Artificial Intelligence for Open-Circuit Switch Fault Detection in Modular Multilevel Converters: A Comprehensive Review

Modular Multilevel Converters (MMCs) have become increasingly popular in medium- and high-voltage power applications due to their highly scalable modularity and excellent harmonic performance. However, the massive number of semiconductor switches in the MMC architecture significantly increases the probability of submodule (SM) failures, particularly open-circuit (OC) faults. Fast and accurate fault detection and diagnosis (FDD) is crucial to prevent catastrophic secondary damage and ensure continuous system reliability. While conventional FDD approaches primarily hardware-based and model-based methods have been widely utilized, they often suffer from the requirement for expensive additional sensors or a heavy reliance on complex, parameter-sensitive mathematical models. Recently, data-driven Machine Learning (ML) and Deep Learning (DL) techniques have emerged as powerful alternatives. By utilizing advanced feature extraction methodologies, these intelligent techniques offer rapid and highly accurate diagnosis without requiring precise analytical system modeling. This paper presents a concise and comprehensive review of OC fault diagnosis strategies in MMCs, with a primary focus on ML- and DL-based methodologies. The conventional and modern methods are categorized and critically compared in terms of computational complexity, diagnosis speed, and sensor requirements, concluding with a discussion on the challenges and future trends of intelligent FDD in power electronics.

کلیدواژه ها:

Modular Multilevel Converter (MMC) ، Open-Circuit Fault ، Fault Diagnosis ، Artificial Intelligence ، Machine Learning ، Power Electronics Reliability