Computational Technology Resources - CCC

T. Janda¹, M. Šejnoha¹, A. Zemanová² and T. Žalská²

¹Department of Mechanics, Faculty of Civil Engineering, Czech Technical University in Prague, Prague, Czechia
²Department of Geotechnics, Faculty of Civil Engineering, Czech Technical University in Prague, Prague, Czechia

doi:10.4203/ccc.8.9.3

T. Janda, M. Šejnoha, A. Zemanová, T. Žalská, "Automatic Differentiation in PyTorch as a Tool for Robust Implementation of Elasto-Plastic Constitutive Model", in P. Iványi, J. Kruis, B.H.V. Topping, (Editors), "Proceedings of the Twelfth International Conference on Engineering Computational Technology", Civil-Comp Press, Edinburgh, UK, Online volume: CCC 8, Paper 9.3, 2024, doi:10.4203/ccc.8.9.3

Keywords: elasto-plasticity, hardening soil model, stress return mapping, automatic differentiation, PyTorch, finite element method.

Abstract

The paper presents a simple and robust approach to an implementation of the hardening soil model into finite element calculations. The implementation of the return stress mapping exploits the automatic differentiation of tensor variables provided by the PyTorch framework. The automatic differentiation allows for a succinct implementation despite the relatively complex structure of the nonlinear equations in the stress return algorithm. The presented approach is not limited to the hardening soil model. It can be utilised in the development and verification of other elasto-plastic constitutive models where expressing and maintaining the Jacobian matrix over different versions of a material model is time-consuming and error-prone.

download the full-text of this paper (PDF, 12 pages, 467 Kb)

go to the previous paper
return to the table of contents
return to the volume description

	Computational & Technology Resources an online resource for computational, engineering & technology publications
	not logged in - login
Front Page Browse CCC CCP CSETS CTR IJRT Other Authors Search Purchase Guide FAQ Contact us	Civil-Comp Conferences ISSN 2753-3239 CCC: 8 PROCEEDINGS OF THE TWELFTH INTERNATIONAL CONFERENCE ON ENGINEERING COMPUTATIONAL TECHNOLOGY Edited by: P. Iványi, J. Kruis and B.H.V. Topping Paper 9.3 Automatic Differentiation in PyTorch as a Tool for Robust Implementation of Elasto-Plastic Constitutive Model T. Janda¹, M. Šejnoha¹, A. Zemanová² and T. Žalská² ¹Department of Mechanics, Faculty of Civil Engineering, Czech Technical University in Prague, Prague, Czechia ²Department of Geotechnics, Faculty of Civil Engineering, Czech Technical University in Prague, Prague, Czechia doi:10.4203/ccc.8.9.3 Full Bibliographic Reference for this paper T. Janda, M. Šejnoha, A. Zemanová, T. Žalská, "Automatic Differentiation in PyTorch as a Tool for Robust Implementation of Elasto-Plastic Constitutive Model", in P. Iványi, J. Kruis, B.H.V. Topping, (Editors), "Proceedings of the Twelfth International Conference on Engineering Computational Technology", Civil-Comp Press, Edinburgh, UK, Online volume: CCC 8, Paper 9.3, 2024, doi:10.4203/ccc.8.9.3 Keywords: elasto-plasticity, hardening soil model, stress return mapping, automatic differentiation, PyTorch, finite element method. Abstract The paper presents a simple and robust approach to an implementation of the hardening soil model into finite element calculations. The implementation of the return stress mapping exploits the automatic differentiation of tensor variables provided by the PyTorch framework. The automatic differentiation allows for a succinct implementation despite the relatively complex structure of the nonlinear equations in the stress return algorithm. The presented approach is not limited to the hardening soil model. It can be utilised in the development and verification of other elasto-plastic constitutive models where expressing and maintaining the Jacobian matrix over different versions of a material model is time-consuming and error-prone. download the full-text of this paper (PDF, 12 pages, 467 Kb) go to the previous paper return to the table of contents return to the volume description
Back to top	©Civil-Comp Limited 2023 - terms & conditions