Fault-Tolerant Voltage Source Inverter for Induction Motor Drive Using Intelligent Techniques

Abstract

Electrical systems can experience faults due to various reasons. These faults may occur as a result of aging components, conditions of use, or manufacturing defects that were undetectable during commissioning. The faults can be broadly classified into two categories: those that occur within the electrical machine such as winding faults and axis tilt, and those that occur outside the electrical machine in the drive chain, such as faults in the mechanical gearbox. One major area of research is focused on monitoring the state of the converter that supplies power to an asynchronous machine. A converter, such as a Pulse Width Modulation (PWM) inverter, may have structural defects like malfunctioning switches (semiconductors) that could damage the entire system. It’s crucial to invest in malfunction detection to prevent such defects from causing irreparable harm to the system. Artificial intelligence (AI) based condition monitoring and diagnosis techniques offer various advantages over traditional methods. These techniques eliminate the need for mathematical models, which reduces engineering and development time significantly. AI-based techniques rely on system datasets or expert knowledge to make accurate predictions. In the case of controlling an induction motor powered by a PWM voltage inverter, AI-based methods can detect open circuit and/or short circuit faults and take corrective measures to ensure the system operates efficiently while maintaining the required level of security. In this thesis, we analyzed the feasibility of using artificial intelligence techniques in detecting, diagnosing, and reconfiguring faults in a three-phase inverter that powers an induction motor. We provided a detailed description of inverter switching faults and developed a simple method to extract characteristics for studying the possibility of detecting and diagnosing these defects. We also attempted to reconfigure the inverter system to prevent faults from occurring.