Machine Learning Based Peripheral Bus Fault Discrimination using Sequence Components in Radial Distribution Network

A novel topology for choosing best method to detect the type of fault occurred at any periphery of a radial power network through machine learning has been presented. Firstly, conventional load flow analysis has been performed, and then, different types of sequence parameters have been found for healthy and different fault conditions at peripheral load buses. Then, sequence component-based fault diagnosis has been performed at the end, and six types of machine learning-based methods have been applied for the discrimination of all types of faults. From a comparative study, the best suitable method has been derived.