Improved Dual-Vector Incremental Model Predictive Voltage Control with Adaptive Parameter Identification for Permanent Magnet Synchronous Motor Drives

To address the limited candidate voltage vectors, insufficient voltage-tracking capability, and parameter sensitivity of conventional model predictive voltage control (MPVC) for permanent magnet synchronous motor (PMSM) drives, this paper proposes an improved dual-vector incremental model predictive voltage control strategy with adaptive forgetting factor recursive least squares (IDV-IMPVC-AFFRLS). First, an incremental voltage prediction model is developed to eliminate the explicit dependence on permanent-magnet flux linkage and reduce parameter dependence to stator resistance and inductance. Second, six virtual voltage vectors are synthesized to expand the candidate set from eight to 14, and a Lagrange-multiplier-based dual-vector duty-cycle optimization is employed to improve reference-voltage approximation. Third, an adaptive forgetting factor recursive least squares (AFFRLS) algorithm based on the prediction-error norm is introduced to identify and update the stator resistance and inductance online. Experimental results show that, under nominal parameters, IDV-IMPVC reduces phase-current THD from 16.57% to 8.45% and torque ripple from 0.10 N·m to 0.04 N·m, corresponding to reductions of 49.0% and 60.0%, respectively. Under resistance and inductance mismatches, IDV-IMPVC-AFFRLS limits THD to 9.15% and torque ripple to 0.04 N·m, demonstrating improved steady-state performance and parameter robustness.