Multi-Parameter Online Identification of Permanent Magnet Synchronous Motor Based on Dynamic Forgetting Factor Recursive Least Squares

To solve the problems of degenerate-rank, slow convergence and low identification accuracy of recursive least squares with forgetting factor in permanent magnet synchronous motor (PMSM) multi-parameter identification model, an online identification method based on dynamic forgetting factor recursive least squares (D-FFRLS) is proposed for PMSM. First, a full-rank parameter identification model is established to realize the online identification of the inductance, stator resistance and permanent magnet flux linkage. Then, the forgetting factor that changes dynamically with the error of the identification model is used to improve the convergence of parameter identification. Finally, the control algorithm is verified by simulation experiments. The results show that the proposed control algorithm can realize multi-parameter online identification within 0.5ms when the parameters of the motor change slowly or abruptly at the same time, or when a single parameter suddenly changes under the rated operating condition. In addition, it has good anti-interference and robustness because of adaptive adjusting dynamic forgetting factor.