Efficient Fault Detection for Open Circuit Faults in HANPC Inverters Using Artificial Neural Network for Motor Drive Applications

Abstract

Open circuit faults are regarded as common faults that affect switching devices of the power converters. The early detection of these faults is a vital task in order to protect the other switching devices and the associated appliances such as motor drive systems. However, the conventional methods require different hardware devices to detect the variation in the current and voltage waveforms. Therefore, in this paper a creative method that use artificial intelligence by performing artificial neural network (ANN) to detect and identify the faulty devices effectively and precisely without the need to adapt additional hardware or modifying the original configuration of the inverter topology. The proposed ANN is trained using offline data and then tested and verified by applying it to hybrid active neutral point clamped (HANPC) topology. This topology is used to drive interior permanent magnet synchronous motor (IPMSM). The proposed method is verified by simulation results.