Bearing remaining useful life estimation based on time–frequency representation and supervised dimensionality reduction. (May 2016)

• Record Type: Journal Article
• Authors: Zhao, Minghang; Tang, Baoping; Tan, Qian
• Is Part Of: Measurement. Volume 86(2016:May.)
• Page Start: 41
• Page End: 55
• Publication Date: 2016-05
• Subjects: LBP local binary pattern -- LDA linear discriminant analysis -- MAE mean absolute error -- MAPE mean absolute percentage error -- PCA principal component analysis -- RMSE root mean square error -- RMS root mean square -- RUL remaining useful life -- STFT short-time Fourier transform -- TFR time–frequency representation; Remaining useful life estimation -- Time–frequency representation -- Supervised dimensionality reduction -- Multiple linear regression -- Bearing; Weights and measures -- Periodicals; Measurement -- Periodicals; Measurement; Weights and measures; Periodicals; 530.8
• DOI: 10.1016/j.measurement.2015.11.047
• Languages: English
• ISSNs: 0263-2241

Abstract

Highlights: Texture feature of time–frequency representation serves as high-dimensional feature. Linear discriminant analysis is used in the dimensionality reduction. Multiple linear regression is used for the final remaining useful life estimation. The proposed approach outperforms the methods using statistical features and/or PCA. Abstract: The extraction of ideal age feature is a challenging task in vibration-based bearing remaining useful life (RUL) estimation. Aiming at this problem, a new approach is proposed on the basis of time–frequency representation (TFR) and supervised dimensionality reduction. Firstly, S transform and Gaussian pyramid are employed to obtain TFRs at multiple scales. Textural features of TFRs are used as the high-dimensional features. Then, a two-step supervised dimensionality reduction technique, i.e. principal component analysis (PCA) plus linear discriminant analysis, is employed to reduce the dimensionality, in which the target dimension and number of classes are taken as variable parameters. Finally, the simple multiple linear regression model is utilized to estimate the RUL. Experimental results indicate that the proposed approach outperforms the methods using traditional statistical features and/or PCA. Additionally, variable conditions of load and speed should be considered in the future to further improve the proposed approach.