The document summarizes research on using an artificial neural network (ANN) approach for fault detection in power transmission lines. It describes training an ANN to recognize normal system conditions from fault conditions based on changes in current and impedance signals. The ANN was trained using the backpropagation algorithm on over 11,000 data points of faults at different locations and inception angles on a simulated 100km transmission line. The trained ANN was able to detect faults with a final error rate of 0.1%, demonstrating the potential for ANNs to enable fast and accurate fault detection compared to conventional relaying techniques.