Robust Arc Detection in Aircraft Electrical Systems Using Adaptive Thresholding and Multi-Band Analysis

The aerospace industry’s transition towards electrical propulsion has led to the adoption of high-voltage DC electrical systems in aircraft, raising critical safety concerns, including the detection of series arcs. This paper presents a novel arc detection algorithm tailored for these demanding environments. The algorithm analyses high-frequency current signals across multiple frequency bands, capitalising on the distinct spectral signatures of arcs. A key innovation is the use of an adaptive threshold, derived from a scaled moving average of the signal power, which allows the algorithm to dynamically adapt to changes in baseline noise. A time threshold is also incorporated to further discriminate between arcs and short-duration transient events, such as switching operations. The algorithm is designed for rapid processing and high statistical accuracy. Experimental validation, using controlled arc generation and switching events under varying conditions, demonstrates the algorithm’s ability to reliably detect arcs while minimising false alarms.