• CN:11-2187/TH
  • ISSN:0577-6686

机械工程学报 ›› 2019, Vol. 55 ›› Issue (13): 113-121.doi: 10.3901/JME.2019.13.113

• 特邀专栏:航空发动机健康监测与故障诊断 • 上一篇    下一篇

基于压缩感知的叶端定时欠采样多频叶片振动盲重构研究

徐海龙1,2, 杨拥民1, 胡海峰1, 官凤娇1, 陈仲生1   

  1. 1. 国防科技大学装备综合保障国防科技重点实验室 长沙 410073;
    2. 株洲中车时代电气股份有限公司 株洲 412001
  • 收稿日期:2018-07-19 修回日期:2018-12-07 出版日期:2019-07-05 发布日期:2019-07-05
  • 通讯作者: 杨拥民(通信作者),男,1966年出生,博士,教授,博士研究生导师。主要研究方向为状态监控与故障诊断,信号处理等。E-mail:yangyongmin@163.com
  • 作者简介:徐海龙,男,1989年出生,博士。主要研究方向航空发动机压气机叶片动力学分析及裂纹检测。E-mail:xhlym1@163.com
  • 基金资助:
    国家重点基础研究发展计划资助项目(973计划,2015CB057400)。

Compressed Sensing-based Blind Reconstruction of Multi-frequency Blade Vibration from Under-sampled BTT Signals

XU Hailong1,2, YANG Yongmin1, HU Haifeng1, GUAN Fengjiao1, CHEN Zhongsheng1   

  1. 1. Science and Technology on Integrated Logistics Support Laboratory, National University of Defense Technology, Changsha 410073;
    2. Zhuzhou CRRC Times Electric Co., Ltd., Zhuzhou 412001
  • Received:2018-07-19 Revised:2018-12-07 Online:2019-07-05 Published:2019-07-05

摘要: 旋转叶片是航空发动机的核心部件,长期工作在复杂的工况下承受交变应力,容易产生振动而导致疲劳失效,因此在线监测叶片振动对发动机运行安全具有十分重要的工程意义。针对传统的接触式应变法无法同时测量所有叶片振动且布线复杂存在安全隐患的问题,采用叶端定时(Blade tip-timing,BTT)非接触方法对叶片振动进行实时在线监测。但是BTT测量信号受叶顶传感器安装限制属于典型的欠采样信号,而叶片由于气动激励及微小裂纹的非线性导致其叶端出现多频振动响应,因此利用欠采样的BTT信号获取未知多频叶片振动是目前遇到的巨大挑战。提出采用压缩感知方法解决BTT欠采样多频叶片振动盲重构。首先分析叶片在气动激励下的多频响应;然后建立BTT测量的压缩感知模型,并采用多重信号分类(Multiplesignal classification,MUSIC)算法进行求解;最后通过数值仿真,并结合旋转叶片BTT测振实验台验证了压缩感知理论解决BTT欠采样多频叶片振动盲重构难题,实现旋转叶片振动BTT非接触在线测量。

关键词: 航发发动机, 欠采样信号重构, 压缩感知, 叶端定时, 叶片振动监测

Abstract: The rotated blade is the core component of aeroengine. However, the blades work in complex working conditions for a long time and are subjected to alternating stresses, which make balde vibrations and result in fatigue failure. Therefore, on-line vibration monitoring is of great engineering significance to operation safety for aeroengine. As the traditional contact strain method cannot measure vibration of all blades at the same time and the arrangement of wire is complex which may bring security risks, Blade tip-timing (BTT) technology is used for the blade real time online vibration monitoring. However, BTT signals are typically under-sampled signals by limited installation of BTT sensors. In addition, blade vibrations are always multi-frequency with less prior knowledge due to crack nonlinearity and complicated aerodynamic excitations. So it is a big challenge that unknowns multi-frequency blade vibrations are reconstructed from BTT under-sampled measurements. The compressed sensing (CS) theory is proposed to reconstruct unknown multi-band blade vibrations from under-sampled BTT measurements. Firstly, it is analyzed that multi-frequency blade vibrations are excited by aerodynamic excitations. Then, The CS model of BTT measurement is built and the MUSIC algorithm is selected to solve the model. Finally, numerical simulations and vibrations measurement of rotated blades by BTT method are used to validate the proposed method.

Key words: aeroengine, blade tip-timing, blade vibration monitoring, compressed sensing, under-sampled signal reconstruction

中图分类号: