基于AHP的贝叶斯网络故障诊断方法研究

针对基于专家知识的故障诊断方法依赖经验的局限，提出一种基于层次分析法（AHP）的贝叶斯网络化工过程故障诊断方法。通过基于关联函数的AHP得到所有变量的权值，对22个变量节点的权值进行排序并将该排序作为K2算法的学习输入建立贝叶斯网络模型，同时结合复杂网络分析指标进行化工过程的故障诊断。通过TE过程故障诊断实例证明本文方法不仅避免了K2算法专家知识的主观因素影响，同时能很好地进行故障定位，找到故障源。

A fault diagnosis method for a Bayesian network based on AHP

A chemical process fault diagnosis method based on the analytic hierarchy process (AHP) is proposed in order to overcome the limitations of experience knowledge based on expert knowledge.The weight of all the variables is obtained by AHP based on the correlation function.The weight of the 22 variable nodes is sorted and the order is used as the learning input of the K2 algorithm to establish the Bayesian network model.At the same time,the chemical process is combined with the complex network analysis index Troubleshooting.The fault diagnosis example of the TE process shows that this method not only avoids the influence of subjective factors in K2 algorithm expert knowledge,but also can locate fault location accurately and find the fault source.