扩展卡尔曼滤波参数辨识下永磁同步电机模型预测转矩控制

Model prediction torque control of PMSM based on extended Kalman filter parameter identification

为了提高模型预测转矩控制的稳态性能,提出一种空间矢量脉宽调制(SVPWM)策略下的无差拍模型预测转矩控制。针对定子电阻与电感对无差拍算法控制效果的重要影响,使用扩展卡尔曼滤波器对模型中的定子电阻、电感参数同时进行在线辨识,以提升算法的参数鲁棒性。首先,对电磁转矩和定子磁链幅值使用无差拍控制,解出两相旋转坐标系下的参考电压矢量,并送入SVPWM进行调制,从而有效抑制了转矩脉动;然后,推导出定子电阻与电感出现不可抗力误差时,dq轴输出电压存在严重偏离理想值的问题;最后,使用扩展卡尔曼滤波器对模型中的电感和电阻参数同时进行在线辨识,并将辨识结果实时反馈至模型中。通过仿真和实验将所提算法的稳态性能、动态性能及参数鲁棒性进行了对比分析,从而证实了所提算法的优越性。

In order to improve the steady-state performance of model predictive torque control,a deadbeat model predictive torque control based on space vector pulse width modulation(SVPWM)modulation strategy was proposed.The important influence of stator resistance and inductance on the control effect of the deadbeat algorithm was analyzed.To solve this problem,extended Kalman filter was used to identify the stator resistance and inductance parameters in the model simultaneously to improve the parameter robustness of the algorithm.Firstly,the reference voltage vector in the two-phase rotating coordinate system was obtained by using the deadbeat control of electromagnetic torque and stator flux amplitude,and then sent to SVPWM for modulation,so as to effectively suppress the torque ripple.Then,it was deduced that the output voltage of dq axis deviates seriously from the ideal value when the stator resistance and inductance have force majeure error.Finally,the inductance and resistance parameters in the model were identified online by using the extended Kalman filter,and the identification results were fed back to the model in real time.The steady-state performance,dynamic performance and parameter robustness of the proposed algorithm were compared with the published results by simulation and experiment,which proves the superiority of the proposed algorithm.