机械工程学报

JIANG Quansheng
School of Mechanical Engineering,
Southeast University,
Nanjing 211189, China

Department of Physics,
Chaohu University,
Chaohu 238000, China

JIA Minping

HU Jianzhong

XU Feiyun
School of Mechanical Engineering,
Southeast University,
Nanjing 211189, China

MODIFIED LAPLACIAN EIGENMAP METHOD FOR FAULT DIAGNOSIS*

Abstract: A novel method based on the improved Laplacian eigenmap algorithm for fault pattern classification is proposed. Via modifying the Laplacian eigenmap algorithm to replace Euclidean distance with kernel-based geometric distance in the neighbor graph construction, the method can preserve the consistency of local neighbor information and effectively extract the low-dimensional manifold features embedded in the high-dimensional nonlinear data sets. A nonlinear dimensionality reduction algorithm based on the improved Laplacian eigenmap is to directly learn high-dimensional fault signals and extract the intrinsic manifold features from them. The method greatly preserves the global geometry structure information embedded in the signals, and obviously improves the classification performance of fault pattern recognition. The experimental results on both simulation and engineering indicate the feasibility and effectiveness of the new method.

Key words: Laplacian eigenmap Kernel trick Fault diagnosis Manifold learning

* This project is supported by National Hi-tech Research and Development Program of China (863 Program, No. 2007AA04Z421) and National Natural Science Foundation of China (No. 50475078, No. 50775035). Received August 22, 2007; received in revised form February 27, 2008; accepted March 3, 2008

浏览（下载）论文全文（PDF格式）

Biographical notes

JIANG Quansheng is currently a PhD candidate in School of Mechanical Engineering, Southeast University, China. His research interests include fault diagnosis, manifold learning, artificial intelligence (AI), etc.
Tel: +86-13913879698; E-mail: jqs1996@163.com

JIA Minping is a professor and doctoral supervisor in School of Mechanical Engineering, Southeast University, China. His current research interests include signals detection and processing, fault diagnosis, artificial intelligence (AI), etc.
Tel: +86-25-52090512; E-mail: mpjia@seu.edu.cn

HU Jianzhong is a PhD in School of Mechanical Engineering, Southeast University, China. His research interests include fault diagnosis, artificial intelligence (AI), manifold learning, etc.
Tel: +86-25-52090512; E-mail: hjz@seu.edu.cn

XU Feiyun is a professor in School of Mechanical Engineering, Southeast University, China. His current research interests include fault diagnosis, signal processing, artificial intelligence (AI), etc.
Tel: +86-25-52090512; E-mail: fyxu@seu.edu.cn

References

[1] ZHONG Binglin, HUANG Ren, JIA Minping. Mechanical fault diagnostics theory[M]. 3rd edition. Beijing: China Machine Press, 2007. (in Chinese)
[2] COX T F, COX M A. Multi-dimensional scaling[M]. London: Chapman and Hall, 1994.
[3] KOHONEN T. Self-organization and associative memory[M]. New York: Springer-Verlag, 1988.
[4] SEUNG H S , DANIEL D L. The manifold ways of perception[J]. Science, 2000 , 290 (5 500): 2 268-2 269.
[5] ROWEIS S, SAUL L. Nonlinear dimensionality reduction by locally linear embedding[J]. Science, 2000, 290(5 500): 2 323-2 326.
[6] TENENBAUM J, SILVA D D, LANGFORD J. A global geometric framework for nonlinear dimensionality reduction[J]. Science, 2000, 290(5 500): 2 319-2 323.
[7] SAMUEL K, MARTIN D L. Face detection in gray scale images using locally linear embeddings[J]. Computer Vision and Image Understanding, 2007, 105 (1): 1-20.
[8] VLACHOS M, DOMENICONI C, GUNOPULOS D, et al. Non-linear dimensionality reduction techniques for classification and visualization[C]// Proceedings of the 8th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, Edmonton. NY: ACM Press, 2002: 645-651.
[9] ZHANG Junping, LI Stan Z, WANG Jue. Manifold learning and applications in recognition[C]// Intelligent Multimedia Processing with Soft Computing. TAN Yap Peng, YAP Kim Hui, WANG Lipo, ed. Heidelberg: Springer-Verlag, 2005: 281-300.
[10] YANG Jianhong, XU Jinwu, YANG Debin, et al. Noise reduction method for nonlinear time series based on principal manifold learning and its application to fault diagnosis[J]. Chinese Journal of Mechanical Engineering, 2006, 42(8): 154-158. (in Chinese)
[11] BELKIN M, NIYOGI P. Laplacian eigenmaps for dimensionality reduction and data representation[J]. Neural Computation, 2003, 15(6): 1 373-1 396.
[12] BELKIN M, NIYOGI P. Laplacian eigenmaps and spectral techniques for embedding and clustering[J]. Neural Information Processing Systems, 2001, 14: 585-591.