Skip navigation

EPR-based material modelling of soils

EPR-based material modelling of soils

Faramarzi, Asaad and Alani, Amir M. (2013) EPR-based material modelling of soils. In: European Geosciences Union General Assembly, EGU2013, 7-12 Apr 2013, Vienna, Austria.

Full text not available from this repository.

Abstract

In the past few decades, as a result of the rapid developments in computational software and hardware, alternative computer aided pattern recognition approaches have been introduced to modelling many engineering problems, including constitutive modelling of materials. The main idea behind pattern recognition systems is that they learn adaptively from experience and extract various discriminants, each appropriate for its purpose. In this work an approach is presented for developing material models for soils based on evolutionary polynomial regression (EPR). EPR is a recently developed hybrid data mining technique that searches for structured mathematical equations (representing the behaviour of a system) using genetic algorithm and the least squares method. Stress–strain data from triaxial tests are used to train and develop EPR-based material models for soil. The developed models are compared with some of the well-known conventional material models and it is shown that EPR-based models can provide a better prediction for the behaviour of soils. The main benefits of using EPR-based material models are that it provides a unified approach to constitutive modelling of all materials (i.e. all aspects of material behaviour can be implemented within a unified environment of an EPR model); it does not require any arbitrary choice of constitutive (mathematical) models. In EPR-based material models there are no material parameters to be identified. As the model is trained directly from experimental data therefore, EPR-based material models are the shortest route from experimental research (data) to numerical modelling. Another advantage of EPR-based constitutive model is that as more experimental data become available, the quality of the EPR prediction can be improved by learning from the additional data, and therefore, the EPR model can become more effective and robust. The developed EPR-based material models can be incorporated in finite element (FE) analysis.

Item Type: Conference or Conference Paper (Paper)
Additional Information: [1] This paper was first presented at the European Geosciences Union General Assembly 2013 (EGU2013) held from 7-12 April 2013 in Vienna, Austria. It was given on 12 April 2013 within the Open Session on Advances in Data, Methods, Models and Their Applications in Geosciences, No: EGU2013-4904. [2] An abstract for this paper is also available at: Geophysical Research Abstracts, Vol. 15, EGU2013-4904, 2013.
Uncontrolled Keywords: evolutionary computing, EPR, soil, material modelling
Subjects: T Technology > TD Environmental technology. Sanitary engineering
Pre-2014 Departments: School of Engineering
School of Engineering > Department of Civil Engineering
Related URLs:
Last Modified: 14 Oct 2016 09:26
Selected for GREAT 2016: None
Selected for GREAT 2017: None
Selected for GREAT 2018: None
URI: http://gala.gre.ac.uk/id/eprint/11132

Actions (login required)

View Item View Item