Abstract: The failure mechanism of soft fault is vague, and fault timely diagnosis is difficult as well as simulation. The soft fault diagnosis method of transmission line based on network model is proposed. The modularity concept of cluster structures in a complex network is introduced to resolve the problem on the similarity measure between the state division and partition criteria. The failure status and characteristics are reflected by network structure, then the fault diagnosis network model is established, and fault diagnosis of the transmission line is transformed into sub-networks to detect problems. For various fault signatures of the same fault, multi-sensor can collect different characteristic value in multi-level and multi-field. High sensitivity state information can be selected to analyze the fault diagnosis better. The experimental results show that the proposed algorithm can achieve higher diagnosis accuracy and has better fault tolerance.

Key words: network model, soft fault, transmission line, fault diagnosis

摘要： 针对输电线路缓变故障的故障机理不明确，全系统的故障及时诊断和仿真困难等问题，提出了基于网络模型的缓变故障诊断方法。引入复杂网络集群划分中的模块性概念，解决了状态划分过程中的数据间相似度测度和划分测度问题。利用网络结构反映故障状态和特征，建立故障诊断网络模型，把输电线路故障诊断转化为子网络探测问题。通过多传感技术针对同一故障的多种故障表征，多层次多领域采集不同的特征量，选择故障反映灵敏度高的状态信息量，从而较好分析诊断故障。通过实验仿真对比，证明了基于网络模型的故障诊断方法对输电线路缓变故障诊断具有较高的正确率和容错性。

关键词: 网络模型, 缓变故障, 输电线路, 故障诊断