论文部分内容阅读
由于机床故障树的复杂性,通过对故障树进行模块化分解得到具有相似模块结构的子模块,并建立模块库进行存储。介绍了基于模块库的对故障特征进行检索的基本流程和原则,同时为了避免重复检索以及NP问题的产生,引入了有效模块的概念,即在遍历模块库时,不仅要考虑事件在故障树中上下层级关系,也要考虑各事件之间的从属关系。最后研究了一种基于贝叶斯理论的重要度最优排序方法并进行重要度校正。核心是引入了有效模块的概念和对最小割集进行基于贝叶斯理论的重要度最优排序。结果表明:该方法能有效缩短求解规模,诊断速度快,重要度排序可信度高。
Due to the complexity of the machine fault tree, the sub-modules with similar module structure are obtained through the modular decomposition of the fault tree, and the module library is established for storage. In this paper, the basic flow and principles of searching for fault features based on module library are introduced. In order to avoid repeated retrieval and NP problem, the concept of effective modules is introduced. When traversing module library, not only the event is considered in the fault tree The relationship between the upper and lower levels, but also consider the affiliation between the various events. Finally, a Bayesian theory based on the importance of optimal sorting method and the importance of correction. The core is to introduce the concept of effective modules and the minimum cut set based on Bayesian theory of the importance of the optimal order. The results show that this method can effectively shorten the solution scale, speed up diagnosis and high reliability of importance ranking.