Early Detection of Vegetation Ignition Due to Powerline Faults

Abstract

High impedance faults through contact with vegetation are one of the main causes of electrically caused wildfires. While detecting these faults is challenging on its own, it is important to do so in the context of the risk of vegetation ignition, as disconnecting the power infrastructure can have unwarranted, damaging consequences during an emergency. Hence, we propose a methodology for prevention of wildfires, through the accurate prediction and early detection of the ignition risk resulting from high impedance faults. Our methodology uses a set of features derived from time- and frequency-domain analyses. To test our methodology, we use a large, publicly available experimental dataset. Our results demonstrate that the methodology allows the detection of ignition risk well before its onset with high accuracy.