Model Predictive Control of 65kW SRM Traction Motor with Offline Torque Optimization

This study investigates the application of a finitehorizon Linear Quadratic Regulator (LQR), also known as Model Predictive Control (MPC), for current regulation in a Switched Reluctance Motor (SRM) under conditions of measurement noise and uncertainties in the machine's inductance profile. To address inductance profile inaccuracies, the recursive least squares (RLS) algorithm is employed to minimize estimation errors. An offline torque optimization algorithm is implemented to enhance torque by adjusting the firing angle and shaping the reference current. The simulated SRM data, modeled using JMAG software, emulates the performance characteristics of a 65 kW Interior Permanent-Magnet Synchronous Motor (IPMSM). Comparative performance analyses using Simulink demonstrate the superiority of the proposed method over traditional hysteresis band current control.