基于H∞及动态阈值的主燃油回路故障检测
Fault Detection of Main Fuel Loop Based on H∞ and Dynamic Threshold
李宏新 1郝彬彬 2董久祥3
作者信息
- 1. 中国航空发动机集团公司,北京 100097
- 2. 中国航发沈阳发动机研究所,沈阳 110015
- 3. 中国航发沈阳发动机研究所,沈阳 110015;东北大学 信息科学与工程学院,沈阳 110819
- 折叠
摘要
为了解决航空发动机主燃油伺服回路传感器故障率高、固定阈值判定算法精度低的问题,提出了一种基于有限频H∞性能指标的扩展型观测器设计方法和一种动态阈值设计算法,实现主燃油伺服回路早期的故障检测.建立了包含非线性特性的主燃油控制回路数学模型,基于执行器及传感器所受扰动的频率分布区间,采用有限频H∞性能指标,设计了观测器增益保证扰动到残差输出的无穷范数最小,增强了故障检测系统的抗干扰性.基于白噪声的概率统计特性提出了一种自适应动态阈值设计方法,降低了传统故障检测方法中的误报率和漏报率,提高了故障的检测精度,并通过仿真对该方法进行了验证.结果表明:针对不同的故障类型,该方法均可提高故障检测实时性,并提高故障检测精度,误报率和漏报率分别从5.5%和20.5%降至1%和9%,对航空发动机的健康维护具有重要意义.
Abstract
To solve the problems of the high fault rate of sensors in the main fuel servo loop of aeroengines and the low accuracy of the fault detection algorithm with a fixed threshold value,an extended observer design method based on finite-frequency H∞ performance index and a dynamic threshold design algorithm was proposed to realize early fault detection of main fuel servo loop.A mathematical model of the main fuel control loop covering nonlinear characteristics was established,based on the frequency distribution of the disturbance to which the actuator and sensors were subjected,the gain of the observer was designed to minimize the infinity norm of the disturbance to the residual output by using the finite frequency H∞ performance index,which enhances the anti-interference properties of the fault detection system.A self-adaptive dynamic threshold design method was proposed based on the probabilistic statistical characteristics of white noise,which reduces the false alarm rate and missed alarm rate in traditional fault detection methods and improves fault detection accuracy,and was verified by simulation.The results show that for different types of faults,the method can improve the real-time performance of fault detection and improve the accuracy of fault detection,the false alarm rate and omission rate are reduced from 5.5%and 20.5%to 1%and 9%,respectively,which is of great significance to the health maintenance of aeroengines.
关键词
主燃油回路/动态阈值/故障检测/航空发动机Key words
main fuel loop/dynamic threshold/fault detection/aeroengine引用本文复制引用
基金项目
国家自然科学基金重点项目(61890925)
出版年
2024
NETL
NSTL
万方数据