摘要
基于定性模型的诊断方法,由于无需系统定量模型,尤其适用于航天器等复杂系统的故障诊断.该方法推理过程包括冲突集的确定及诊断集的生成.基于BHS-树的方法,是一种由冲突集生成诊断集的有效途径,但存在诊断结果冗余及反向递归过程复杂的问题.采取异常/正常特征量相结合的措施,有效地解决了诊断结果冗余的问题,能够准确确定诊断集;同时在系统单故障的假设前提下,对反向递归过程进行适当简化,降低了推理的复杂程度,并通过实例对上述改进进行了验证.
Because the qualitative model-based diagnosis approach has no need of quantitative system model,it is of particular for the fault diagnosis in many complicated fields such as spacecraft.Its diagnosis process can be divided into two parts： determination of the set of conflicts and generation of the set of diagnoses.A fault diagnosis approach based on BHS-tree can deal with the latter.However,this approach has two drawbacks： result redundancy and reversed recursion complexity.An improved approach is proposed to deal with two problems in this paper.The information of abnormal and normal observations is related to solve the former,and the process of reversed recursion is simplified on the assumption of single fault.The improvement is validated by an example.
出处
《空间控制技术与应用》
2011年第2期21-25,共5页
Aerospace Control and Application
关键词
故障诊断
定性模型
BHS-树
结果冗余
反向递归
fault diagnosis
qualitative model
BHS-tree
result redundancy
reversed recursion
