基于零序电流的配网单相接地故障综合辨识方法

A Comprehensive Fault Identification Method for Single-Phase Ground Faults in Distribution Networks Based on Zero-Sequence Current

中文关键词:配电网  零序电流  故障辨识  波形相似度  残差网络

keywords:distribution network  zero-sequence current  fault identification  waveform similarity  residual network

中文摘要:针对配电网高阻故障下故障定位和故障原因辨识困难的问题，提出了一种基于零序电流的配电网综合故障辨识方法。首先，分析了配电网高阻故障下节点处基频零序电流波形相似度与配电网典型故障的零序电流波形特征。其次，提取节点处基频零序电流故障后首个周波波形，利用其波形余弦相似度特征实现故障区段定位。再次，利用残差网络深度学习模型提取故障区段零序电流波形的图像特征，辨识故障原因。最后，在仿真平台搭建多分支混合线路配电网模型，构建故障仿真样本库用于训练和测试。算例分析表明，在多分支混合线路配电网中，所提方法对于故障区段定位准确率为100%，对于故障原因辨识的准确率可达99.8%，在噪声和励磁涌流等干扰的影响下，仍具有良好的抗扰性和适应性，有助于指导故障重合闸，提高配电网故障恢复效率，提高供电可靠性。

Abstract:To solve the difficulties of fault location and fault cause identification under high-resistance faults in distribution networks，a comprehensive fault identification method for distribution networks based on zero-sequence current is proposed. First，the paper analyzes the waveform similarity of fundamental-frequency zero-sequence current at nodes and the zero-sequence current waveform characteristics of typical faults in distribution networks under high-resistance faults. Second，the paper extracts the waveform of the first cycle of fundamental-frequency zero-sequence current at nodes after a fault，and uses the cosine similarity feature of this waveform to achieve fault section location. Third，it employs a residual network （a deep learning model） to extract image features from the zero-sequence current waveform of the fault section，thereby identifying the fault cause. Finally，a multi-branch hybrid line distribution network model is built on a simulation platform，and a fault simulation sample library is constructed for training and testing. Case analysis shows that in the multi-branch hybrid line distribution network，the proposed method achieves a 100% accuracy rate for fault section location and a 99.8% accuracy rate for fault cause identification. Good anti-interference performance and adaptability are remained even under the influence of disturbances such as noise and inrush current，which helps guide fault reclosing，improve the efficiency of distribution network fault recovery，and enhance power supply reliability.