一种基于截断奇异值分解的强跟踪平方根卡尔曼滤波估计燃气轮机健康参数的方法  

Estimation Method of Health Parameters of Gas Turbine by Strong Tracing and Square Root Kalman Filter based on Truncated Singular Value Decomposition

作　　者：房友龙 贺星[2] 刘东风[2,3] 刘冲培 周志才 Fang Youlong;He Xing;Liu Dongfeng;Liu Chongpei;Zhou Zhicai(Institute of Construction and Development,Chinese People's Armed Police Force Research Institute,Beijing 100012,China;College of Power Engineering,Naval University of Engineering,Wuhan 430033,China;Unit No.91663,Chinese People's Liberation Army,Shandong Qingdao 266012,China)

机构地区：[1]中国人民武装警察部队研究院建设发展研究所,北京100012 [2]海军工程大学动力工程学院,武汉430033 [3]中国人民解放军91663部队,山东青岛266012

出　　处：《燃气轮机技术》2025年第1期24-32,共9页Gas Turbine Technology

摘　　要：燃气轮机健康参数的估计对其健康管理非常重要。在采用扩展卡尔曼滤波对燃气轮机健康参数进行估计过程中,针对燃气轮机监测参数少于健康参数而导致系统可观性不足的问题,采用截断奇异值分解降低健康参数估计维数的方法改进系统的可观性;针对模型和噪声统计特性的误差和递推过程中计算机舍入误差导致滤波发散和计算发散问题,提出一种强跟踪平方根卡尔曼滤波改善滤波稳定性的方法,并应用于实例分析。分析结果表明本文提出的方法可有效改善估计系统可观性,提高滤波稳定性。Estimation of gas turbine health parameters is very important for its health management.In the process of estimating the health parameters of gas turbine by using extended Kalman filter,aiming at the problem that the monitoring parameters of gas turbine were less than the health parameters,which leaded to the insufficient observability of the system,the method of truncation singular value decomposition(SVD)to reduce the estimated dimension of the health parameters was adopted to improve the observability of the system.Aiming at the problems of filtering divergence and calculation divergence caused by the error of statistical characteristics of model and noise and the rounding error of computer in recursive process,a method of strong tracking and square root Kalman filter was proposed to improve filtering stability and applied to an example analysis.The analysis results show that the proposed method can effectively improve the observability of the estimation system and improve the filtering stability.

关 键 词：燃气轮机 健康参数 卡尔曼滤波 截断奇异值分解 强跟踪 平方根 

分 类 号：TK472[动力工程及工程热物理—动力机械及工程]