Lateral load bearing capacity modelling of piles in cohesive soils in undrained conditions; an intelligent evolutionary approach

Ahangar Asr, A; Javadi, AA; Johari, A; Chen, Y

Abstract

The complex behaviour of fine-grained materials in relation with structural elements has received noticeable attention from geotechnical engineers and designers in recent decades. In this research work an evolutionary approach is presented to create a structured polynomial model for predicting the undrained lateral load bearing capacity of piles. The proposed evolutionary polynomial regression (EPR) technique is an evolutionary data mining methodology that generates a transparent and structured representation of the behaviour of a system directly from raw data. It can operate on large quantities of data in order to capture nonlinear and complex relationships between contributing variables. The developed model allows the user to gain a clear insight into the behaviour of the system. Field measurement data from literature was used to develop the proposed EPR model. Comparison of the proposed model predictions with the results from two empirical models currently being implemented in design works, a neural network-based model from literature and also the field data shows that the EPR model is capable of capturing, predicting and generalising predictions to unseen data cases, for lateral load bearing capacity of piles with very high accuracy. A sensitivity analysis was conducted to evaluate the effect of individual contributing parameters and their contribution to the predictions made by the proposed model. The merits and advantages of the proposed methodology are also discussed.

Citation

Ahangar Asr, A., Javadi, A., Johari, A., & Chen, Y. (2014). Lateral load bearing capacity modelling of piles in cohesive soils in undrained conditions; an intelligent evolutionary approach. Applied Soft Computing, 24, 822-828. https://doi.org/10.1016/j.asoc.2014.07.027

• Journal Article Type: Article
• Acceptance Date: Jul 9, 2014
• Online Publication Date: Sep 8, 2014
• Publication Date: Nov 1, 2014
• Deposit Date: Dec 15, 2015
• Publicly Available Date: Jul 9, 2016
• Journal: Applied Soft Computing
• Print ISSN: 1568-4946
• Publisher: Elsevier
• Volume: 24
• Pages: 822-828
• DOI: https://doi.org/10.1016/j.asoc.2014.07.027
• Publisher URL: http://dx.doi.org/10.1016/j.asoc.2014.07.027
• Related Public URLs: http://www.journals.elsevier.com/applied-soft-computing/