Continuous-Control-Set Model-Free Predictive Control Using Time-Series Subspace for PMSM Drives
Fengxiang Wang, Yao Wei, Héctor Young, Dongliang Ke, Dongxiao Huang, José Rodríguez
Abstract
Recently, data analysis is used in model-free predictive control to mitigate the effects of parameter mismatches in parametric models. However, the finite-control-set (FCS) type cannot fully satisfy high-quality requirements due to the variable switching frequency, and it is necessary to consider the continuous-control-set (CCS) type to achieve better control performances. Nevertheless, the use of conventional time series structures in CCS model-free predictive control algorithms poses a challenge due to the complex design of control laws. To address this issue, this article proposes a CCS model-free predictive control based on a time-series subspace, which is then applied to a permanent magnet synchronous motor (PMSM) driving system. This method constructs a time-series subspace model from data and creates a suitable control law using the recursive least squares algorithm and Lagrange method without any time-varying physical parameters, to predict the future behavior of the stator voltage. The stability of the proposed method is analyzed through Bode diagrams and zero/pole maps under different conditions. A complete set of experiments proves the feasibility and advantages including improved current quality, tracking performances, and system noises compared to the conventional control strategies