A Novel Approach to Fault Diagnosis of High Voltage Transmission line - A Self Attentive Convolutional Neural Network Model

A novel self attention convolutional neural network (SAT-CNN) model is presented in this paper for fault detection and classification (FDC) of high voltage transmission line that continuously experiences the number of shunt faults. The effect of fault rises the instability, line restoration cost and damages the load in real time application. Therefore, for the rapid restoration of faulty phases, a precise model is required to detect and classify the faults. In this paper, we propose a SAT-CNN with time series imaging based feature extraction model which is capable to perform precise detection and classification of faults. The effectiveness of the proposed SAT-CNN model is tested by using the number of input signals namely voltage, current and combined voltage and current signal, under various sampling frequencies. A comparison between proposed and other AI based method is also conducted in this paper to confirm the high performance of SAT-CNN as compared to others.