Enhanced Fault Classification and Localization in Microgrids Using Machine Learning

Abstract

The identification and positioning of faults are crucial in microgrids to enhance their performance and control. However, conventional protection methods are not effective due to significant variations in fault currents caused by diverse operational scenarios in microgrids. Additionally, they cannot locate the fault. Thus, the authors propose a deep learning-based system that uses discrete wavelet transform, wavelet energy entropy, and artificial neural networks to classify and locate the faults in the distribution network of the microgrid. The system is designed to quickly isolate the fault and restore power supply. MATLAB/Simulink is used to simulate the microgrid and train the neural networks. The study shows that the proposed system achieves high accuracy in fault classification and localization within a short period.