Founded in 1971 >
Chinese Sci-tech Core Periodicals >
British Science Abstracts (SA, INSPEC) Indexed Journals >
United States, Cambridge Scientific Abstract: Technology (CSA: T) Indexed Journals >
United States, Ulrich's Periodicals Directory(UPD)Indexed Journals >
United States, Cambridge Scientific Abstract: Natural Science (CSA: NS) Indexed Journals >
Poland ,Index of Copernicus(IC) Indexed Journals >
International Standard Serial Number:
ISSN 1001-4551
Sponsor:
Zhejiang University;
Zhejiang Machinery and Electrical Group
Edited by:
Editorial of Journal of Mechanical & Electrical Engineering
Chief Editor:
ZHAO Qun
Vice Chief Editor:
TANG ren-zhong,
LUO Xiang-yang
Tel:
86-571-87041360,87239525
Fax:
86-571-87239571
Add:
No.9 Gaoguannong,Daxue Road,Hangzhou,China
P.C:
310009
E-mail:
meem_contribute@163.com
Abstract: A fault diagnosis method of typical parts of mechanical equipment(gear, bearing)based on Triplet Network was proposed,for the problem that the fault diagnosis accuracy of deep neural network in typical parts of mechanical equipment(gear、bearing)waslow under small sample conditions. Firstly, the original timing signal was converted into time-frequency signal by using short-time Fourier transform.Then, the model based on triplet network was used to extract the features of the same fault and different fault samples from the time-frequency signal. By comparing the similarity of the features of the same fault and different fault samples, the model parameters were optimized to achieve the effect that the feature similarity of the extracted same fault samples was higher and higher, and the feature similarity of different fault samples was lower and lower.Finally, fault recognition was realized by comparing the feature similarity between unknown samples and known fault samples; the effectiveness of the fault diagnosis method was verified by using the bearing fault data set of Jiangnan University and the gear fault data set of the University of Connecticut. The research results show that when there are only 5 training samples in each type, the recognition rate of bearing faults can reach 68%, and the recognition rate of gear faults can reach 96.8%, which is better than the traditional deep neural network method.
Key words: gear fault diagnosis; rolling bearing fault diagnosis; deep neural network; feature similarity; fault recognition rate; time-frequency signal; small samples
-
- Chinese Core Periodicals
-
- Chinese Sci-tech Core Periodicals
-
- SA, INSPEC Indexed
-
- CSA: T Indexed
-
- UPD:Indexed
-
Copyright 2010 Zhejiang Information Institute of Mechinery Industry All Rights Reserved
Technical Support：Hangzhou Bory science and technology
