Fault Detection and Classification in Overhead Transmission Lines Using Discrete Wavelet Transform and Artificial Neural Networks

Abstract

Transmission lines are fundamental components of modern power systems, as they enable the transfer of bulk electrical energy over long distances. However, their wide exposure to environmental conditions and operating stresses makes them highly vulnerable to various faults, which in turn threaten system stability, reliability, and security. To address this challenge, this paper introduces a protection scheme that utilizes only sinusoidal current signals obtained from the sending end of the line, avoiding the need for additional measurementsor complex setups. The measured signals are decomposed into positive and negative components, and their comparison forms the basis for fault detection. Once a fault is identified, classification is carried out using features derived through the discrete wavelet transform (DWT), which are subsequently analyzed by an artificial neural network (ANN) to ensure accurate categorization. Simulation studies performed in both PSCAD and MATLAB confirm the effectiveness of the proposed method, showing its robustness, adaptability, and high precision in detecting and classifying a wide range of fault conditions, thereby presenting a reliable solution for transmission line protection.