Extended State Observer-Based Hierarchical Objective Optimization Model-Free Predictive Control for Three-Level NPC Inverter

Finite set model predictive control (FCS-MPC) is widely used in power converters, but its dependence on accurate models and complex selection of weighting factors can lead to performance degradation. To solve this problem, this paper proposes an extended state observer-based hierarchical objective optimization model-free predictive control for three-level neutral point-clamped (NPC) inverter without weighting factors. Specifically, extended state observer is introduced into the finite-set model predictive control to compensate for model inaccuracies, including the suppression of the neutral potential and the improvement of the robust performance. Secondly, the repeated adjustment of weighting factors is avoided by the hierarchical objective optimization mechanism. The main feature of the control method proposed in this paper is that the reliability and robustness of the controlled system can be significantly improved without knowing the exact mathematical model and parameters of the system. Finally, the effectiveness of the proposed method is experimentally verified on a three-level NPC inverter.