An Adaptable Intelligent LSTM-based Technique for Fault Detection in Domestic Electrical Networks

Abstract

Fault detection is crucial in ensuring the safety and reliability of electrical systems. In this paper, we present a novel deep learning-based technique for the detection of faults in domestic electrical networks, particularly Distribution Panels using Long Short-Term Memory (LSTM) Networks. The LSTM network, among the other Recurrent Neural Network (RNN) models is chosen due to its ability to handle sequential data and long-term dependencies, making it suitable for analyzing time series data. The LSTM model is trained on data collected from sensors that monitor temperature, humidity, and the theoretical elasticity value of the wires in the distribution panel, which are critical factors in determining the health of the electrical junction boxes. The model is trained on data for a period of one year, with readings taken every four hours. The use of LSTM networks for fault detection in electrical systems is a relatively new area of research, and this study aims to explore its potential. The results of this study stands as a promising tool for future researchers and practitioners in the field of electrical engineering who are interested in developing automated fault detection systems in electrical systems.