Affiliation:
1. Laboratoire Génie de Production, Institut National Polytechnique/Ecole Nationale d’Ingénieurs de Tarbes, Université de Toulouse, 47 Av d’Azereix, F-65016 Tarbes, France
Abstract
Finite element (FE) simulations have been effective in simulating thermomechanical forming processes, yet challenges arise when applying them to new materials due to nonlinear behaviors. To address this, machine learning techniques and artificial neural networks play an increasingly vital role in developing complex models. This paper presents an innovative approach to parameter identification in flow laws, utilizing an artificial neural network that learns directly from test data and automatically generates a Fortran subroutine for the Abaqus standard or explicit FE codes. We investigate the impact of activation functions on prediction and computational efficiency by comparing Sigmoid, Tanh, ReLU, Swish, Softplus, and the less common Exponential function. Despite its infrequent use, the Exponential function demonstrates noteworthy performance and reduced computation times. Model validation involves comparing predictive capabilities with experimental data from compression tests, and numerical simulations confirm the numerical implementation in the Abaqus explicit FE code.
Subject
Computational Mathematics,Computational Theory and Mathematics,Numerical Analysis,Theoretical Computer Science
Reference37 articles.
1. Abaqus (1989). Reference Manual, Hibbitt, Karlsson and Sorensen Inc.
2. Modeling of flow stress of 42CrMo steel under hot compression;Lin;Mater. Sci. Eng. A,2009
3. A critical analysis of plastic flow behaviour in axisymmetric isothermal and Gleeble compression testing;Bennett;Comput. Mater. Sci.,2010
4. Thermo-mechanical simulation using gleeble system-advantages and limitations;Kumar;J. Metall. Mater. Sci.,2016
5. Refining constitutive relation by integration of finite element simulations and Gleeble experiments;Yu;J. Mater. Sci. Technol.,2019