基于遗忘因子多新息随机梯度算法PMSM参数辨识

Parameter Identification for PMSM Based on Forgetting Factor Multi-Innovation Stochastic Gradient Identification Algorithm

针对传统辨识算法结果不精确以及电机参数变化问题,提出了基于遗忘因子多新息随机梯度辨识算法。结合永磁同步电机系统电压方程,构建离散辨识模型。采用矢量控制方法控制电机,获得辨识模型输入输出数据,对转子电阻和电感参数进行在线辨识。仿真结果表明,该算法能够实现对永磁同步电机参数的辨识。

In view of the problem that the results of traditional identification algorithm are not accurate and parameters variation of motor, a new algorithm based on the forgetting factor multi-innovation stochastic gradient identification algo- rithm is proposed. Combined with permanent magnet synchronous motor （PMSM） system voltage equation, a discrete iden- tification model is constructed. By using vector control method, the input and output data of the identification model are ob- tained to identify the rotor resistance and inductance. The simulation results show that this identification algorithm can ob- tain the accurate parameters of the PMSM model.