Generalized Unbiased Parameters Identification for PMSM Based on Information Enrichment and Convergence Correction

Accurate multiparameter estimation in permanent magnet synchronous motors (PMSMs) is essential for high-performance control. However, gradient-descent-based methods, widely adopted for this purpose, are inherently biased, and the simultaneous identification of multiple parameters without external signal injection remains challenging. This work introduces an information enrichment strategy and a convergence correction method to overcome these limitations. An evaluation function is first established to quantitatively relate the information content of experimental conditions to estimation accuracy, providing a rigorous foundation for multiparameter identification. By adjusting the drive system and implementing a customized data preselection strategy, the proposed approach ensures the acquisition of information-rich data necessary for accurate estimation of four key PMSM parameters. Moreover, this study analytically derives, for the first time, the inherent bias in gradient-descent-based estimators within the PMSM. Leveraging these findings, a generalized, unbiased, and iterative identification framework is proposed to enable high-precision multiparameter estimation. Comparative validations confirm the effectiveness and robustness of the proposed method.