Computationally Efficient and Loss-Minimizing Model Predictive Control for Induction Motors in Electric Vehicle Applications

This paper introduces a loss-minimizing Model Predictive Control (MPC) strategy for induction motors in electric vehicle applications designed to track a specified speed reference. The proposed control incorporates three key features that enhance efficiency and minimize losses. Firstly, an inverter selection vector strategy minimizes electromagnetic torque ripple, additional inverter switching frequency, and computational cost. Secondly, every element in the proposed control is based on the induction motor model, including consideration for iron losses. Thirdly, the MPC stator flux reference is optimized for total electric loss minimization, given any electromagnetic torque and mechanical speed reference, with no additional computational cost. The loss-minimizing function is derived from the induction motor model and accounts for all motor losses, including iron losses. Its straightforward implementation and pre-computed algebraic form ensure easy integration into various systems while reducing real-time computational overhead. The proposed control is tested and compared to a classical MPC through dynamic case studies, demonstrating satisfactory results in reducing total electric losses and electromagnetic torque ripple. During testing for electric vehicle applications within relevant standardized urban driving cycles, the proposed control showcases excellent energy efficiency results, reducing total electric losses by 49% compared with classical MPC.