Model-Free Predictive Current Control for Induction Motor Drive Using Recursive Least Squares Algorithm

A conventional model predictive current control (MPCC) scheme for an induction motor (IM) drive performs as an effective controller at dynamic and steady-state conditions. However, the performance of this control scheme is highly dependent on the mismatch in the motor parameters. In addition, the modeling knowledge of the IM and its accurate parameters measurement are required in the controller. In this paper, a new model-free PCC (MF-PCC) scheme has been proposed to overcome the aforementioned problems. The proposed scheme uses an auto-regressive exogenous (ARX) model instead of the motor model to form a model-free controller. The ARX model works only on the motor's input-output transfer function. The unknown coefficients or parameters in the ARX model are estimated by using a recursive least squares algorithm. To obtain an accurate prediction, an ARX structure has been modified into an observable canonical state-space model. Simulation results show that the proposed control scheme is robust against motor parameters mismatch under different operating conditions in comparison with the conventional MPCC scheme.