 The proposed method combines wavelet transform and machine learning techniques to detect, classify, and locate electrical faults on transmission lines. It first uses wavelet transform to extract features from current or voltage signals, which are then used to calculate coefficients and convert them to an RGB image. This image is then fed into a GoogleNet model to classify the fault type, and a convolutional neural network is used to quickly locate the fault. Simulation results demonstrate that this method has high accuracy and fast processing times, making it a valuable tool for analyzing system stability in the electricity industry. This article was authored by Nguyen Nhan Banh and Le Van Dye.