OPTIMIZATION OF POWER DISTRIBUTION NETWORKS USING MACHINE LEARNING FOR FAULT DETECTION AND PREVENTIVE MAINTENANCE

Abstract

This study investigates the optimization of power distribution networks using machine learning (ML) techniques, with a specific focus on fault detection and preventive maintenance. The research evaluates the performance of three machine learning models—deep learning, decision trees, and support vector machines (SVM)—in terms of accuracy, fault detection capabilities, time efficiency, and cost-effectiveness. The results reveal that deep learning models outperformed traditional methods, achieving an accuracy of 92%, compared to 83% for decision trees and 85% for SVM. In terms of fault detection, deep learning detected 96% of faults, whereas decision trees and SVM detected only 85% and 89%, respectively. Moreover, machine learning models, particularly deep learning, significantly reduced fault detection time by 40%, and demonstrated a 25% reduction in operational costs due to fewer emergency repairs and optimized maintenance schedules. Machine learning offers the power distribution networks a chance to enhance operational performance and minimize operating downtime and optimize resource allocation.  The implementation success demands resolution of data quality requirements and infrastructure compatibility alongside initial installation costs and implementation expenses.  This study demonstrates the fundamental need to deploy machine learning technology in electricity distribution systems for enhancing performance while improving stability.