Gaussian Naïve Bayes Algorithm Based Transmission Line Fault Classification by Using Single-Ended Parameters

The transmission line is the most important component of the power system. Classifying the faults occurring on the transmission line helps the system operator activate the mechanism of unsymmetrical circuit breaker tripping. In this work, different types of faults occurring on the transmission line are classified with the help of single-ended parameters only. Firstly, the data required to generate features is created by simulating the practical transmission line in PSCAD software. The data consists of all three phase voltages and currents on one side of the transmission line when the line is subjected to all possible faults including no fault condition. While simulating the faults on the transmission line, parameters such as fault resistance, fault inception angle, and fault location are varied in all types of faults. The generated data is uploaded to Python software to calculate different features such as time series parameters and frequency series parameters. This feature set is applied to different machine learning algorithms to classify the faults on the transmission line. The performance of the different algorithms is compared to select the best algorithm.