Detecting Series Arc Faults Using High-Frequency Components of Branch Voltage Coupling Signal
Zhipeng He, Zixiao Xu, Hu Zhao, Weilin Li, Yan Zhen, Wenjun Ning
Abstract
Series arc faults (SAFs) are important safety issues in low-voltage ac distribution networks. However, the load type and circuit topology are increasing, and arcing current features are easily affected by these factors, which makes SAF detection very challenging. This article proposes a new method for detecting SAFs in real scenarios containing multiple load types and complex circuit topologies. The proposed method uses the branch voltage coupling signal (BVCS) as the detection signal. First, the wavelet transform is employed to preprocess the BVCS to simplify the representation of fault components. Second, the absolute value sum of wavelet transform detail coefficients (AVSDCs) is utilized to characterize them. Then, the complete detection algorithm is designed based on the difference of AVSDC between normal conditions and arc faults and the continuity of SAFs. Finally, an online arc fault detection device (AFDD) is developed to validate the accuracy of the proposed method and its generalization ability. The results show that the proposed method has good detection accuracy under complex working conditions with multiple load types and circuit topologies.