诊断键合图在液体火箭发动机故障诊断中的应用

Application of Diagnostic Bond Graph to Fault Diagnosis for Liquid-Propellant Rocket Engine

研究了通过诊断键合图模型建立液体火箭发动机解析冗余关系的故障诊断方法。通过在液体火箭发动机键合图模型中加入观测信号和虚拟传感器的方法,建立其诊断键合图模型,并利用此模型构建了系统的解析冗余关系。提出了根据诊断键合图因果路径并利用解析冗余关系残差趋势分析建立具有残差变化方向故障特征矩阵的方法,使故障特征值能更好地反映故障特征。最后利用仿真数据对诊断结果进行验证和分析,表明该方法建模简单、诊断速度快、准确性高。

The liquid-propellant rocket engine＇s fault isolation method using analytical redundancy relations （ARRs） generating from diagnostic bond graph （DBG） has been studied. The DBG model of LRE has been constructed by introducing observed signal and virtual sensors into bond graph model of the system. Then the ARRs for LRE can be established using the DBG model. The method to build fault signature matrix by incorporating with the causality of DBG and the tendency analysis with the residuals of ARRs has been proposed. The fault eigenvalue obtained from this fault signature matrix can represent fault signature in a better and more precise way. By using the simulation data, the fault diagnosis results have been analyzed and verified. It is shown that the fault diagnosis method is simple, fast and with high veracity.