Model-Free Predictive Control Strategy for Inverter Based on Adaptive Super- Twisting Sliding Mode Observer

The application of model predictive control (MPC) in inverters needs to consider the influence of parameter accuracy. The mismatch between the modeling circuit parameters and the actual circuit parameters will restrict the performance of the traditional model predictive control. Model-free predictive control (MFPC) using ultra-local models has been extensively studied. However, the model-free predictive control of algebraic methods requires a large amount of historical voltage and current sampling data, resulting in slower dynamic characteristics. Therefore, this paper proposes a model-free predictive control method based on super-local model and adaptive super-twisting sliding mode observer. Firstly, the influence of parameter variation on the conventional model predictive control is analyzed, and the first-order ultra-local model of the inverter is established. Secondly, a super-twisting sliding mode observer with an adaptive factor is used to estimate the dynamic part of the ultra-local model. Finally, in the simulation, the proposed method is compared with MPC and algebra-based MFPC to verify the effectiveness and superiority. The results show that the proposed method can significantly reduce the current THD value in steady state and shorten the adjustment time in dynamic state.