Model-Free Predictive Control Strategy for NPC Three-Level Inverter-Fed Linear Induction Machine Drive System at Low Switching Frequency

Due to the large air gap, end effects, and low switching frequency, the linear induction machine (LIM) system used in urban rail transit has problems such as severe nonlinear parameter variation and significant thrust fluctuation. Although model predictive control strategies can suppress thrust fluctuation at low switching frequency through weighted switching actions, parameter identification, and other processes, there are still problems such as poor parameter robustness and limited effectiveness in suppressing thrust fluctuation. Therefore, a strong robustness model-free predictive control (MFPC) strategy suitable for the neutral point clamped (NPC) three-level inverter-fed LIM drive system is proposed. Firstly, the current and flux prediction model is established, and then a cost function with multiple objectives including primary current control, neutral point voltage balance, and low switching frequency is derived. Then, an MFPC strategy based on state observer is designed according to the ultra-local model. Finally, the simulation results show that the proposed method avoids the influence of parameters mismatch and modeling errors, effectively improving the parameter robustness and suppressing thrust fluctuation of the LIM drive system.