基于融合状态递推的非线性联邦滤波器故障检测算法  

Nonlinear Federated Filter Fault Detection Algorithm Based on Fusion State Propagator

作　　者：卢航 郝顺义[1] 沈飞[1] 李保军 Lu Hang;Hao Shunyi;Shen Fei;Li Baojun(Aeronautics and Astronautics Engineering College, Air Force Engineering University, Xi＇an 710038, China)

机构地区：[1]空军工程大学航空航天工程学院,西安710038

出　　处：《航天控制》2018年第2期88-92,共5页Aerospace Control

基　　金：陕西省自然科学基金(2017JQ6034);民机专项(MJZ-2014-S-47);航空科学基金(20110896009)

摘　　要：传统的联邦滤波器故障检测算法一般是针对线性系统模型提出的,现在非线性条件下对算法进行推广,推导了针对非线性模型的状态χ~2故障检测算法,并对其进行改进,提出一种基于融合状态递推的新型联邦滤波故障检测算法。该算法利用联邦滤波器的全局最优融合状态通过状态递推器获得融合状态递推,再由各子系统通过非线性滤波得到的状态估计与融合状态递推之间的残差构造检验量,进行故障检测。由于全局最优融合状态的精度要高于局部滤波值,故相比于传统联邦故障检测法,该算法的检测灵敏度更高;最后设计了一组仿真实验,验证了该检测算法的有效性。Traditional federal filtering fault detection algorithm is implemented for linear system model,this algorithm is extended and the state chi-square test is deduced under the condition of nonlinear model. A new federal filtering fault detection algorithm based on fusion state propagator is presented for improving the traditional algorithm structure. This arithmetic take advantage of federated filter＇s global optimal fusion state estimation to gain fusion state recursion by state propagator,and the residual error is used between state estimation gained by nonlinear filter and fusion state recursion for building detection values and implement the fault detection. Because of the accuracy of global optimal fusion state estimation is higher than that of local estimation,the detection sensitivity of this algorithm is higher accordingly. Finally,a simulation experiment is designed and the effectiveness of this algorithm is verified.

关 键 词：故障检测 联邦滤波 非线性模型 

分 类 号：V249.3[航空宇航科学与技术—飞行器设计]