Fault detection and prognosis for multilevel inverter

Abstract

Multilevel inverters have become increasingly popular due to their numerous advan tages, such as providing high-voltage generation and improved power quality. However, due to their complexity and the presence of numerous components, they are more likely to suffer faults and failures. The purpose of this dissertation is to develop efficient fault detection and prognosis techniques for multilevel inverters employed by Non-Intrusive Load Monitoring (NILM) applications. This study aims to identify and predict failure by analyzing the output voltage of inverters, ensuring reliable and uninterrupted power supply. The proposed methodologies make use of advanced signal processing techniques, machine learning algorithms, and statistical analysis to accurately detect and diagnose faults while also predicting the remaining useful life of faulty components.