Premium
Numerical modeling and neural networks to identify model parameters from piezocone tests: I. FEM analysis of penetration in two‐phase continuum
Author(s) -
Obrzud Rafał F.,
Truty Andrzej,
Vulliet Laurent
Publication year - 2011
Publication title -
international journal for numerical and analytical methods in geomechanics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.419
H-Index - 91
eISSN - 1096-9853
pISSN - 0363-9061
DOI - 10.1002/nag.972
Subject(s) - finite element method , geotechnical engineering , inverse problem , boundary value problem , inverse , numerical analysis , numerical modeling , test data , population , numerical models , engineering , structural engineering , computer simulation , mathematics , simulation , mathematical analysis , statistics , geometry , software engineering , demography , sociology
This study presents a numerical approach designed for material parameter identification for the coupled hydro‐mechanical boundary value problem (BVP) of the piezocone test (CPTU) in normally and lightly overconsolidated clayey soils. The study is presented in two related papers and it explores the possibility of using neural networks (NNs) to solve the complex inverse problem of the penetration test, including partially drained conditions. It has been demonstrated that the development of NN‐based inverse models can be based on training data sets that consist of pseudo‐experimental measurements derived from numerical simulations of the piezocone test. The first paper presents the development of the FE model of the studied problem, which can be used to generate a training data population corresponding to typical piezocone measurements that are obtained for clayey soils. The paper contains a detailed description of the numerical model with a sensitivity analysis with respect to different model parameters including the effect of partial drainage. The analysis also includes the model verification by means of a comparative analysis with numerical models of penetration proposed in the literature, as well as experimental evidence. Finally, owing to the loss of accuracy observed when applying a ‘rough’ frictional interface in the Updated Lagrangian formulation, an equivalent semi‐numerical model for the piezocone test is proposed, taking into account a possible occurrence of partial drainage during penetration. Copyright © 2010 John Wiley & Sons, Ltd.