Inverter Fault Diagnosis Based on PNN Neural Network

Abstract

A three-phase inverter is widely used in every aspect of life and production, and fault diagnosis becomes more important. In this paper, the working characteristics and process of the traditional three-phase three-wire inverter when an open-circuit fault occurs are analyzed in-depth, and the Simulink model is built to simulate it. At the same time, the M language written by MATLAB software is used to build the diagnosis model of single-phase open-circuit fault of three-phase inverter IGBT based on a PNN neural network. The first 20 harmonics of the three-phase output voltage were selected as the eigenvalues to construct the input training samples, which were trained to make them have a certain ability of fault diagnosis.