动态贝叶斯网络在电力通信网故障诊断中的应用

Application of Dynamic Bayesian Network in Fault Diagnosis of PowerCommunication Network

为解决当前电力通信网网络管理中过分依赖人工,故障诊断的时效性和精确度难以保证,存在影响网络的安全运行,难以适应数字化建设等问题,本文提出将动态贝叶斯网络应用于电力通信网的故障诊断中。首先,考虑噪声对识别的干扰,对噪声进行处理,建立网络疑似故障节点与探测集合间的动态贝叶斯模型;然后,定义疑似故障节点阈值,过滤噪声探测,同时缩小网络规模保证识别的时效性;最后,根据网络规模和单节点故障概率限制同时发生故障的节点个数进一步缩小计算规模,利用贪心思想计算得到最有可能故障节点集。实验结果表明,在网络的不同噪声水平中,该算法均能在满足网络故障诊断时效性要求的同时,具有更高的准确率,可以为电力通信网故障诊断提供依据。

In order to solve the problems of over-reliance on manual labor,the timeliness and accuracy of fault diagnosis are difficult to guarantee,the safe operation of the network is affected,and it is difficult to adapt to digital construction in the current power communication network management,in this paper,the application of dynamic Bayesian network in fault diagnosis of power communication network is proposed.Firstly,considering the interference of noise on identification,the noise is processed,and the dynamic Bayesian model between the network suspected fault nodes and the detection set is established.Then,the threshold of the suspected fault node is defined,and the noise detection is filtered,and the network size is reduced to ensure the timeliness of identification.Finally,according to the network scale and the probability of single node failure,the number of nodes that fail at the same time is further reduced to further reduce the calculation scale,and the greedy idea is used to calculate the set of most likely faulty nodes.Simulation results show that in different noise levels,the algorithm can meet the timeliness requirements of network fault diagnosis,while having a higher accuracy rate,which can provide basis for fault diagnosis of power communication network.