• 论文 •    

基于故障行为的模型诊断方法

张立明, 欧阳丹彤,, 赵相福,   

  1. 1.吉林大学 计算机科学与技术学院,吉林长春130012;2.吉林大学 符号计算与知识工程教育部重点实验室,吉林长春130012
  • 出版日期:2008-09-15 发布日期:2008-09-25

Model diagnosis based on failure behavior

ZHANG Li-ming, OUYANG Dan-tong, ZHAO Xiang-fu   

  1. 1.School of Computer Science & Technology, Jilin University, Changchun 130012, China; 2.Ministry of Education Key Lab of Symbolic Computation and Knowledge Engineering, !Jilin University, Changchun 130012, China
  • Online:2008-09-15 Published:2008-09-25

摘要: 提出一种利用故障行为求解所有极小诊断的新方法,并结合带有终止节点的集合枚举树形式化地表达计算过程,逐步生成所有的极小诊断。该方法不用求解冲突集及冲突集的碰集,一次直接求出所有的极小诊断,避免了由于调用基于假设的真值维护系统时的组合爆炸而引起的NP-完全问题。在集合枚举树中添加了终止节点,从而避免了非极小诊断的产生,且不会因剪枝而丢失正确的解。实验结果表明,该算法程序容易编制,且效率较好,可以满足复杂的被诊断对象的实时性要求。

关键词: 故障分析, 故障行为, 基于模型的诊断, 极小诊断, 集合枚举树

Abstract: A new method using failure behavior to resolve all minimal diagnoses, and the computation procedure was formalized by combining revised Set Enumeration tree (SE-tree) with closed nodes to produce all the resolutions gradually. It could directly compute all the minimal diagnoses without computing all the conflict sets and therefore the hitting sets of the collection of the corresponding conflict sets like the classical methods. And then the combinatorial explosion caused by calling Assumption-based Truth Maintenance System (ATMS), known as an NP-complete problem, could be avoided as well. As the closed nodes were added to the SE-tree, the non-minimal diagnoses could never be produced, and the true resolutions could not be missed by pruning, either. Results showed that the corresponding algorithm could be easily implemented, the efficiency was greatly improved and real-time requirements for complicated diagnosed objects were satisfied.

Key words: failure analysis, failure behavior, model-based diagnosis, minimal diagnosis, set enumeration tree

中图分类号: