论文部分内容阅读
针对传统的传递系统模型在航天器自主诊断系统中只能表示单一粒度诊断知识,知识表示的完备性差,导致诊断结果的分辨率不高,且诊断系统执行效率低的问题,提出了分层传递系统模型和诊断方法。为了保证不同粒度诊断知识的完备性表示,将系统模型按不同粒度和结构关系进行分层。通过自顶向下的递归搜索,先监测系统异常,再匹配故障类型,逐层缩小搜索空间,最终找出故障候选集和故障的层次关系。同时为了减小计算规模,在诊断过程中引入了分离策略,降低实时诊断计算量。应用该方法建立了某卫星测控分系统模型并进行故障仿真,结果表明该方法可以增加模型知识表示的完备性,并能有效提高诊断效率和结果的分辨率。
Aiming at the problem that the traditional transfer system model can only represent the single granularity diagnostic knowledge and the completeness of the knowledge representation in the spacecraft autonomous diagnosis system, the resolution of the diagnostic result is not high and the execution efficiency of the diagnostic system is low, a hierarchical transfer System model and diagnosis method. In order to ensure the completeness of the diagnosis knowledge of different granularity, the system model is stratified according to different granularity and structure. Through top-down recursive search, the system is firstly monitored for anomalies, matched with the fault type, and the search space is narrowed down one by one to finally find out the hierarchical relationship between the fault candidate set and the fault. At the same time, in order to reduce the scale of computation, a separation strategy is introduced in the diagnosis to reduce the computational complexity of real-time diagnosis. This method is used to establish a satellite observing and controlling subsystem model and to simulate the fault. The results show that this method can increase the completeness of the model knowledge representation, and can effectively improve the diagnostic efficiency and the resolution of the result.