Adaptive Model Predictive Control with a Torque-Reactance Observer for High-Efficiency PMSM Drives in Battery Electric Vehicles

Effective power electronic battery management is crucial for enhancing the performance and lifespan of battery electric vehicle (BEV) systems. This paper presents a modified adaptive torque and reactance parameters linear model predictive (ATRP-LMP) approach. The proposed method aims to improve motor drive efficiency and optimize battery discharge in BEVs. Specifically, ATRP-LMP controls the permanent magnet synchronous motor (PMSM) within the BEV powertrain to achieve the desired speed under the required traction torque. The developed wheel torque is considered an unmeasured input; therefore, it is estimated using an average third-order generalized integrator (ATOGI) flux motor observer. As a result, ATOGI incorporates stator reactance parameters estimated from the motor flux. Simulation results demonstrate that the proposed ATRP-LMP significantly reduces battery discharge current, watt-hour per kilometer, capacity rate, and the number of parallel cell units compared to the classical LMP approach while increasing battery lifetime and vehicle efficiency. Specifically, the ATRP-LMP achieves a 9.6% reduction in average current draw, extends battery discharge time by 86%, reduces energy consumption by 15%, decreases the required number of parallel battery units by 16%, and lowers the discharge rate by 10%, demonstrating its effectiveness in optimizing battery utilization and extending operational range.