中国民航大学学报 ›› 2021, Vol. 39 ›› Issue (3): 29-33.

• 民用航空 • 上一篇    下一篇

基于 GA-LSSVM 的航空发动机气路故障诊断 

田 静,胡鹤翔    

  1. (中国民航大学航空工程学院,天津 300300) 
  • 收稿日期:2020-05-20 修回日期:2020-05-20 接受日期:2020-04-21 出版日期:2021-06-25 发布日期:2021-11-28
  • 作者简介:田静(1972—),女,辽宁兴城人,教授,博士,研究方向为飞机系统。

Fault diagnosis of aero-engine gas path based on GA-LSSVM 

TIAN Jing, HU Hexiang    

  1. (College of Aeronautical Engineering, CAUC, Tianjin 300300, China) 
  • Received:2020-05-20 Revised:2020-05-20 Accepted:2020-04-21 Online:2021-06-25 Published:2021-11-28

摘要: 针对航空发动机气路故障数据样本少、非线性的问题,将遗传算法(GA,genetic algorithm)优化的最小二乘支持 向量机(LSSVM,least square support vector machine)应用于航空发动机气路故障诊断。 首先对 LSSVM 算 法中的关键参数进行分析,并用 GA 对 LSSVM 中的正则化参数(C)、核参数(g)进行优化;其次采用 GALSSVM 对某型双转子涡喷发动机进行气路故障诊断; 最后从诊断精度、 抗噪能力和训练耗时 3 方面与 SVM、LSSVM 和 GA-SVM 比较。 结果表明:GA-LSSVM 在诊断精度和抗噪强度方面都高于其他 3 种算法, 但由于 GA-LSSVM 有寻优过程,故训练时间较长。

关键词: 航空发动机, 气路故障诊断, 最小二乘支持向量机, 遗传算法

Abstract: Aiming at the lack of data sample and nonlinearity of aero-engine gas-path fault, the genetic algorithm (GA) optimized least square support vector machine (LSSVM) is applied to aero-engine gas-path fault diagnosis. Firstly, the key parameters of LSSVM algorithm is analyzed, and the regularization parameter(C) and kernel parameter ( g ) of LSSVM are optimited with GA. Secondly, GA-LSSVM is used to diagnose the gas path fault of a biaxial turbofan engine. Finally, SVM, LSSVM, GA-SVM and GA-LSSVM are compared from three aspects of diagnosis accuracy, anti-noise ability and training time. Results show that GA-LSSVM is superior to the other three algorithms in diagnosis accuracy and anti noise intensity, but its training time is longer due to the optimization process of GA-LSSVM.

Key words: aero-engine, gas path fault diagnosis, least square support vector machine(LSSVM), genetic algorithm (GA) 

中图分类号: