Influence of Grey System Parameter Identification Method on Prediction of Bearing Capacity of Piles

Abstract:

An unequal interval grey model GM (1,1) was established according to the variation characteristics of sequence data about the bearing capacity of overlength piles, and a difference equations was usually adopted to replace the grey differential equation for determining the system parameters of the model; however, great errors would occur when the model thus established was used to predict the bearing capacity of overlength piles, that is, the prediction results would be overestimated or underestimated. In order to improve the prediction accuracy of the model, we established an error objective function based on optimization theory in this study, employed the method of nonlinear least squares to identify the system parameters in the grey differential equation for the bearing capacity of overlength piles, and built an optimization-based grey optimization model. The model system built with optimization method was used to predict the bearing capacity of overlength piles, and the predicted values fit the test values well. In addition, the model system had a higher accuracy, compared with the grey difference model built with difference method; therefore, the model system built with optimization method could provide reference for prediction of the bearing capacity of overlength piles.

• Periodical: Advanced Materials Research (Volumes 243-249)
• Pages: 2044-2049
• DOI: https://doi.org/10.4028/www.scientific.net/AMR.243-249.2044
• Online since: May 2011
• Authors: Fang Yang, Lin Zhu Sun, Zi Ling Xie, Ya Gang Zhou
• Keywords: Bearing Capacity of Overlength Pile, Difference Method, Model GM (1,1), Optimization Method, Prediction

References

[1] Deng, Z.Y., Lu, P.Y. Comparative Analysis of Several Models for Predicting Vertical Ultimate Bearing Capacity of Single Pile [J]. Rock and Soil Mechanics, 2002, 23(4): 428-431.

[2] Lee, I.M., Lee, J,H. Prediction of Pile Bearing Capacity Using Artificial Neural Networks [J]. Computers and Geotechnics, 1996, 18(3): 189-200.

DOI: 10.1016/0266-352x(95)00027-8

[3] Dong, Q.H. Determination and Prediction of Ultimate Load of Foundation Piles [J]. Rock and Soil Mechanics, 2001, 22(4): 463-466.

[4] Zhang, S.G., Han, B., Xu, Y., et al. Study of Methods of Predicting Ultimate Bearing Capacity of Pile Foundation [J]. Journal of Hohai University, 2001, 29: 72-74.

[5] Liu, J.L. Discussion on Hyperbolic Prediction of Ultimate Bearing Capacity of Single Pile [J]. Geotechnical Engineering Technique, 2001, (4): 204-207.

[6] Zhao, M.H., Hu, Z.Q. Adjusted Hyperbolic Method for Predicting Bearing Capacity of Trial Pile [J]. Journal of Building Structures, 1995, (3): 47-52.

[7] Wang, X.Q., Gao, W.H., Yang, L.D. Neural Network Prediction of Ultimate Bearing Capacity and Settlement of Pile Foundation [J]. Building Science, 2003, 19(1): 58-60.

[8] Liu, L.M., Shu, X., Xiong, J.H. Theory Development and Practice of Pile Foundation Engineering [M]. Beijing: China Building Materials Industry Press, 2002.

[9] Lin, Y.L. Applied Stochastic Processes [M]. Beijing: Tsinghua University Press, 2002. 78-103.

[10] Li, D., Su, X.H., Ma, S.Y. Stock Price Prediction Algorithms based on Novel Grey Markov Model [J]. Journal of Harbin Institute of Technology, 2003, 35(2): 244-248