Litcius/Paper detail

A Machine Learning–Based Approach for Fault Detection in Power Systems

Pathan Ilius, Mohammad AlMuhaini, Muhammad Javaid, M. A. Abido

2023Engineering Technology & Applied Science Research25 citationsDOIOpen Access PDF

Abstract

Machine learning techniques are becoming popular for monitoring the health and faults of different components in power systems, including transformers, generators, and induction motors. Normally, fault monitoring is performed based on predetermined healthy and faulty data from the corresponding system. The main objective of this study was to recognize the start of a system fault using a Support Vector Machine (SVM) approach. This technique was applied to detect power system instability before entering an unstable condition. Bus voltages, generator angles, and corresponding times before and after faults were used as training data for the SVM to detect abnormal conditions in a system. Therefore, a trained SVM would be able to determine the fault status after providing similar test data once a disturbance has been resolved.

Topics & Concepts

Support vector machineFault (geology)Electric power systemEngineeringFault detection and isolationTransformerControl engineeringInduction motorCondition monitoringArtificial intelligenceComputer sciencePower (physics)Pattern recognition (psychology)VoltageActuatorElectrical engineeringGeologyPhysicsSeismologyQuantum mechanicsPower System Optimization and StabilityPower Systems Fault DetectionPower Transformer Diagnostics and Insulation
A Machine Learning–Based Approach for Fault Detection in Power Systems | Litcius