A novel unsupervised learning method for intelligent fault diagnosis of rolling element bearings based on deep functiona
- PDF / 3,144,149 Bytes
- 15 Pages / 595.22 x 842 pts (A4) Page_size
- 46 Downloads / 176 Views
DOI 10.1007/s12206-020-1002-x
Journal of Mechanical Science and Technology 34 (11) 2020 Original Article DOI 10.1007/s12206-020-1002-x Keywords: · Rolling element bearings · Improve intelligent diagnostics · Deep functional auto-encoder · Unsupervised learning enhancement · Massive raw data analysis
Correspondence to: Jianping Xuan [email protected]
Citation: Aljemely, A. H., Xuan, J., Jawad, F. K. J., Al-Azzawi, O., Alhumaima, A. S. (2020). A novel unsupervised learning method for intelligent fault diagnosis of rolling element bearings based on deep functional autoencoder. Journal of Mechanical Science and Technology 34 (11) (2020) 4367~4381. http://doi.org/10.1007/s12206-020-1002-x
Received December 3rd, 2019 Revised
February 12th, 2020
Accepted August 4th, 2020 † Recommended by Editor No-cheol Park
A novel unsupervised learning method for intelligent fault diagnosis of rolling element bearings based on deep functional auto-encoder Anas H. Aljemely1, Jianping Xuan1, Farqad K. J. Jawad2, Osama Al-Azzawi2 and Ali S. Alhumaima3 1
School of Mechanical Science and Engineering, Huazhong University of Science and Technology, Wu2 han 430074, China, School of Civil Engineering and Mechanics, Huazhong University of Science and 3 Technology, Wuhan 430074, China, School of Electronic Engineering and Computer Science, South Ural State University, Chelyabinsk 454016, Russia
Abstract
Recently, several studies tried to develop fault identification models for rolling element bearing based on unsupervised learning techniques. However, an accurate intelligent fault diagnosis system is still a big challenge. In this study, a deep functional auto-encoders (DFAEs) model with SoftMax classifier was designed for valuable feature extraction from massive raw vibration signals. To maximize the unsupervised feature learning ability of the proposed model, various activation functions were applied in an effective methodology, these hidden activation functions enhance significantly the sparsity of the training data-set. The proposed method was validated using the raw vibration signals measured from the machine with different bearing conditions. The achieved results showed that the high-superiority of the proposed model comparing to standard deep learning and other traditional fault diagnosis methods in terms of classification accuracy even with massive input data sets.
1. Introduction
© The Korean Society of Mechanical Engineers and Springer-Verlag GmbH Germany, part of Springer Nature 2020
Rotating machinery is extensively used in the newly industry. The key components of the rotating machinery can develop diverse faults under rigorous working situations such as large load, strong impact, high speed, and high background noise [1, 2]. Rolling element bearings are vital components of most rotating machinery and electrical apparatuses. The latter’s failures may cause significant losses and serious economic casualties. Therefore, it is worth to diagnose the different bearing faults precisely and automatically that may occur in rotat
Data Loading...
