Model-Free Predictive Control of IPMSM with Adaptive Voltage Vector Coefficient Based on a Predictor-Based Neural Network

In model predictive control (MPC), the voltage vector coefficient in the predictive model is closely linked to the model parameters, significantly affecting control performance. However, existing model-free predictive control strategies neglect this dependency on model parameters. To address this issue, this paper proposes a novel adaptive model-free predictive control strategy of interior permanent magnet synchronous motors (IPMSMs) with online updating of voltage vector coefficient. The proposed method employs an online predictorbased neural network (PNN) to estimate the system function, while updating the voltage vector coefficient through a stochastic approximation (SA) algorithm. It can significantly enhance the robustness and reliability of the control system in IPMSM under parametric uncertainties. Finally, the effectiveness of the proposed control strategy is validated through experiments conducted on an IPMSM platform driven by a three-level converter.