A Wavelet Characters based Adaptive Prediction Horizon for Multi-Step Predictive Torque Control on PMSM Drives

The realized control performance and development for the predictive control are limited by the short prediction horizon (PH). To solve this problem, a multi-step predictive torque control (MPTC) with an adaptive prediction horizon based on continuous wavelet transform (CWT) is proposed in this paper, and applied to a permanent magnet synchronous motor (PMSM) driving system as a torque controller. The prediction horizon is auto-tuned by the obtained clustering character of CWT, and the final value of the prediction horizon is calculated based on the current operating state. The stability of the long prediction horizon for the proposed method is analyzed and demonstrated on the discrete-time domain. The correctness is demonstrated by validations, as well as advantages on speed integrated time and absolute error (ITAE) and stator current total harmonic distortion (THD).