Affiliation of Author(s):民航学院

Journal:Hangkong Dongli Xuebao

Abstract:A method of multi-features fusion based on normalized Euclidean distance was proposed to meet the requirement of monitoring aero-engine rolling bearing real time condition. Firstly, the bearing fault simulation experiments were carried out while the fault sensitivity of the features before and after fusion was analysed by introducing a quantitative evaluation index of fault sensitivity. Then, the proposed method was compared with the principal component analysis, support vector data description and support vector distribution estimation. Finally, the bearing fatigue accelerated experiment was carried out, and the proposed fusion method was applied to the aero-engine rolling bearing condition monitoring. Experimental results show that compared with principal component analysis, support vector data description and support vector distribution estimation, the fault sensitivity of fusion value based on normalized Euclidean distance is higher. For different types and different stages of the bearing fault, the fusion value obtained by proposed method is more sensitive, which is more suitable to be an index of the bearing condition monitoring, compared with the effective value. © 2017, Editorial Department of Journal of Aerospace Power. All right reserved.

ISSN No.:1000-8055

Translation or Not:no

Date of Publication:2017-09-01

Co-author:Lin, Tong,Zhang, Quande,Wang, Hongwei,Chen, Libo

Correspondence Author:cg