A Novel Autonomous Technique for Early Fault Detection on Overhead Power Lines

Abstract

Providing uninterrupted, reliable and high-quality power supply to consumers have been very challenging for power utilities around the world. The utilities spend a considerable amount for the maintenance, inspection, testing and monitoring of power infrastructure to achieve the above. The overhead power lines are the main power carriers for the delivery of electricity to the consumers. These overhead power lines are prone to risk of fires and faults caused by vegetation encroachment and touches on the power corridor, resulting in damage to the conductors, power blackouts and catastrophic events such as wildfires or bushfires. Therefore, utilities have to regularly inspect and monitor powerlines for the prevention of damage. Utilities have been adopting foot patrol, aerial inspections and more recently LIDAR technology to inspect the overhead power lines on regular basis. However, these approaches only give point of time information and are time consuming and costly. This paper presents a novel technique which uses directional antennas mounted on poles carrying overhead conductor to detect the partial discharge (PD) signals generated by incipient faults in an experimental set up and also field testing. These PD signals are processed by FPGA signal processing unit for peak detection and filtering and digest of the signals is delivered to a secured cloud server for post-data processing such as trend analysis, pattern recognition and localisation. This paper demonstrates the effectiveness of this Early Fault Detection (EFD) technique to accurately locate vegetation encroachment and partial discharge originated from broken insulators, defective gas switches and transformers. This technology has been successfully employed at various locations on distribution feeders in Australia and the results are very encouraging.