摘要
基于柴油机动力装置故障树的诊断模式,引入了目前处理概率知识基础上不确定性问题的最有力的推断方法之——贝叶斯网络,但是如何建立贝叶斯网络是其广泛应用的“瓶颈”。介绍了贝叶斯网络的基本理论并提出了基于故障树的贝叶斯网络建造方法,建立了柴油机故障树诊断模型,并把它转化成贝叶斯网络,利用贝叶斯网络对柴油机系统进行了故障诊断,从而改善了诊断结果,并按照最优原则充分利用观测信息,进行了网络的更新和知识积累。
Based on the model of Fault Tree for the Diesel Engine, the Bayesian Network, which is one of the most powerful methods in solving the uncertainty problems, is introduced. But how to establish the BN is the "bottleneck" of the application in the fault diagnosis. This paper discusses about the Bayesian Network and the method of establishing the BN based on FT. The diagnosis model of Bayesian Network is established based on FT, and the diesel engine system is diagnosed by BN. The diagnosis results is improved, and according to the superior principle, the network is updated and the knowledge is piled with the observation.