Research on fault line selection technology of distribution network based on EMD and WVD

Currently, in neutral ungrounded power systems, the fifth harmonic zero-sequence power directional method for fault line selection faces practical challenges. The inherent low fifth harmonic content in such systems, combined with harmonic interference from load sources and difficulties in parameter setting, leads to relatively low accuracy in actual fault identification. Considering the rich transient fault information available during faults, this paper proposes an improved method combining Empirical Mode Decomposition (EMD) and Wigner–Ville Distribution (WVD). The approach first extracts high-frequency signal components through EMD decomposition and then calculates signal instantaneous energy using WVD. By integrating three key parameters—fault amplitude, phase angle, and instantaneous energy—the method employs phase angle information for fault mode determination. Amplitude and energy values are combined to define correctness weights. Specifically, the algorithm fuses correctness weights from 11 sampling points within 2 ms after the fault to identify the fault line. Simulation results demonstrate that this EMD–WVD multi-point fusion method for single-phase grounding fault detection features reasonable weight definition criteria, achieves high selection accuracy with rapid response, and shows strong potential for practical engineering applications.