基于AHP的因果图故障诊断研究

动态不确定因果图(DUGG)是一种最新的图形化概率推理模型,可简洁、直观地表达不确定的因果知识,提供更有效的概率推理。层次分析法可通过构建层次分析模型,计算出每一层元素的权重数值。文章将层次分析法(AHP)引入DUGG,并创新地应用于舰船柴油主机滑油系统,建立起相应的DUGG模型,应用AHP计算出所需的先验概率,再代入DUGG模型求得在观察到的证据下假设事件的排序概率,进而可对滑油系统温度过高这一故障进行诊断。

Research on fault diagnosis based on AHP and DUCG

Dynamic uncertain causality graph is a new graphical probabilistic reasoning model. It can express the uncertain causal knowledge intuitively, and provide more efficient probabilistic reasoning. Analytic hierarchy process can calculate the weight value of elements in each layer by constructing AHP model. This paper brings AHP into DUCG and the method is innovatively applied to the marine lube oil system. First establish the corresponding DUCG model and apply AHP to calculate the required prior probability, and then bring the DUCG model to obtain the sorting probability of the hypothetical event under the observed evidence. The fault of excessive temperature of lubricating oil system can be diagnosed.