Model Inaccuracy Analysis and Compensation of Stationary Frame-Based Deadbeat Predictive Current Control for High-Speed PMSM Drives

Abstract

Due to explicit modeling of delay effects and fast dynamic response, the recently reported stationary frame-based deadbeat predictive current control (DBPCC) is particularly attractive for high-speed permanent magnet synchronous machine (PMSM) drives. However, its actual control performances would be heavily dependent on model accuracy. Therefore, this article presents an analytical and systematical investigation into the influences of model inaccuracy on the stationary frame-based DBPCC. Steady-state and transient performance degradations under machine parameter mismatches and inverter nonlinearity over the whole speed range have been studied. Importantly, the analyses show that the steady-state control errors will increase greatly under parameter mismatch as the speed rises; while the transient current tracking errors, only caused by inductance mismatch, can be quickly suppressed after two time-steps. To further improve the parameter robustness of the stationary frame-based DBPCC for high-speed drives, this article also proposes a simple but effective method, called adaptive reference correcting current injection (ARCCI). The proposed method can effectively eliminate all the steady-state current errors caused by model inaccuracy. It can be easily implemented and applicable at high speeds with low switching-to-fundamental frequency ratios (SFRs) as well. Both the presented analyses and compensation methods have been validated by extensive simulations and experiments.