Sensorless Control of a Permanent-Magnet Synchronous Motor Based on Incremental Deadbeat Predictive Current Control and an Adaptive Sliding-Mode Observer

To address the strong dependence of conventional deadbeat predictive current control on motor parameters and the effect of observation errors on rotor-angle estimation during sensorless operation, this paper investigates a permanent-magnet synchronous motor (PMSM) control method that combines incremental deadbeat predictive current control with an adaptive sliding-mode observer. First, an incremental predictive model is derived by taking the difference between the discrete current equations of two consecutive sampling periods, thereby reducing the influence of parameter mismatches in the permanent-magnet flux linkage and stator resistance on predictive control. In addition, the one-sampling-period delay inherent in digital control is taken into account, and compensating voltage commands are generated using historical current and voltage information. Subsequently, a sliding-mode observer based on adaptive back-electromotive-force estimation and a phase-locked loop is designed, and its stability is analyzed using a Lyapunov-based approach. Simulation results show that, under the permanent-magnet flux-linkage and stator-resistance mismatch conditions considered in this study, the proposed incremental deadbeat predictive current control achieves a smaller steady-state q-axis current deviation than conventional deadbeat predictive current control. Under the specified speed-tracking and sudden-load-application conditions, the adaptive sliding-mode observer exhibits lower speed ripple and a smaller steady-state rotor-angle estimation error than the conventional sliding-mode observer, while its estimated speed tracks the actual motor speed more closely.