Parameter-Free Model Predictive Control With Series Structure for Permanent Magnet Synchronous Motors

To balance the computational burden, switching frequency, and steady-state performance of traditional single-step model predictive control (MPC), and to improve the parametric robustness of MPC, a novel two-step MPC strategy with cascading structure is proposed. First, the preselection process of alternative voltage vectors is analyzed in detail by calculating the possible optimal voltage vector and suboptimal voltage vector at the next two prediction times. Second, the predicted current value of the next two moments is calculated according to the alternative voltage vector obtained in the previous step. The predicted current of the next two moments is calculated using the same voltage vector, and the optimal output voltage vector is determined by the design cost function. Both the reference voltage model and the predicted current model in the whole structure are reconstructed based on the ultra-local model, and the unknown terms are estimated using an improved sliding mode observer. Finally, the proposed cascade MPC scheme is verified by experiments, and the experimental results prove its effectiveness.