Fault Identification Method of Distribution Networks Considering Multiple Disturbance Factors and Travelling Wave Transmission Characteristics

In the context of increasingly complex distribution networks where accurate fault identification is vital for power supply reliability, conventional denoising methods face significant challenges, including information loss under multi‐disturbance conditions and inadequate characterisation of weak faults due to insufficient feature saliency. To address these issues, this study proposes a novel fault identification methodology that comprehensively considers multi‐disturbance factors and leverages travelling wave (TW) propagation characteristics. The proposed method uses adaptive local iterative filtering integrated—singular spectrum analysis (ALIF‐SSA) for signal denoising, preserving essential fault information while suppressing noise, and extracts spectral features from reconstructed signals via frequency‐domain transformation, focusing on harmonic distributions and dominant frequency components. A dual‐band evaluation strategy (10–100 kHz and 1–5 MHz) is employed to enhance feature separability in interference‐intensive environments, prioritising low‐frequency components (10–100 kHz) for detection due to their stable transmission properties and analysing high‐frequency components (1–5 MHz) through normalised amplitude ratio comparisons. This framework combines the stability of low‐frequency signals with the discriminative resolution of high‐frequency components for complementary diagnostics. Comparative case studies validate that the proposed approach outperforms conventional single‐criterion methods in identification accuracy, offering a more reliable solution for fault identification in distribution networks.