基于可信度的故障区段诊断方法

Approach of faulty section diagnosis based on creditability

针对非故障树干上可能存在反向馈电的问题,利用故障指示器的拒动率和误动率,提出了一种基于可信度的故障区段诊断方法。采用基于可辨识矩阵和逻辑运算的属性约简算法对原始决策表进行了约简,分析了满足约简决策表条件下的区段故障可信度特点,把确定最大可疑故障区段问题变成寻找区段故障可信度极大值和递推计算局部范围内区段故障可信度。同时,利用图形构件的拓扑关系,研究了故障模式和反向馈电概率的自动生成方法,使基于故障指示器的配电网故障寻址系统具有通用性。

As there may be power feedback in healthy branches,an approach of faulty section diagnosis based on creditability is proposed,which uses the misoperation rate and refuse-to-operate rate of fault indicators. An attribute reduction algorithm based on distinctness matrix and logical operation is used to reduce the original decision table. The creditability of faulty section satisfying the conditions of reduced decision table is analyzed and the task to determine the most suspected faulty section has two steps：to recursively calculate the fault creditability of local section and to find the maximum fault creditability. The automatic generation of fault mode and power feedback probability is studied based on the topological relation among graphic components,which makes the indicator-based fault addressing system of distribution network universal.