BP Neural Network Based Research On Servo Control System Of Permanent Magnet Synchronous Motor

Servo system has been widely used in various sectors of the national economy.With the development of economy and social progress,the requirements for the quality of servo system control are constantly improving.Therefore,the research of high-performance servo control algorithm has important theoretical and practical significance.In this paper,permanent magnet synchronous motor(PMSM)is taken as the research object.Based on the principle of vector control,the non-linear problem of PMSM is studied.BP artificial neural network is used to control PMSM.The main work is as follows:(1)The mathematical model of permanent magnet synchronous motor(PMSM)is analyzed,and it is clarified that PMSM is a non-linear system.Variation of motor parameters,external load disturbance and unmodeled dynamics in practical application will affect the servo performance of permanent magnet synchronous motor(PMSM).Then the common servo control schemes are analyzed and summarized.It is found that the common servo control schemes are not completely applicable to the control of non-linear systems.Therefore,the artificial neural network algorithm is introduced to solve the problem of non-linear control.(2)By studying the vector control method,the mathematical expressions of PMSM in D and Q coordinates are analyzed,and the inverse dynamic model of PMSM vector control in discrete time is obtained,and then an artificial neural network controller is constructed based on the inverse dynamic model.The artificial neural network controller is trained by back propagation(BP)algorithm.Its input and output are motor speed,q-axis current and the values of their adjacent sampling time.The training speed is accelerated by off-line learning method,and the robustness is improved by on-line learning method.(3)A servo control experimental platform is built,which has complete functions,can run with 220 V AC power supply,and can drive two motors at the same time.The platform software is complete,and the host computer software and debugging software are self-made.Speed tracking experiments on the experimental platform show that the proposed control algorithm has good speed tracking effect and strong robustness.In the case of sudden changes in speed input,load,inertia and resistance of stator circuit,no static error tracking of speed can be realized,which shows that the method is feasible and superior.