Abstract: Aiming at the problems of large computation, obvious torque fluctuation, and prediction error due to parameter mismatch in traditional dual-vector model predictive torque control for permanent magnet synchronous motor, an improved dual-vector model predictive torque control strategy based on novel extended sliding-mode disturbance observer is proposed. Firstly, in order to reduce the control computation, a fast voltage vector selection tableis proposed with a graph of 12 sector voltage vector, the number of iterations is greatly reduced from 14 to 3. Then, to address the undesirable effects of motor parameter mismatch, the stator inductance value is estimated by the amount of error variation between the predicted and actual measured values at adjacent moments, the inductance updating mechanism is integrated, and a novel extended sliding-mode observer is proposed to enhance the robustness of the prediction model in the case of parameter mismatch. The MATLAB/Simulink simulation results show that the proposed strategy can reduce the rotor torque volatility, effectively reduce the unfavorable influence of the parameters change, and improve the stability of the prediction model.

Key words: permanent magnet synchronous motor(PMSM), model predictive torque control, extended sliding-mode observer, parameter mismatch