卫星电源系统的有向图分层诊断方法

Hierarchical diagnosis for satellite power system based on diagraph

在卫星电源故障诊断问题的研究中,针对系统的知识获取难问题,提出了一种将有向图和CLIPS专家系统相结合的分层诊断方法。这种方法利用有向图知识描述容易和CLIPS推理引擎快速的优势,设计了一个以事实描述系统模型知识的自定义模板结构,同时采用有向图分层策略减小故障搜索空间,然后利用CLIPS提供的快速匹配算法完成诊断推理过程,最后结合已有的专家经验知识给出卫星电源的故障诊断结果。实验结果表明,新方法具有很好的准确性和快速性。

To tackle the difiqculty of knowledge acquisition for fault diagnosis of satellite power system, a hierarchical diagnosis method combining diagraph with CLIPS expert systems was presented. Taking advantage of easy knowledge description in diagraphs and high-speed inference engine in CLIPS, a deftemplate structure was defined in CLIPS to describe diagraph. Search space for failure sources was reduced by the hierarchical strategy for diagraph. By defining a set of rules in CLIPS, a hierarchical diagraph-based fault diagnosis algorithm was achieved. Rete pattern matching algorithm provided by CLIPS was used to improve the diagnosis speed. The power system of a certain satellite was illustrated as a case study to validate the effectiveness of the method proposed.