LAUSR

The poor parameter robustness of current loop is the main limitation for traditional deadbeat predictive current control (DPCC) to become a high-performance control method. Among various improvement methods aimed at addressing this issue, the incremental prediction model has become a research hotspot as a simple and feasible method to eliminate magnetic flux. Based on this, this article proposes a parameter adaptive incremental model-based DPCC (PAIDPCC). Unlike other methods based on incremental model, the proposed method takes the prediction error of IDPCC as the objective function and combines the finite-time gradient descent (FGD) method to adaptively update the inductance in real time. It not only eliminates the impact of magnetic flux and resistance parameter mismatch but also eliminates the impact of inductance parameter mismatch, achieving robustness to all parameters. Moreover, a deadbeat predictive speed controller (DPSC) instead of a proportional–integral (PI) controller is used as the speed outer loop, improving the dynamic responsibility. Finally, the strong parameter robustness and fast dynamic response capability of the proposed method were verified on a 1-kW surface permanent magnet synchronous motor (SPMSM) test bench.