Model-Free Predictive Current Control for Voltage Source Inverter: A Comparative Investigation

This study presents a comprehensive comparative investigation between the Model-Free Predictive Control (MFPC) approach based on an Ultra-Local Model (ULM) and conventional Model Predictive Control (MPC) within the context of a two-level voltage-source inverter (2L-VSI). The MFPC strategy harnesses historical data and the ULM for enhanced adaptability in uncertain scenarios. Simulation results reveal that the MFPC approach, employing Algebraic Identification Approaches (API), Extended State Observer (ESO), and Integral sliding mode observer (ISMO), maintains commendable performance under steady-state conditions, even when encountering parameter mismatches. Notably, during quick changes in current reference, the MFPC approach, specifically MFPCC-ISMO, showcases superior dynamic performance. Furthermore, when contrasted with MFPC and conventional MPC methods, the ULM-augmented strategy exhibits notable superiority. The incorporation of the ULM substantiates the MFPC's efficacy, supporting robust power conversion systems conducive to the transition to cleaner and sustainable energy solutions.