Dual Time Scale Weighted Recursive Least Squares for IPMSM Parameter Identification with Inverter Nonlinearity

This research paper delves into the realm of parameter identification for Interior Permanent Magnet Synchronous Motors (IPMSMs), with a particular focus on the influence of inverter nonlinearity. A novel online parameter estimation approach is presented, integrating the dual time scale (DTS) methodology with the weighted recursive least squares (WRLS) algorithm. The DTS framework strategically partitions motor parameters into two groups, tailored to their inherent dynamic characteristics, thereby significantly reducing computational complexity while ensuring system stability. Within this framework, the WRLS algorithm, deployed during system steady-state conditions, mitigates the risk of convergence to local extrema by adaptively assigning weights to parameters based on their confidence levels suppressing the effects of noise. Furthermore, the modeling process meticulously accounts for often-neglected VSI nonlinearities, such as dead time, switching time, and IGBT voltage drop, thereby enhancing the accuracy of the estimation model. Extensive simulations are conducted, analyzing the impact of VSI nonlinearity compensation and evaluating the efficacy of the proposed DTS-WRLS method. The outcomes underscore the superior performance of the proposed approach in enhancing parameter identification accuracy and stability.