Electromagnetic radiation detection and monitoring in high-voltage transmission lines using machine learning techniques

Abstract

Electromagnetic radiation (EMR) from high-voltage transmission lines (HVTL) poses significant risks to both human health and electrical infrastructure. Accurate detection and monitoring of EMR are essential for assessing its impact, severity, and potential mitigation strategies. This study investigates EMR data collected from transmission lines operating at 400 kV, 230 kV, 110 kV, 22 kV, and 11 kV at multiple locations, leveraging Machine Learning (ML) techniques based on Artificial Intelligence (AI) for classification and regression analysis. The dataset comprises electric and magnetic field measurements as input features, while transmission line voltage, EMR impact, and severity serve as target variables. To achieve precise classification and prediction, multiple ML models, including Random Forest (RF), Decision Trees (DT), Support Vector Machines (SVM), k-Nearest Neighbors (k-NN), Ensemble methods, and Artificial Neural Networks (ANN), were employed. A comparative performance analysis demonstrated that the Ensemble Bagged Trees algorithm outperformed other models in terms of accuracy, sensitivity, specificity, false positive rate (FPR), and F1 score. The model achieved an impressive accuracy of 90.1 % in classifying transmission line voltage levels and 99.4 % in predicting EMR severity, making it a highly effective tool for real-time monitoring. By integrating ML-based classification and prediction frameworks, this research provides a robust and scalable approach to real-time EMR assessment, enhancing power grid reliability and electromagnetic safety. The findings contribute to improved safety protocols for power line workers, UAV operations, and proactive fault detection in power systems.