taptap下载安装安卓学报

• 民用航空 • 上一篇    下一篇

基于SVM 的放气活门调节模型多元回归分析

曹惠玲,王新   

  1. (taptap下载安装安卓航空工程学院,天津300300)
  • 收稿日期:2016-11-23 修回日期:2017-02-20 出版日期:2017-10-25 发布日期:2017-12-14
  • 作者简介:曹惠玲(1962—),女,河北唐山人,教授,工学博士,研究方向为航空发动机性能分析与故障诊断.
  • 基金资助:
    中央高校基本科研业务费专项(3122014D010)

Multiple regression analysis of bleed valve regulation model based on SVM

CAO Huiling, WANG Xin   

  1. (College of Aeronautical Engineering, CAUC, Tianjin 300300, China)
  • Received:2016-11-23 Revised:2017-02-20 Online:2017-10-25 Published:2017-12-14

摘要: 航空发动机2.5 级放气活门是重要的防喘机构。本研究针对轴流式压气机喘振机理进行分析,使用支持向量机(SVM,support vector machine)方法,通过对PW4077D 发动机30 个无故障航班QAR(quick access recorder)数据的处理与多元回归分析,分别建立了3 个航段下2.5 级放气活门的辨识模型。随机选取一次航班的QAR数据,使用已建立的模型对一次航班的2.5 级放气活门开度进行计算并与实际开度值进行对比,验证了通过回归分析得到的辨识模型具有较高的准确度,对于航空发动机状态监控和故障诊断具有一定的应用价值。

关键词: 支持向量机, 回归分析, 辨识模型, QAR, PW4077D, 2.5 级放气活门

Abstract: Aero engine's 2.5 bleed valve is an important mechanism to eliminate the risk of stall and surge. Stall and surge mechanism of axial-flow compressor are analyzed; after 30 trouble-free flights'QAR (quick access recorder)data analyzed and disposed, 3 flight sections'identification models of 2.5 bleed valve control models are set up by using SVMmethod. Applicating the models on a flight's QAR data, the bleed valve opening is calculated and compared with the actual value, proving that the models have higher accuracy and certain application value in the condition monitoring and fault diagnosis of aero-engine.

Key words: SVM, regression analysis, identification model, QAR, PW4077D, 2.5 bleed valve

中图分类号: 

Baidu
map