基于RS-BN的电力信息通信系统故障诊断方法

Fault Diagnosis Method of Electric Power Information and Communication System Based on RSBN

针对电力信息通信系统集成的软硬件系统种类繁多、网络拓扑结构复杂,引起故障告警的原因具有复杂、不确定的特点,提出了一种结合粗糙集和贝叶斯网络的电力信息通信系统快速故障诊断方法。首先,对系统运行记录数据进行预处理,抽取出特征属性值,利用粗糙集对冗余的进行化简,获得最小的特征属性集,然后,根据运行记录中的词频信息计算获得各特征属性集的条件概率,在此基础上建立最小属性集的贝叶斯网络故障诊断模型,实现故障原因的快速定位分析。实验结果表明了方法的可有效、快速、准确地定位电力信息通信系统的故障原因,对保障智能电网的运行具有重要价值。

Electric power information communication system is an important supporting platform of the construction of strong smart grid. The system has various types of software and hardware, complex topology structure. So the reasons of malfunction alarm are the complex and uncertain. Aiming at this problem, this paper proposes a combination of rough sets and Bayesian network （RSBN） to fast fault diagnosis for electric power information communication system. Firstly, the system running record data is preprocessed to extract feature attribute values. Then the features are reduced with rough set to get the smallest attributes sets. The conditional probability of each feature is calculated by word frequency information from the running record. Finally, the Bayesian network of fault diagnosis is built with minimum attribute sets to realize fast fault diagnosis. The experimental results show that the method is effective, rapid, accurate positioning the cause of the problem of electric power information and communication system. This method is important to guarantee the operation of the smart grid.