Reliable Online Fault Detection and Fault-Tolerant Control of Grid-Connected Modular Multilevel Converters Based on Active Disturbance Rejection and Improved Model Predictive Control

Abstract

A Modular Multilevel Converter (MMC) represents an emerging topology for high-power and high-voltage energy conversion systems, due to its modularity structure, which is designed on the cascade connection of Sub-Modules (SMs). Its use of a large number of semiconductor devices increases the probability of failure of SMs during operation. The reliability of systems using SMs has become the obsession of researchers in this field. Rapid diagnosis and accurate fault location are essential to ensure continuous system operation. In this paper, a comprehensive fault-tolerant diagnostic and control solution, including fault detection and location, fault tolerance, and re-configuration of the MMC in case of faults, has been proposed to overcome Open Circuit (OC) and Short Circuit (SC) faults of IGBTs. This method uses a computational method based on three separate algorithms, combined to give the optimal solution and minimize computation time. This solution improves both the reliability and availability of the system while keeping the capacitor voltages of the SMs healthy. Rigorous simulations are performed to demonstrate the validity and effectiveness of the proposed control technique.