Abstract
his research project aims to investigate a cost-effective and efficient fault monitoring that detect, diagnose and classify events arising from the neutral section assembly of a 25kV single-phase electrified railway line network. Condition monitoring and fault detection on the railways is vital, particularly on detecting faults from the state of origin to provide an accurate root cause analysis, whilst predicting inevitable failures. A concept encompassing wireless sensors and machine learning techniques is developed to detect, diagnose and predict failures resulting from a poor contact wire and pantograph interaction through cloud processing. The literature explores the advantage of employing wireless sensor networks against hard-wiring over the long stretch of the overhead wires. The clustering, prediction of tasks using machine learning techniques from the enormous volume of sensed data onsite is simplified. The fault detection and diagnosis of this experiment is conducted simultaneously through cloud processing, employing ThingSpeak™...
M.Phil. (Electrical Engineering)