Application of Artificial Neural Network for High Impedance Fault Detection in Power System

Abstract:
The High impedance fault (HIF) occurs when an energized conductor comes in contact with surfaces having relatively high resistance, such as concrete, tree limbs, etc, and results in low magnitude fault currents than conventional relaying systems’ threshold. As a result, HIF remains undetected by the conventional relaying system. This paper presents an HIF detection method using artificial neural networks (ANN). The power system and fault scenarios are modelled using MATLAB software R2024b. RMS voltages and currents of three phases are used to train the ANN. A trained ANN-based detection model is tested under various fault conditions. A microcontroller-based hardware prototype is implemented to validate the performance and accuracy of the implemented HIF detection logic. The ANN-based detection algorithm and prototype accurately detect high impedance faults.