基于FP-Growth算法的10kV配电网分支线断线故障诊断与定位方法

Open-line Fault Diagnosis and Positioning Method for 10kV Power Distribution Network Branch Line Based on FP-growth Algorithm

针对10 kV配电网分支线断线不接地故障难以被诊断的问题,利用配电网多源数据,提出通过数据间关联规则的挖掘进行断线故障诊断与定位的方法。首先根据已有数据源,建立兼具电气量特征与时空量特征的故障诊断特征库,然后通过m RMR(最小冗余最大相关)算法进行特征选择,筛选出用于断线故障诊断的主要特征,最后基于FP-Growth(频繁模式增长)算法,进行关联规则挖掘,进而确定是否发生了分支线断线不接地故障并判断故障位置。实际算例分析表明,所提方法能快速、准确地对10 kV配电网分支线断线不接地故障的发生及地点做出诊断,且具有较好的容错性。

To solve the problem of open-line non-grounding fault diagnosis for 10 k V power distribution network branch line, a diagnosis and positioning method through data association is proposed based on multiple data source. According to the data source, a feature database including electrical features and time-space features for fault diagnosis is established. The main features for fault diagnosis are selected through minimum redundancy and maximum relevance. A fault diagnosis and positioning method is proposed based on frequent pattern-growth(FP-Growth). Practical analysis shows that the method can diagnose and position open-line fault of 10 k V power distribution network branch line quickly and accurately with good fault tolerance.