Modeling, Parameter Estimation, and Current Ripple Suppression for Winding-Discontinuous-Segmented PMLSM

Due to the end effect and variations in the primary–secondary coupling area, the permanent magnet linear synchronous motor (PMLSM) with discontinuously segmented windings exhibits inherent electromagnetic parameter asymmetry (EMPA). This asymmetry includes inductance imbalance and asymmetric no-load flux (NF) coupled by the segmented primary, leading to model and parameter errors when motor control is based on nominal parameters and a general model. Considering EMPA, this article investigates the modeling and parameter estimation to mitigate the current ripple. A time-varying nonlinear current-flux dynamic model is developed, based on an unconventional dq-axis reference frame, incorporating time-varying dq-axis inductance and NF, to minimize model and parameter errors. Additionally, a series of self-commissioning procedures are proposed, utilizing only the distributed drive system, to offline estimate time-varying dq-axis parameters and establish fitting models. Experimental results under six operating conditions are presented to evaluate the current tracking accuracy and residual disturbance dynamics of the robust deadbeat predictive current control (DPCC) using the improved dynamic and parameter models, validating the effectiveness of the proposed models.