Multi-Kernel PCA with Discriminant Manifold for Hoist Monitoring
Abstract:
Dimensional reduction is crucial to machine condition monitoring and diagnosis since the extracted features are often redundant and heterogeneous as well as high dimensional data often are embedded in lower-dimensional manifold. Inspired by manifold learning and multiple-kernel learning theory, multi-kernel principal component analysis with discriminant manifold (DMMKPCA) is proposed for host fault monitoring and diagnosis. The method not only preserves the local and global structures of data set but also handles heterogeneous characteristic sets, which inherits the excellences of LPMVP and multiple kernel learning. A two-stage iterative optimization algorithm is proposed to obtain the optimal combining weights of multiple kernel functions and parameters of each kernel functions. A case study of hoist illustrates the efficiency of the proposed algorithm on the information extraction and fault detection.
How to cite this article
Shijin Ren, Ping Ling, Maoyun Yang, Yinlong Ni and Zhihuan Song, 2013. Multi-Kernel PCA with Discriminant Manifold for Hoist Monitoring. Journal of Applied Sciences, 13: 4195-4200.
REFERENCES
Ghoraani, B. and S. Krishnan, 2011. Time-frequency matrix feature extraction and classification of environmental audio signals. IEEE Trans. Audio Speech Language Process., 19: 2197-2210.
CrossRef
He, Q., 2013. Time-frequency manifold for nonlinear feature extraction in machinery fault diagnosis. Mech. Syst. Signal Process., 35: 200-218.
CrossRef Direct Link
Jiang, Q., M. Jia, J. Hu and F. Xu, 2009. Machinery fault diagnosis using supervised manifold learning. Mech. Syst. Signal Process., 23: 2301-2311.
CrossRef Direct Link
Lin, Y.Y., T.L. Liu and C.S. Fuh, 2011. Multiple kernel learning for dimensionality reduction. IEEE Trans. Pattern Anal. Machine Intell., 33: 1147-1160.
CrossRef Direct Link
Liu, L., W. Zuo, D. Zhang, N. Li and H. Zhang, 2012. Combination of heterogeneous features for wrist pulse blood flow signal diagnosis via multiple kernel learning. IEEE Trans. Inform. Technol. Biomedicine, 16: 598-606.
CrossRef Direct Link
Liu, X.P., G.Y. Xu and S.J. Ren, 2012. Research on hoist machinery fault monitoring based on HHT-DDKICA and SVDD method. Chin. J. Machinery Eng., 39: 21-32.
Wang, G.B., 2010. Manifold learning based fault diagnosis for rotating machine. Ph.D. Thesis, Central South University.
Wang, S., W. Huang and Z.K. Zhu, 2011. Transient modeling and parameter identification based on wavelet and correlation filtering for rotating machine fault diagnosis. Mech. Syst. Signal Process., 25: 1299-1320.
CrossRef Direct Link
Wang, Z., S.C. Chen and T. Sun, 2008. MultiK-MHKS: A novel multiple kernel learning algorithm. IEEE Trans. Pattern Anal. Mach. Intell., 30: 348-353.
CrossRef Direct Link
Xie, X. and H. Shi, 2012. Dynamic multimode process modeling and monitoring using adaptive gaussian mixture models. Ind. Eng. Chem. Res., 51: 5497-5505.
CrossRef Direct Link
Yu, J., 2012. Semiconductor manufacturing process monitoring using gaussian mixture model and bayesian method with local and nonlocal information. IEEE Trans. Semiconduct. Manuf., 25: 480-493.
CrossRef Direct Link
Zhang, M.G. and Z.H. Song, 2009. LPMVP algorithm and its application to fault monitoring. ACTA Automatic Sinica, 35: 766-771.
Zhu, W., X.F. Hu and Z. Zhen, 2009. MPCA-AR-based Fault predictionalgorithm. Chin. J. Device Instrument, 3: 1778-1782.
Lei, Y., Z. He and Y. Zi, 2011. EEMD method and WNN for fault diagnosis of locomotive roller bearings. Expert Syst. Appl., 38: 7334-7341.
CrossRef Direct Link
© Science Alert. All Rights Reserved