Model Predictive Control Method Of PSMS Based On Data-Driven Approach

Permanent magnet synchronous motor is widely used in AC servo system because of its high efficiency and high power density.With the expansion of its applications,the industry has more strict requirements for its stability and security.At present,MPC algorithm is regarded as a potential high-performance motor control algorithm because of its ability to handle multiple input and multiple output variables and additional constraints.However,the performance of the algorithm can not be independent of the controlled object model.The inaccurate model will lead to errors in the system prediction,which will have a negative impact on the dynamic and static performance of the system.This paper studies the electromagnetic parameter mismatch that may occur during the operation of the motor.Firstly,the classical MPC algorithm is applied to the design of motor drive controller,and the adjustable parameters of MPC controller are simulated and analyzed according to the system optimization objective.Based on this,a data-driven MPC strategy is proposed based on the classical MPC algorithm.Using the historical data of each variable of the system to build the motor prediction model,the prediction error can also be reduced in the case of parameter mismatch,so as to improve the performance of the system.Furthermore,the Lyapunov method is used to analyze the stability of the DDMPC controller,and the boundary conditions for the stable operation of the system are defined.Finally,under different working conditions,the DDMPC controller with different range of historical data is simulated and analyzed.The results show that with the scope of collecting historical data grows,the controller has a stronger ability to resist parameter mismatch.Compared with the classical MPC,DDMPC can eliminate the static error of the PMSM control system and improve the control performance of the system,which verifies the superiority and feasibility of the algorithm.