航空发动机部件性能退化参数的分布式估计算法  

Distributed Estimation Algorithm for Aero-engine Deviation Parameters

作　　者：尹大伟[1] 吕日毅[1] 常斌[1] 颜仙荣[1] 

机构地区：[1]海军装备研究院,上海200436

出　　处：《航空学报》2013年第12期2716-2724,共9页Acta Aeronautica et Astronautica Sinica

基　　金：某"十二五"预研项目~~

摘　　要：航空发动机部件性能退化参数(EDPs)估计是发动机性能寻优控制(PSC)的关键技术之一。针对应用传统的集中式Kalman滤波算法估计EDPs存在计算效率不高、容错性差等不足,提出了采用分布式滤波的思想估计性能退化参数。以集中式Kalman滤波算法估计EDPs的状态变量模型(SVM)和递推算法为基础,引入信息融合,设计了一种结构简洁的联邦滤波器。根据联邦滤波器的结构和递推公式,从理论上分析了这类分布式估计算法的优势。最后以某型涡扇发动机为例,对联邦滤波器的估计能力进行了仿真验证,应用设计的联邦滤波器估计EDPs,并与集中式Kalman滤波算法的估计结果进行比较。仿真结果表明,分布式计算模式的联邦滤波算法能迅速收敛,且估计精度明显高于集中式Kalman滤波算法。本文所做的研究对发动机PSC的发展具有一定的理论意义和工程应用价值。Aero-engine deviation parameters （EDPs） estimation is one of the key technologies of performance seeking con- trol （PSC）. The idea of using distributed filtering to estimate EDPs is proposed, because there are some disadvantages in the traditional centralized Kalman filtering,such as low computational efficiency, poor fault-tolerance, etc. A simple federa- ted filtering with information fusion is designed to estimate the EDPs, which is based on the traditional state variable model and Kalman filtering algorithm. The advantages of the designed federated filtering are analyzed in theory in terms of the construction and recurrence equations. Finally, the capability of federated filtering is certified using a given turbo fan engine by simulation, The EDPs are estimated by applying the designed federated filtering, and the estimation results are compared with those of traditional Kalman filtering＇s, which show that federated filtering with distributed calculation can realize conver- gencde more quickly, and the estimation precision is evidently higher. This study may have some theoretical significance and practical value for the development of PSC.

关 键 词：航空发动机 参数估计 性能寻优控制 KALMAN滤波 分布式估计 联邦滤波 

分 类 号：V233.7[航空宇航科学与技术—航空宇航推进理论与工程]