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: Effective failure prediction and health management (PHM) for wind turbine gearbox bearings can reduce the failure rate of wind turbine bearing components. Aiming at the problem of accurate prediction of remaining residual service life (RUL) of high-speed bearing of wind turbine, a data-driven RUL prediction method for wind turbine high-speed bearings based on spectral shape factor (SSF) and Elman neural network (ENN) data was proposed. Firstly, Teager energy operator (TEO) was introduced to preprocess the original vibration signal of high-speed shaft of wind turbine. Secondly, based on short-time Fourier transform (STFT), an SSF was constructed to transform each fault characteristic index of the bearing. Then, based on the monotonicity, trendability and predictability of TEO energy signal, the fitness function was constructed, the transformed indexes were screened, and the fault characteristic index most suitable for predicting the bearing RUL was determined. The RUL prediction experiment of a specific wind turbine highspeed bearing in actual operation was carried out by using ENN and measured data. Finally, the method based on SSF and ENN was quantitatively compared with three existing data-driven techniques. The results show that after TEO pretreatment and SSF transformation, the fitness of logarithmic entropy of the vibration signal is the highest. At the same time, compared with the other three kinds of data-driven techniques, the RUL prediction method has higher RUL prediction accuracy, which can maintain a high level within 35 days. Thus, the applicability and rationality of the RUL prediction method can be well proven.
Key words: high-speed bearings; spectral shape factor (SSF); Elman neural network(ENN); remaining useful life(RUL); shorttime Fourier transform(STFT); datadriven; logarithmic entropy; Teager energy operator (TEO)