Scholarlink

<p>To enhance the reliability of ground fault detection in distribution networks and their ability to adapt to complex operating environments, this paper proposes an optimized method for automatic detection of ground wire faults that integrates transient signal analysis and controllable signal injection. Firstly, an improved adaptive noise complete integrated empirical mode decomposition combined with time–frequency peak filtering (ICEEMDAN-TFPF) is used to denoise the transient zero sequence current signal, effectively suppressing noise interference; On this basis, an improved dynamic time bending distance algorithm is used to construct a comprehensive cross-correlation coefficient matrix for each line signal waveform, and fault line selection is achieved through K-center point clustering, significantly improving the accuracy of line selection. After determining the faulty line, inject controllable signals, extract harmonic sequence components based on the symmetrical component method, establish harmonic balance equations and solve them to achieve accurate positioning of the fault distance. The experimental results show that the proposed method can effectively restore transient signal characteristics and accurately select fault lines in strong noise environments; The maximum error of fault location is only 0.57&#xa0;km, which is better than the allowable error standard of 1&#xa0;km, and maintains high positioning accuracy under different grounding resistance and harmonic combinations, verifying the robustness and practicality of this method in complex distribution network environments.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Optimization of Automatic Detection of Ground Faults in Distribution Networks Based on Transient Signals and Injection Method

  • 期刊论文

  • 发表日期：

    2025年11月11日

查看全文
  • Tao Du

摘要

To enhance the reliability of ground fault detection in distribution networks and their ability to adapt to complex operating environments, this paper proposes an optimized method for automatic detection of ground wire faults that integrates transient signal analysis and controllable signal injection. Firstly, an improved adaptive noise complete integrated empirical mode decomposition combined with time–frequency peak filtering (ICEEMDAN-TFPF) is used to denoise the transient zero sequence current signal, effectively suppressing noise interference; On this basis, an improved dynamic time bending distance algorithm is used to construct a comprehensive cross-correlation coefficient matrix for each line signal waveform, and fault line selection is achieved through K-center point clustering, significantly improving the accuracy of line selection. After determining the faulty line, inject controllable signals, extract harmonic sequence components based on the symmetrical component method, establish harmonic balance equations and solve them to achieve accurate positioning of the fault distance. The experimental results show that the proposed method can effectively restore transient signal characteristics and accurately select fault lines in strong noise environments; The maximum error of fault location is only 0.57 km, which is better than the allowable error standard of 1 km, and maintains high positioning accuracy under different grounding resistance and harmonic combinations, verifying the robustness and practicality of this method in complex distribution network environments.