基于超螺旋积分滑模观测器的无模型预测电流控制在ANPC逆变器驱动PMSM中的应用

To address the issue of control accuracy degradation caused by parameter mismatch in a permanent magnet synchronous motor (PMSM) driven by an active neutral point clamped (ANPC) three-level inverter, a model-free predictive current control strategy based on a super-twisting integral sliding mode observer (MFPCC-STISMO) is proposed. A parameter-independent first-order ultralocal model is constructed to replace the traditional PMSM mathematical model, and an STISMO with an integral sliding mode surface is designed to estimate the lumped disturbance, thereby suppressing the impact of parameter mismatches on motor control performance. Meanwhile, in the neutral point voltage balancing method, the traditional three-level model predictive current control (MPCC) relies on weighting factors, so a neutral point voltage balancing strategy without weighting factors is proposed. By leveraging the characteristic that a pair of positive and negative small vectors in the three-level inverter generate neutral point currents with equal magnitude and opposite directions, the complicated weighting factors adjustment process is eliminated through reasonable allocation of small vector action times. Finally, comparative experiments with traditional methods on an experimental platform show that the proposed method exhibits excellent dynamic and steady-state performance, and effectively enhances the disturbance rejection capability of the PMSM.