High?speed vehicle adopt different control strategies for different fault events in the event of external en?vironment disturbance and servo failure event leading to state instability. In order to enable the vehicle to identify the type of event that causes the abnormal state autonomously, and accordingly call the corresponding processing strategy, this paper proposes a decision system based on decision network for vehicle anomaly diagnosis. Based on the Petri net method, the abnormal event diagnosis system of the vehicle is constructed. By analyzing the influence of the different event on the flight state, the system can make the vehicle autonomously locate the event types that cause the abnormal state quickly by using the matrix reasoning ability of Petri net, with sensor reading of each state quantity as input and event type as output, provides the basis for the follow?up control behavior. Finally, a RLV re?entry section is taken as an example. After the various types of faults are injected into it, the correctness of the e?vent diagnosis system is verified by C++software. The result proves that the diagnosis system can diagnose and dis?tinguish different types of events correctly.%高速飞行器在遇到外部大气扰动和伺服故障事件导致状态失稳时,对于不同的故障事件采用不同的控制策略.为了使飞行器能够自主的辨别导致状态异常的事件类型,并针对性调用相应的处理策略,提出了一种基于决策网络进行飞行器异常事件诊断的系统.基于Petri net方法构建的飞行器异常事件诊断系统,通过分析差异事件对飞行器状态的影响规律,以各状态量传感器读数为输入,发生的事件类型为输出,可以使飞行器自主利用Petri net的矩阵推理运算能力快速在线定位出引起状态异常的事件类型,为后续采取的控制行为提供依据.最后以某RLV再入段为例,对其注入各种类型故障后,应用C++软件对事件诊断系统正确性进行仿真验证,结果证明建立的诊断系统能够正确的诊断和区分不同类型事件.
展开▼
