Fault recognition of multilevel inverter using artificial neural network approach

Abstract

This thesis focuses on the development of a diagnostic tool for newlinedetecting IGBT power electronic switch flaws caused by both open and short newlinecircuit faults in multi-level inverter time-frequency output voltage newlinespecifications. High-resolution laboratory virtual instrument engineering newlineworkbench software testing tool with a sample rate data collection system, as newlinewell as specialized signal processing and soft computing technologies, are newlineused in this proposed method. On a single-phase cascaded H-Bridge newlinemultilevel inverter, simulation and experimental investigations of both open newlineand short issues of the IGBT components are performed out. In all newlineconceivable switch issues, the output voltage signals are evaluated for newlinedifferent modulation index values. Fast Fourier Transform and Discrete newlineWavelet Transform methods are used to study the frequency domain newlineproperties of output voltage signals. In the Artificial Neural Network, the newlineBack Propagation Training technique was employed, and the generated neural newlineparameter values were used in the LabVIEW real-time fault diagnosis model. newlineIn the area of high-power implementations, multilevel inverters newline(MLIs) can be the most successful system, and they have been exploited in a newlinebroad range of industrial purposes that need less working costs and higher newlinequality output like electric cars, drive systems, voltage profile compensators, newlinevoltage regulators, sustainable energy applications, etc. The MLI is used newlineamongst the most common and dependable categories of MLI, in which the newlinestair cased pulse-width synthesized inverter voltage output could be made newlinethrough a number of cells that are connected in series H-bridge voltage cells. newline