Multiparameter and VSI Nonlinearity Identification for PMSM Drives with Dual Extended Kalman Filter and Discrete Optimal Resonant Observation Loop

Precise electrical parameter identification is becoming an essential technology in transportation electrification due to the increasing demands for fault diagnosis and online monitoring. However, in the permanent magnet synchronous machine (PMSM), multiparameter identification encounters significant challenges due to the dead-time effect (DTE) resulting from voltage source inverter (VSI) nonlinearity and the degraded signal-to-noise ratio under low-speed conditions. To address this issue, this paper proposes a method that integrates an observation loop and a compensation loop to identify the DTE-induced distortion voltage. By coordinating high-frequency (HF) small-signal injection with the dual extended Kalman filter (DEKF), simultaneous multiparameter identification is achieved. Compared to traditional methods, the proposed method eliminates reliance on precise device parameters or polarity detection, avoids introducing additional delay, and exhibits superior steady-state and transient performance. Notably, it achieves reliable identification even at zero speed and zero load with all parameter initial values set to zero. Furthermore, the design procedure and implementation flowchart of the proposed method are systematically analyzed. Finally, experimental results comprehensively validate its effectiveness under various critical operating conditions.