水电机组基于贝叶斯网络的故障树故障诊断分析研究

发布时间：2018-01-23 00:27

本文关键词： 水电机组故障诊断贝叶斯网络故障树模型转化　出处：《中国农村水利水电》2017年08期 　论文类型：期刊论文

【摘要】：为克服贝叶斯网络应用在水电机组故障诊断中存在的结构和概率不易确定的困难,研究过程中将故障树转化为贝叶斯网络模型。首先比较两种模型的类似,通过与或门故障树例子验证了贝叶斯网络方法和故障树分析法计算结果相同。再将现有故障树结构的发电机故障模型转化为贝叶斯网络模型,利用柱状图对比通过专家系统确定的先验概率和通过贝叶斯网络的反向推理获得各个底层事件在顶层事件发生故障下的后验概率,得到最有可能导致系统故障的底层事件,据此进行诊断,为复杂系统的可靠性设计和故障诊断提供依据。
[Abstract]:In order to overcome the difficulty that the structure and probability of the Bayesian network are difficult to determine in the fault diagnosis of hydroelectric generating set, the fault tree is transformed into the Bayesian network model in the research process. Firstly, the similarity of the two models is compared. The results of Bayesian network and fault tree analysis are the same, and the existing fault tree generator fault model is transformed into Bayesian network model. Using histogram to compare the priori probability determined by expert system and the backward reasoning of Bayesian network to obtain the posteriori probability of each underlying event under the top-level event fault. The underlying events which are most likely to lead to the system failure are obtained, and the diagnosis is carried out, which provides the basis for the reliability design and fault diagnosis of the complex system.
【作者单位】：深圳蓄能发电有限公司;武汉大学动力与机械学院;
【基金】：国家自然科学基金项目(51379160)
【分类号】：TV738
【正文快照】： 0引言由于组成构件之间及构件内部存在很多不确定关系,各种可能的因素充斥其间,使得故障诊断在大型设备领域应用中准确率比较低下,导致故障原因不易确定。因此如何从众多不确定信息中获得最终的最有可能的故障原因是故障诊断领域发展所需要克服的主要问题。目前利用故障树分析

【相似文献】