基于深层知识规则的液体火箭发动机故障诊断

本文从深层知识的表示方法之一——符号有向图出发，提出一种基于规则的故障诊断方法。该方法首先建立液体火箭发动机系统的符号有向图，然后根据各个故障根源得出系统的子符号有向图，最后根据这些子图导出故障诊断规则，用这些规则去诊断系统故障。该方法具有较好的解释性和诊断的快速性。

ROCKET ENGINE FAULT DIAGNOSIS BASED ON RULES OF DEEP KNOWLEDGE REPRESENTATION

From one method of the deep knowledge representation,the signed directed graph(SDG),a qualitative method of fault diagnosis based on rules is proposed in this paper In the method,SDG of the rocket engine system is drawn first,and then the sub SDGs corresponding to the fault origins are derived from the SDG of the engine system,in the end rules are derived from the sub SDGs,the rules can be used to diagnosis the faults of the engine system The method has a better interpretation and can identify the faults quickly