基于概率阈的非冗余多故障系统诊断策略优化

针对传统的单故障假设无法诊断复杂系统多故障并发的情况,提出了一种基于概率阈的非冗余系统多故障诊断策略。首先对系统的相关性模型进行扩展,并删除低于概率阈的故障状态,建立非冗余系统的多故障测试诊断模型;其次在信息熵算法的基础上建立Rollout算法,获得了最优测试序列;然后建立故障诊断树并计算测试代价;最后以某机载电子系统为例验证了该方法的有效性。该方法可以在保证测试费用最小的情况下获得非冗余系统的最优测试序列。

Optimization of diagnostic strategy in non-redundantmulti-fault system based on probability threshold

This paper propsoed a multiple fault diagnostic strategy for non-redundant system based on probability threshold. Firstly, it expanded system dependency models, removed fault states which were below probability threshold, and established multiple fault diagnostic models of non-redundant system. Then using entropy algorithm as base algorithm, it presented Rollout algorithm to obtain best test sequence. At last, it built the fault diagnosis tree and calculated test cost. It put forward an airborne electronic system to prove its effectiveness. It can obtain the best test sequence by minimums cost test.