摘要
以大型液体火箭发动机故障诊断系统框架为基础，按照发动机不同工作阶段的特点建立了相应的故障检测与诊断算法。利用发动机的高队数学模型，对基于推广的卡尔曼滤波器技术的新息检测算法进行了研究。根据发动机系统工作过程的特点，建立了降阶故障模式响应模型，并发展了相应的故障模式检验方法。为了适应在线实时检测的需要，利用发动机的试验数据，分别研究了基于人工神经元网络辨识模型的发动机启动过程检测算法和基于时间序列分析方法的发动机主机工作过程检测算法。通过试验数据的检测验证，证明了这些实时算法的快速性、准确性和鲁律性。
The algorithms of the failure detection and diagnosis of liquld propellant rocketengines play a key role in the heaIth monitoring techniques. According to the stage of the engineoperation, deferent detection and diagnosis algorithms are developed. A model based detectionalgorithm is constructed via using the Extended Karlman Filtering methed' Reduced order failureresponse medels are used to develop medel based failure diagnosis algorothms. To improve the realtime detection performance, neural network based identification model is constructed for the failuredetection of the starting process of the engine, and time sequence medel is developed for detection ofthe failure during the engine main-stage process. The efficiency of the detection algorithms is verifiedthrough test data based computation.
出处
《推进技术》
EI
CAS
CSCD
北大核心
1997年第1期13-17,共5页
Journal of Propulsion Technology
关键词
液体推进剂
火箭发动机
故障诊断
故障检测
Liquid propellant rocket engine, Engine fault, Fault diagnosis, Fault detection,Artificial neural network, Real time algorithm
