Arc faults present a significant risk to electrical power systems, potentially causing equipment damage, fires, and service interruptions if not detected quickly. Traditional protection methods often struggle to detect arc faults because of their nonlinear, low-current, and nonstationary behaviour. Wavelet-based analysis has proven effective due to its strong time–frequency resolution. By decomposing voltage and current signals into multiple frequency bands, wavelet transforms extract transient features linked to arc initiation and extinction. Indicators such as wavelet coefficients, high-frequency energy, and entropy help distinguish arc faults from normal conditions and disturbances. Discrete, continuous, and wavelet packet transforms, combined with intelligent classifiers, enhance detection accuracy, speed, and robustness in modern distribution systems.
Kumar et al. (Thu,) studied this question.