基于扩展故障树的运载火箭故障诊断专家系统

Fault Diagnosis Expert System for Launch Vehicle Based on Augmented Fault Tree

针对运载火箭故障诊断专家系统中知识获取的瓶颈问题,通过将扩展故障树分析法和基于规则的诊断专家系统有机结合,建立基于扩展故障树的运载火箭故障知识获取及表示方法,实现了从扩展故障树到诊断知识的自动转换和诊断知识的规范化表示,解决了基于规则的诊断专家系统的知识获取难题。在此基础上,结合扩展故障树给出了运载火箭故障诊断专家系统快速推理策略。该策略基于诊断优先系数,实现了对发生概率大、结构重要度高、危害严重的故障的优先诊断,并通过浅层推理与深层推理相结合,保证了推理过程的精确性和严密性。

In view of the difficulty of knowledge acquisition, a fault diagnosis expert system for launch vehicle combined with the augnlented fault tree analysis is established in the paper. On the part of diagnosis knowledge, the automatic knowledge acquisition and normative representation method for launch vehicle is presented based on the augmented fault tree analysis, which solved the problem of knowledge acquisition for expert system based on rules. Furthermore, the inference engine of the fault diagnosis expert system for launch vehicle is given also. It can priorly diagnose the faults with the high occurrence probability, structure importance and hazardness based on the diagnosis priority. In addition, the accuracy and strictness of reference process can be ensured by the combining the superficial reference and deep reference.