AC Arc Detection Method based on Periodicity of High-Frequency Current Components

This paper presents an algorithm for detecting arcs by analyzing high-frequency signals of current. Series AC arcs were induced through experiments, capturing the current waveform during arc occurrences. In an AC system, arcs function akin to a switch, intermittently interrupting the circuit. The ignition and extinguishing phases of the arc generate periodic high-frequency component in the current, resulting from abrupt changes in its flow. Leveraging these distinctive features, this paper suggests an arc detection algorithm by scrutinizing the duration and occurrence intervals of high-frequency components. To enhance detection accuracy, loads were categorized as resistive or non-resistive, and specific algorithms were applied accordingly. A notch filter was employed to analyze specific frequency ranges in the time domain. To validate the algorithm, detection times were measured using arc current data for each type of load. © 2024 Korean Institute of Electrical Engineers. All rights reserved.