Model-Free Predictive Current Control of PMSM Based on Sliding Mode Disturbance Observer

Abstract

Model predictive current control (MPCC) based on sliding mode disturbance observer (SMDO) effectively addresses the issues arising from parameter mismatches and unmodeled nonlinear elements in Permanent magnet synchronous motor (PMSM). However, conventional sliding mode faces a trade-off between reaching speed and chattering, and their closed-loop transfer functions are often challenging to derive. To overcome these limitations, this paper proposes a model-free predictive current control method based on an adaptive sliding mode disturbance observer (ASMDO-MFPCC). By exploiting a novel decoupled PMSM ultra-local model and a novel sliding mode control law, a discrete-time observer is designed to enhance convergence speed and suppress chattering. The stability of the observer is analyzed, and its parameter ranges are theoretically determined. Moreover, by linearizing the control law near the designed sliding surface, a complex vector linear model is established for the entire system, allowing closed-loop transfer function analysis to confirm that deadbeat performance can be preserved under accurate parameter settings. To further improve the system’s robustness, a simple yet effective inductance identification algorithm is proposed for real-time calibration of the predictive control system. This algorithm operates independently of the observer design. Experimental results validate the effectiveness of the proposed current control scheme.