Separation of Bridge Deflection Signals Based on ICA

Abstract:

In order to reject environmental factor effect from long-span bridge deflection signal, this document introduced the basic theory of independent component analysis (ICA) and applied FastICA algorithm to separate background engineering simulation signal, which was based on using vector-dimension-augmenting technique to process the signal after adopting fir lowpass filter to rejecte the high frequency component, and finally relized the separations of live load deflection, temperature difference deflection and long-term deflection. Campared separation results with deflection source signal, it was shown that the separation result was good when the correlation coefficients between separation signal and source signal were larger than 0.8.

• Periodical: Advanced Materials Research (Volumes 374-377)
• Pages: 2090-2095
• DOI: https://doi.org/10.4028/www.scientific.net/AMR.374-377.2090
• Online since: October 2011
• Authors: Chun Hui Tang, Hong Yang, Xia Ping Liu, Zhuo Sun
• Keywords: Bridge, Deflection, Filtering, ICA

References

[1] Liu Gang: Condition Assessment Research to Large-span Bridges Based on Long-term Static Monitoring Data[D]. Chongqing, Chongqing University, 2010. (in Chinese)

[2] Chen De-wei, Jing Guo-qiang, Huang Zheng. Prediction of bridge structural performance under temperature by artificial neural net work[J]. Structural Engineers, 2006, 04(5): 24-28. (in Chinese)

[3] A. Deraemaeker, E. Reynders, G. De Roeck, J. Kullaa. Vibration-based structural health monitoring using output-only measurements under changing environment [J]. Mechanical Systems and Signal Processing. 2008, 22:34-56.

DOI: 10.1016/j.ymssp.2007.07.004

[4] Liang Zongbao. Research on structural safety assessment of bridges based on statistical analysis of monitoring information [D].Chongqing: Chongqing University,2006. (in Chinese)

[5] WEI L, RAJAPAKSE J C. Approach and Applications of Constrained ICA[J]. IEEE Transaction on NeuralNetworks, 2005, 16 (1): 203-212

[6] Koldovsky Z, Tichavsky P,Oja E. Efficient Variant of Algorithm FastICA for Independent Component Analysis Attaining the CRLB[J]. IEEE Transactions on Neural Networks, 2006, 17 (5): 1265-1277.

DOI: 10.1109/tnn.2006.875991

[7] Wang Ronghui, Chi Chun, Chen Qingzhong, Zhen Xiaoxia. Study on the Model of the Fatigue-loaded Vehicles in Guangzhou Trestle Bridges[J]. Journal of South China University of Technology(Natural Science), 2004, 12(12): 94-96. (in Chinese)