Research On Current Predictive Control Of Permanent Magnet Synchronous Motor Servo System

The current loop performance of Permanent Magnet Synchronous Motor (PMSM) servo system determines the quality of the whole system. Predictive control which has excellent dynamic performance makes it become a hot topic of current control strategy for high-performance servo system. Compare with deadbeat predictive control, direct predictive control (DPC) has no need for Space Vector Pulse Width Modulation (SVPWM. Current loop adopting DPC has advantages of fast dynamic response and high bandwidth. Aiming at DPC and deadbeat predictive control for PMSM current control, the main work of this dissertation is as follows:In the vector control architecture, based on PMSM current predictive control mathematical model, this dissertation analyzes the DPC strategy for PMSM current loop in depth. The expression of optimal voltage vector which leads a minimum value of the evaluation function is deduced. The DPC process for PMSM current loop is elaborated. To overcome the poor accuracy of current tracking and weak robustness when system parameters mismatch in conventional DPC, an improved DPC that integrates integral action is proposed. Under the improved DPC strategy, the closed loop transfer functions of d、q-axis in z domain are deduced. The integral gain factors of improved DPC are obtained by configuring the desired closed-loop poles of the system. The selection of sampling time and the integral gain factors of DPC are discussed. The dissertation also analyzes deadbeat predictive current control strategy and introduces its flow chart and control block diagram.The PMSM current control models of DPC and deadbeat predictive control are built on the Matlab/Simulink platform. Aiming at conventional DPC and improved DPC, comparative simulations study on performance of current、speed control and robustness of the model parameters mismatch has done. Under the same switching frequency, comparative simulations study on current and torque control performance between the improved DPC and deadbeat predictive control has done. Finally, based on GS2000 servo driving platform, comparative experiment verification for DPC and deadbeat predictive control has done.The simulation and experimental results consistently show that:The improved DPC has improved the current tracking accuracy of d、q-axis. At the meantime, the robustness of model parameters mismatch is enhanced. Under the same switching frequency, compared with deadbeat predictive control, the improved DPC has larger bandwidth and better torque response.