of conventional proportionalintegraldifferential (PID) controller and the inconvenience of manual adjustment of PID parameters, the feedforward control structure and radial basis function (RBF) neural network were researched. A compound PID control structure of "threeclosedloop + feedforward" was proposed.RBF neural network was used for online identification of control system and gradient descent method was used to automatically adjust the PID parameters of the controller. Then contrastive experiments of conventional threeclosedloop PID controller and "threeclosedloop + feedforward" compound PID controller and selftuning experiment of position loop PID parameters were carried out on the experimental platform. The results indicate that compared with the conventional PID control structure, the position response performance of the compound control structure is improved by 12% and the speed response performance is improved by 31%. In addition, the selftuning of the PID parameters can be realized by using RBF neural network and the tuning effect is good.

Key words: compound control; RBF neural network; gradient descent method; PID selftuning