Modified model predictive torque control for induction motors with improved robustness against mutual inductance mismatching

Abstract

This article proposes a new model predictive torque control (MPTC) for induction motors to improve its robustness against parameter mismatching when current model (CM) for flux estimation is employed. The main advantages of the proposed method are the improved range of stable operation during the occurrence of mutual inductance Lm, mismatched value, lower computational burden, and reduced switching losses. The improvement in robustness is achieved via adaptation of direct torque control (DTC) features thus eliminating any additional parameter estimation mechanism. Reference transformation is used to remove torque error from the cost function to achieve weighting-factor-free MPTC formulation. Then, based on the estimated and reference stator flux vector positions, it is established that flux positional error is directly proportional to the torque error. The positional error is then employed to reduce the admissible voltage vectors, which decreases parameter dependence of MPTC. The simulation results are compared to two other well-established MPTC techniques and based on the results, it is concluded that the proposed MPTC without parameter estimation and compensation technique have a wider tolerability of 56% for Lm mismatching.