基于模糊匹配的配电网短路故障区段定位方法

Short-circuit Fault Location Method of Distribution Network Based on Fuzzy Matching

摘要:

针对传统10 kV配电网短路故障方法难以适应非健全信息环境的难题，文中利用配电网配用电信息系统的多源数据，以Elman神经网络模型为中心，建立基于模糊匹配的短路故障区段定位方法。首先，以配用电信息系统数据库为基础，对配电网短路故障相关的信号与电气量进行分析，建立配电网短路故障诊断特征库。然后通过I-Relief算法进行主要特征的筛选选取，来作为Elman神经网络的数据输入，并基于Elman神经网络模型对多源数据和配电网短路故障类型及位置进行模糊匹配。最后通过西南某地区实际算例分析，证明所提模型能高效快速地对10 kV配电网短路故障进行区段定位，且具有较好的容错性和实用性。

Abstract:

Aiming at the difficulty that the shortcircuit fault method for traditional 10 kV distribution network is difficult to adapt to the unsound information environment, the multisource data of electricity information system in distribution network is used to establish a short-circuit fault zone location method based on fuzzy matching with the Elman neural network model as the center. Firstly, based on the distribution information system database, the signals and electrical quantities related to short-circuit faults in distribution network are analyzed to establish a shortcircuit fault diagnosis feature library for distribution network. Then, the main features are selected by I-Relief algorithm as the data input of Elman neural network, and the fuzzy matching for shortcircuit fault type and location of multi-source data and distribution network is performed based on Elman neural network model. Finally, through the analysis of an actual case in a southwest region, the proposed model can efficiently and quickly locate the shortcircuit faults in 10 kV distribution network with good fault tolerance and practicality.