Deep Learning Based Identification and Categorization of Transmission Line Fault

Abstract

Fast and accurate fault identification is instrumental in the reliable functioning of power trans mission systems. When done promptly, it can significantly limit the damages to equipment as well as power interruptions. Typically, Identification and categorization of faults Techniques are founded on mathematical models and threshold operations. However, these methods experience difficulties when dealing with varying system conditions, noise, and non-linear behavior. This study introduces a deep learning based solution to the problem of the identification and classification of transmission line faults. The approach involves gathering voltage and current signals from transmission lines under both ideal and flawed circumstances. The signals are then reprocessed and fed into They are trained using deep learning models like Long Short Term Memory (LSTM) networks and Convolutional Neural Networks (CNN). These models are capable of automatically extracting the most relevant features from the time, domain and frequency, domain representations of the signals, thereby obviating the need for manual feature engineering. The developed network can not only locate a fault but also identify its category, For instance, single line, to, ground, double line, to, ground, and three-phase faults. Simulation studies conducted under varying fault locations, fault resistances, and noise levels demonstrate that the proposed deep learning framework achieves high accuracy, robustness, and faster response compared to conventional methods. The results confirm that deep learning techniques provide an effective and reliable solution for real-time fault identification and categorization in modern power transmission systems.