电网故障诊断的时间因果贝叶斯网模型

Temporal causal Bayesian network model for power system fault diagnosis

在使用保护和断路器的动作信息进行电网故障诊断时,这些警报的时序特性的加入能够给诊断提供更丰富的信息,使得诊断结果更准确。提出了电网故障诊断的时间因果贝叶斯网模型,采用模糊方式对时间因果关系进行离散化处理,采用模糊运算来合成多个时间因果关系,通过概率计算获得最大可能的故障假说。理论与算例表明该方法有效可行。

Taking account into the temporal information of alarm messages can afford more useful information for power system fault diagnosis,and makes the diagnosis results more accurate.A temporal causal Bayesian network model is proposed for power system fault diagnosis.The fuzzy method is used to discretize the temporal causal relations between faults and alarms,fuzzy operation is used to combine these causal relations,and the fault hypothesis with maximum likelihood is obtained by probability calculating.The theory and example demonstrate that this approach is correct and available.