A Dual-Modal Pantograph–Catenary Arc Detection Method Based on Sound-Optic Fusion

The pantograph–catenary (P–C) system is the primary power supply channel for high-speed trains. When the pantograph goes offline, it can cause P–C arc phenomena, which may lead to damage to the P–C system. Detection of these arcs is essential for the safe and reliable operation of high-speed trains. However, current P–C arc detection methods still suffer from a high false negative rate (FNR) and a high false positive rate (FPR). To address these problems, this article proposes a dual-modal P–C arc detection method based on sound-optic fusion. First, the sound signal and optical signal in the P–C area are analyzed respectively to obtain the duration and relative intensity of the arc. Second, the time window analysis method is used to judge whether the arc occurs according to the typical characteristics of the arc. Finally, aiming at the uncertainty of the time window analysis method, a fusion method of fuzzy logic and DS evidence theory is proposed to blur the arc characteristics in the sound-optic signal, and then the confidence is calculated to determine whether the arc occurs. The experimental data indicate that the FPR and FNR of the method are reduced to 0.79% and 0.6%, and the accuracy is improved to 99.21%. The method has good accuracy and robustness.