| 引用本文: | 刘畅宇,王小君,张大海,等.基于时间卷积网络的配电网高阻接地故障检测及可解释性分析方法[J].电力系统保护与控制,2026,54(03):109-120.[点击复制] |
| LIU Changyu,WANG Xiaojun,ZHANG Dahai,et al.High-impedance ground fault detection and interpretability analysis in distribution networks based on temporal convolutional networks[J].Power System Protection and Control,2026,54(03):109-120[点击复制] |
|
| 摘要: |
| 数据驱动型算法可有效降低配电网多重随机性及噪声干扰对高阻故障检测阈值的影响，但由于模型“黑箱”特性致使其可解释性不足。为此，提出一种基于时间卷积网络(temporal convolutional networks, TCN)的配电网高阻接地故障检测及可解释性分析方法。首先，利用改进自适应噪声完备集合经验模态分解对零序电流进行分解与重构，过滤噪声干扰的同时增强故障特征表达。其次，构建TCN对处理后的波形进行时序特征提取，提升模型对高阻故障及典型扰动工况的识别能力。然后，构建分数加权的类激活映射方案对模型的检测依据展开分析，结合波形关键区域的归因指标刻画高阻“零休”特性与模型决策关注区域的匹配度，提升模型可解释性。最后，在MATLAB/Simulink仿真模型及真型试验场数据的基础上，进一步验证了所提方案的有效性和可靠性。 |
| 关键词: 配电网 高阻接地故障 改进自适应噪声完备集合经验模态分解 时间卷积网络 可解释性 |
| DOI：10.19783/j.cnki.pspc.250672 |
| 投稿时间：2025-06-23修订日期：2025-12-12 |
| 基金项目:国家自然科学基金面上项目资助(52377071)
较近，易接触地表植被、混凝土、树枝等非金属导电介质引发高阻故障(high impedance fault,?HIF)。HIF过渡电阻较大导致其故障特征微弱，且与电容投切(capacitor switching,?CS)、负荷投切(load switching, |
|
| High-impedance ground fault detection and interpretability analysis in distribution networks based on temporal convolutional networks |
| LIU Changyu1,WANG Xiaojun1,ZHANG Dahai1,SHANG Boyang1,ZHANG Yongjie1,ZHANG Yongjie2 |
| (1. School of Electrical Engineering, Beijing Jiaotong University, Beijing 100044, China; 2. Economic Research
Institute of State Grid Sichuan Electric Power Company, Chengdu 610095, China) |
| Abstract: |
| Data-driven algorithms can effectively reduce the influence of multiple uncertainties and noise interference on detection thresholds for high-impedance faults in distribution networks. However, the “black-box” nature of these models limits their application interpretability. Thus, a method for high-impedance ground fault detection and interpretability analysis in distribution networks based on temporal convolutional networks (TCNs) is proposed. First, an improved adaptive noise-complete ensemble empirical mode decomposition is employed to decompose and reconstruct the zero-sequence current, suppressing noise interference while enhancing fault feature expression. Then, a TCN is developed to extract temporal features from the processed waveforms, thereby improving the model’s ability to distinguish high-impedance faults from typical disturbance conditions. Subsequently, a fractional-weighted class activation mapping scheme is designed to analyze the model’s decision basis. By combining attribution indicators of key waveform regions, the method characterizes the correspondence between the distinctive “zero-off” features of high-impedance faults and the model’s decision-focused regions, thereby enhancing interpretability. Finally, based on MATLAB/Simulink simulation models and field test data, the effectiveness and reliability of the proposed method are validated. |
| Key words: distribution network high-impedance ground fault improved adaptive noise-complete ensemble empirical mode decomposition temporal convolutional network interpretability |
