In order to obtain a sensitive bearing degradation indicator with significant trend, a new approach is proposed. For the capability of preserving the local structure of data manifold, locality preserving projections are used for features dimension reduction without missing the original information. Since continuous hidden Markov model (CHMM) can take multiple fault features into consideration and also capture the stochastic characteristics of fault features in the time domain, CHMM-based negative log likelihood probability is built as a new health assessment indication. The experiment results show that, compared with the general features, the proposed degradation indicator shows great trend of the bearing health and is more sensitive to incipient defects.
Rolling elements bearings Locality preserving projections Continuous hidden Markov model Negative log likelihood probability Fault features Degradation indicator
This is a preview of subscription content, log in to check access.
The work described in this paper was supported by a grant from the National Defence Researching Fund (No. 9140A27020413JB11076).
A. Malhi, R. Yan, R.X. Gao, Prognosis of defect propagation based on recurrent neural networks. Prognosis of defect propagation based on recurrent neural networks. IEEE Trans. Instrum. Meas. 60(3), 703–711 (2011)CrossRefGoogle Scholar
Z. Tian, M.J. Zuo, Health condition prediction of gears using a recurrent neural network approach. IEEE Trans. Reliab. 59(4), 700–705 (2010)CrossRefGoogle Scholar
H. Qiu, J. Lee, J. Lin, G. Yu, Robust performance degradation assessment methods for enhanced rolling element bearing prognostics. Adv. Eng. Inform. 17(3–4), 127–140 (2003)CrossRefGoogle Scholar
R. Huang, L. Xi, X. Li, C.R. Liu, H. Qiu, J. Lee, Residual life predictions for ball bearings based on self-organizing map and back propagation neural network methods. Mech. Syst. Signal Process. 21(1), 193–207 (2007)CrossRefGoogle Scholar
J. Yu, Bearing performance degradation assessment using locality preserving projections and Gaussian mixture models. Mech. Syst. Signal Process. 25(7), 2573–2588 (2011)CrossRefGoogle Scholar
J. Yu, A nonlinear probabilistic method and contribution analysis for machine condition monitoring. Mech. Syst. Signal Process. 37, 293–314 (2013)CrossRefGoogle Scholar
L. Linxia, Discovering prognostic features using genetic programming in remaining useful life prediction. IEEE Trans. Ind. Electron. 61(5), 2464–2472 (2014)CrossRefGoogle Scholar
Q. Liu, M. Dong, Y. Peng, A novel method for online health prognosis of equipment based on hidden semi-Markov model using sequential Monte Carlo methods. Mech. Syst. Signal Process. 32, 331–348 (2012)CrossRefGoogle Scholar
J. Yan, C. Guo, X. Wang, A dynamic multi-scale Markov model based methodology for residual life prediction. Mech. Syst. Signal Process. 25(4), 1364–1376 (2011)CrossRefGoogle Scholar
D. Yu, L. Deng, X. He, A. Acero, Large-margin minimum classification error training: a theoretical risk minimization perspective. Comput. Speech Lang. 22(4), 415–429 (2008)CrossRefGoogle Scholar
S. Roweis, L. Saul, Nonlinear dimensionality reduction by locally linear embedding. Science 290, 2323–2326 (2000)CrossRefGoogle Scholar
R.I.A. Davis, B.C. Lovell, Comparing and evaluating HMM ensemble training algorithms using train and test and condition number criteria. Formal Pattern Anal Appl. 6, 327–336 (2003)Google Scholar