航空学报 > 2013, Vol. 34 Issue (11): 2529-2538   doi: 10.7527/S1000-6893.2013.0225

基于非线性自适应滤波的发动机气路部件健康诊断方法

鲁峰1,2, 黄金泉1, 吕怡秋1, 仇小杰2   

  1. 1. 南京航空航天大学 能源与动力学院, 江苏 南京 210016;
    2. 中航工业航空动力系统控制研究所, 江苏 无锡 214063
  • 收稿日期:2013-01-16 修回日期:2013-04-26 出版日期:2013-11-25 发布日期:2013-06-09
  • 通讯作者: 黄金泉,Tel.:025-84895995 E-mail:jhuang@nuaa.edu.cn E-mail:jhuang@nuaa.edu.cn
  • 作者简介:鲁峰 男,博士,副教授,硕士生导师。主要研究方向:航空发动机预测健康管理、建模与仿真、信息融合等。 Tel:025-84892203-2103 E-mail:lufengnuaa@126.com;黄金泉 男,博士,教授,博士生导师。主要研究方向:航空发动机建模、控制与故障诊断等。 Tel:025-84895995 E-mail:jhuang@nuaa.edu.cn
  • 基金资助:

    国家自然科学基金(51276087);中国博士后科学基金(2013M530256);江苏省自然科学基金(BK20130802);江苏省博士后科学基金(201202063)

Aircraft Engine Gas-path Components Health Diagnosis Based on Nonlinear Adaptive Filters

LU Feng1,2, HUANG Jinquan1, LV Yiqiu1, QIU Xiaojie2   

  1. 1. College of Energy and Power Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, China;
    2. AVIC Aviation Power Control System Institute, Wuxi 214063, China
  • Received:2013-01-16 Revised:2013-04-26 Online:2013-11-25 Published:2013-06-09
  • Supported by:

    National Natural Science Foundation of China (51276087);China Postdoctoral Science Foundation (2013M530256);Natural Science Foundation of Jiangsu Province (BK20130802);Postdoctoral Science Foundation of Jiangsu Province (201202063)

摘要:

针对发动机气路突变故障诊断精度不高以及算法工程化验证周期长的问题,提出了线性自适应卡尔曼滤波算法,且将其扩展至非线性系统,并在快速原型试验平台上实现算法快速验证。在非线性滤波算法的状态方程中引入状态突变因子,采用统计意义的广义似然比检验方法,通过测量残差对气路部件健康参数的突变与否进行检验,解决了发动机气路健康参数突变的准确估计,搭建基于NI CRIO的航空发动机气路性能分析快速原型试验平台,实现了非线性自适应滤波算法在快速原型验证平台的部署及快速验证。以某型大涵道比涡扇发动机为对象,通过数字仿真与快速原型平台验证了非线性自适应滤波算法相比于常规扩展卡尔曼滤波(EKF)具有更好的突变诊断能力,同时具有较高的渐变诊断能力。

关键词: 航空发动机, 气路分析, 扩展卡尔曼滤波, 自适应滤波, 快速原型

Abstract:

To deal with the issue of poor accuracy of gas-path abrupt fault diagnosis and the long term required for algorithm validation, a detailed algorithm of linear adaptive Kalman filter is presented and extended to the nonlinear system, and then validated on a rapid prototyping platform. A tuning factor is introduced to the state equation of the nonlinear filter, and a generalized likelihood ratio test is used to detect and estimate an abrupt fault by monitoring the residuals. Gas-path abrupt faults can be diagnosed by the shift of the tuning factor in the nonlinear filter algorithm. Then the proposed algorithm is validated on the NI CRIO test of aircraft engine gas-path analysis, and it is realized by the rapid prototyping with arrangement and downloads. Tests on a high bypass ratio turbofan engine through digital simulation and rapid prototyping platform show that the adaptive filter algorithm can obtain estimates of both abrupt and gradual deteriorations more accurately than the conventional extended Kalman filter (EKF) algorithm.

Key words: aircraft engine, gas-path analysis, extended Kalman filter (EKF), adaptive filter, rapid prototyping

中图分类号: