Power converter system fault diagnosis based on AI tech

Abstract

To improve the working stability and reliability of a three-phase converter, this article presents a novel method for detecting IGBT open-circuit faults in a three-phase two-level power converter. The method employs a combination of the Extreme Learning Machine (ELM) and Whale Optimization Algorithm (WOA) and uses simulation data of the converter's output current to train the WOA-ELM model. The WOA algorithm is used to determine the optimal weight and bias matrix for the ELM, leading to a fault diagnosis model with high accuracy and efficiency. An optimal time window is also incorporated to balance diagnostic speed and accuracy. Additionally, the proposed approach is robust to voltage ripple, harmonics, speed, and load fluctuations, making it a reliable and practical solution for fault diagnosis in power converters.