Similarity-Based Fast Open-Circuit Fault Diagnosis Method for Modular Multilevel Converters

Abstract

Due to the large number of semiconductor devices within modular multilevel converters (MMCs), MMCs will encounter higher possibility of fault occurrance than other two-level converters. Therefore, it is essential to apply fault diagnosis to enhence the reliability of MMCs. An rapid, real-time, and submodule (SM) capacitor voltage similarity-based open-circuit fault diagnostic method for MMCs is proposed in this paper. The presented fault detection and location (FDL) scheme utilizes the submodule capacitor voltage waveforms under both normal and fault conditions to conduct the similarity analysis. The voltage of capacitor within the faulty submodule will vary from those within normal SMs because of the lost of the discharging circuit caused by the open-circuit fault. The correlation coefficients of the capacitor voltages within the same arm are designed to identify this feature. Taking the advantage of correlation coefficients, the open-circuit fault can be identified rapidly to avoid the capacitor voltage of failed SM being charged to a high level. The proposed FDL method do not require extra data and only use the sampled signals that are already available for the control algorithm of MMCs. Thus, the proposed approach can be implemented without additional measurement or hardware circuitry. The proposed FDL method is validated by experimeantal results which indicate that the open-circuit fault can be detected and indentified quickly and accurately within one fundamental period.