Javadi, A.A., Johari, A., Ahangar-Asr, A., Faramarzi, A. and Toll, D.G. (2009) Using evolutionary polynomial regression in modelling of unsaturated soils. In: Proceedings of the 17th UK Conference on Computational Mechanics. Spencer Institute of Theoretical and Computational Mechanics, Unversity of Nottingham, Nottingham, UK, pp. 353-356. ISBN 9780853582557

Full text not available from this repository.

Abstract

Investigation of unsaturated soil behaviour is a very wide, complicated and demanding area of research in geotechnical engineering. A number of constitutive models have been developed to describe the complex behaviour of unsaturated soils. Despite the complexity of the existing constitutive theories, none of the existing constitutive models can describe various aspects of behaviour of unsaturated soils under different stress paths and loading conditions. In this paper a new approach is presented, based on evolutionary polynomial regression (EPR), for modelling of unsaturated soil behaviour. EPR is an evolutionary data mining technique that generates a transparent and structured representation of the behaviour of a system directly from data. Like neural network, this method can operate on large quantities of data in order to capture nonlinear and complex interactions between variables of the system. It has the additional advantage that it allows the user to gain insight into the behaviour of the system. The capabilities of the EPR methodology are illustrated by application to modelling of the behaviour of unsaturated soils. EPR models are developed and validated using results from a comprehensive set of triaxial tests on samples of compacted unsaturated soils. The developed models are used to predict different aspects of behaviour of unsaturated soil for conditions not used in the model building process. The results show that the proposed approach is very effective and robust in modelling the behaviour of unsaturated soils. It provides a unified framework for constitutive modelling of unsaturated soils. It is also shown that a suitably trained EPR is able to predict entire stress paths of a triaxial test incrementally (point-by-point) with a high degree of accuracy. The merits and advantages of the proposed approach are highlighted.

Item Type:	Conference Proceedings
Title of Proceedings:	Proceedings of the 17th UK Conference on Computational Mechanics
Additional Information:	[1] This paper was first presented at the 17th UK Conference on Computational Mechanics (ACME-UK), held from 4-6 April 2009 in Nottingham, UK.
Uncontrolled Keywords:	unsaturated soil, constitutive modelling, data mining, evolutionary computation
Subjects:	T Technology > T Technology (General)
Pre-2014 Departments:	School of Engineering School of Engineering > Department of Civil Engineering
Related URLs:	Organisation Organisation Organisation
Last Modified:	14 Oct 2016 09:26
URI:	http://gala.gre.ac.uk/id/eprint/11155

Actions (login required)

View Item