Torque jerk suppression for HEV mode transition based on PMSM vector control and dual-strategy coordinated control

During mode transition in Hybrid Electric Vehicles (HEVs), conventional strategies assume ideal motor torque output and employ coarse phase divisions, neglecting both the distinct dynamic response demands across the four engine start-up phases (reverse drag, ignition, torque ramp-up, stable output) and the adverse effects of motor torque fluctuations. This results in compound disturbances from command mismatch and execution error, exacerbating longitudinal jerk. To address this, a dual-strategy coordinated torque jerk suppression scheme is proposed: at the execution layer, a modified feedforward decoupling (MFD) strategy with forgetting-factor recursive least squares (FFRLS) enables real-time parameter identification and torque fluctuation suppression; at the coordination layer, two phase-adaptive model predictive control strategies—MP-MPC and MP-AMPC—are developed to match phase-specific dynamics. Co-simulation results show the integrated schemes reduce positive peak jerk from 8.079 to 2.280 m/s 3 and 2.008 m/s 3 , and negative peak jerk from −8.193 to −3.834 m/s 3 and −4.320 m/s 3 , respectively, verifying significant improvement in longitudinal transition smoothness.