Combination of ICA and SOM for Classification of Machine Condition Patterns

Abstract:

Nonlinear independent component analysis (NICA) is a powerful method for analyzing nonlinear and nongaussian data. Artificial neural network (ANN), especially self-organizing map (SOM) based on unsupervised learning, is an excellent tool for pattern clustering and recognition. A novel multi-NICA network is proposed for feature extraction of different mechanical patterns, followed by a typical ANN that is one of Multi-Layer Perceptron (MLP), or Radial Basis Function Network (RBFN), or self-organizing map (SOM), which implements the final classification. Using NICA and appropriate strategies for further feature extraction, nonlinear and higher than second order features embedded in multi-channel vibration measurements can be captured effectively. Mechanical fault patterns can be recognized correctly. Results from the contrast classification experiments show that the new compound ICA-SOM classifier can be constructed in a simpler way and it can classify various fault patterns with high accuracy, both of which imply a great potential in health condition monitoring of machine systems.

• Periodical: Key Engineering Materials (Volumes 295-296)
• Pages: 643-648
• DOI: https://doi.org/10.4028/www.scientific.net/KEM.295-296.643
• Online since: October 2005
• Authors: Shu Xiang Yang, W.D. Jiao, Z.T. Wu
• Keywords: Artificial Neural Network (ANN), MLP, Nonlinear ICA, RBFN, Residual Mutual Information, SOM

References

[1] A. Ypma, D.M.J. Tax and R.P.W. Duin: Proc IEEE International Workshop on Neural Networks for Signal Processing, Wisconsin, USA, 1999, p.67.

[2] P. Comon: Signal Processing, Vol. 36 (1994), p.287.

[3] S. Chitroub, A. Houacine and B. Sansal: Proc Image Processing 2000 International Conference, Vancouver, Canada, Vol. 3 (2000), p.328.

[4] G. Gelle, M. Colas and G. Delaunay: Mechanical Systems and Signal Processing, Vol. 14 (2000), p.427.

[5] A. Bell and T. Sejnowski: Neural Computation, Vol. 7 (1995), p.1129.

[6] J.F. Cardoso and A. Souloumiac: IEE Proceedings F, Vol. 140 (1993), p.256.

[7] L. Jiao: Application and Implementation of Neural Network (Xi'an Electron Science & Technology University, Xi'an 1996).

[8] A. Bell and T. Sejnowski: Neural Computation, Vol. 7 (1995), p.1129.

[9] J.F. Cardoso and A. Souloumiac: IEE Proceedings F, Vol. 140 (1993), p.256.

[10] A. Hyvärinen: Neural Computing Surveys, Vol. 2 (1999), p.94.

[11] R. Tinos and M.H. Terra: Proc 5 th Brazilian Symposium on Neural Network, 1998, Belo Horizonte, Brazil.

[12] L.B. Almeida: Proc. of the 2001 Int. Joint Conf. on Neural Networks, Washington, D.C., 2001. Normal Gear pitting fault Segments Classifiers Amount of test segments Correctly classified Segments Mistaken Classified Segments Classification Accuracy (%) Amount of test segments Correctly classified Segments Mistaken Classified Segments Classification Accuracy (%) MLP 32 30 2 93. 75% 32 26 6 81. 25% RBFN 32 30 2 93. 75% 32 28 4 87. 50% SOM 32 29 3 90. 63% 32 26 6 81. 25.

DOI: 10.3109/9780203213810-2