Constrained neural network training and its application to hyperelastic material modeling

Neural networks (NN) have been studied and used widely in the field of computational mechanics, especially to approximate material behavior. One of their disadvantages is the large amount of data needed for the training process. In this paper, a new approach to enhance NN training with physical knowledge using constraint optimization techniques is presented. Specific constraints for hyperelastic materials are introduced, which include energy conservation, normalization and material symmetries. We show, that the introduced enhancements lead to better learning behavior with respect to well known issues like a small number of training samples or noisy data. The NN is used as a material law within a finite element analysis and its convergence behavior is discussed with regard to the newly introduced training enhancements. The feasibility of NNs trained with physical constraints is shown for data based on real world experiments. We show, that the enhanced training outperforms state-of-the-art techniques with respect to stability and convergence behavior within FE simulations.

[1]  Paul J. Werbos,et al.  Applications of advances in nonlinear sensitivity analysis , 1982 .

[2]  L. Treloar.  Stress-Strain Data for Vulcanized Rubber under Various Types of Deformation , 1944 .

[3]  M. Lefik,et al.  Artificial neural network as an incremental non-linear constitutive model for a finite element code , 2003 .

[4]  R. Rivlin.  Large elastic deformations of isotropic materials IV. further developments of the general theory , 1948, Philosophical Transactions of the Royal Society of London. Series A, Mathematical and Physical Sciences.

[5]  Yong Wang,et al.  A generalized-constraint neural network model: Associating partially known relationships for nonlinear regressions , 2009, Inf. Sci..

[6]  G Pande,et al.  Enhancement of data for training neural network based constitutive models for geomaterials , 2002 .

[7]  D. Moreira,et al.  Comparison of simple and pure shear for an incompressible isotropic hyperelastic material under large deformation , 2013 .

[8]  Carsten Könke,et al.  Neural networks as material models within a multiscale approach , 2009 .

[9]  George Cybenko,et al.  Approximation by superpositions of a sigmoidal function , 1989, Math. Control. Signals Syst..

[10]  Youssef M A Hashash,et al.  Numerical implementation of a neural network based material model in finite element analysis , 2004 .

[11]  A. Ibrahimbegovic.  Nonlinear Solid Mechanics , 2009 .

[12]  H. Parisch,et al.  Efficient non‐linear finite element shell formulation involving large strains , 1986 .

[13]  Peter Wriggers,et al.  A simple method for the calculation of postcritical branches , 1988 .

[14]  L. Treloar,et al.  The properties of rubber in pure homogeneous strain , 1975 .

[15]  P. Wolfe.  Convergence Conditions for Ascent Methods. II , 1969 .

[16]  Klaus-Robert Müller,et al.  Efficient BackProp , 2012, Neural Networks: Tricks of the Trade.

[17]  Werner Wagner,et al.  A robust non‐linear mixed hybrid quadrilateral shell element , 2005 .

[18]  Patrick van der Smagt.  Minimisation methods for training feedforward neural networks , 1994, Neural Networks.

[19]  Wolfgang Graf,et al.  A material description based on recurrent neural networks for fuzzy data and its application within the finite element method , 2013 .

[20]  R. Ogden.  Large deformation isotropic elasticity – on the correlation of theory and experiment for incompressible rubberlike solids , 1972, Proceedings of the Royal Society of London. A. Mathematical and Physical Sciences.

[21]  James H. Garrett,et al.  Knowledge-Based Modeling of Material Behavior with Neural Networks , 1992 .

[22]  Bernhard A. Schrefler,et al.  Artificial Neural Networks in numerical modelling of composites , 2009 .

[23]  P. Wolfe.  Convergence Conditions for Ascent Methods. II: Some Corrections , 1971 .

[24]  Philip E. Gill,et al.  Practical optimization , 1981 .

[25]  Geoffrey E. Hinton,et al.  Learning internal representations by error propagation , 1986 .

[26]  Sven Klinkel,et al.  A continuum based three-dimensional shell element for laminated structures , 1999 .

[27]  Christopher M. Bishop,et al.  Curvature-driven smoothing: a learning algorithm for feedforward networks , 1993, IEEE Trans. Neural Networks.

[28]  Yaser S. Abu-Mostafa,et al.  Learning from hints in neural networks , 1990, J. Complex..

[29]  P. Wriggers.  Nonlinear Finite Element Methods , 2008 .

[30]  Ridha Hambli,et al.  Multiscale methodology for bone remodelling simulation using coupled finite element and neural network computation , 2011, Biomechanics and modeling in mechanobiology.

[31]  Zeliang Liu,et al.  A Deep Material Network for Multiscale Topology Learning and Accelerated Nonlinear Modeling of Heterogeneous Materials , 2018, Computer Methods in Applied Mechanics and Engineering.

[32]  Wei Chen,et al.  A framework for data-driven analysis of materials under uncertainty: Countering the curse of dimensionality , 2017 .

[33]  Paul Steinmann,et al.  Hyperelastic models for rubber-like materials: consistent tangent operators and suitability for Treloar’s data , 2012 .

[34]  Christian Kanzow,et al.  Theorie und Numerik restringierter Optimierungsaufgaben , 2002 .

[35]  D K Smith,et al.  Numerical Optimization , 2001, J. Oper. Res. Soc..

[36]  William L. Ko,et al.  Application of Finite Elastic Theory to the Deformation of Rubbery Materials , 1962 .

[37]  R. Ogden.  Non-Linear Elastic Deformations , 1984 .

[38]  Raimund Rolfes,et al.  Neural network assisted multiscale analysis for the elastic properties prediction of 3D braided composites under uncertainty , 2018 .

[39]  Pascal Fua,et al.  Imposing Hard Constraints on Deep Networks: Promises and Limitations , 2017, CVPR 2017.

[40]  Sven Klinkel,et al.  A mixed shell formulation accounting for thickness strains and finite strain 3d material models , 2008 .

[41]  F. Gruttmann,et al.  An advanced shell model for the analysis of geometrical and material nonlinear shells , 2020, Computational Mechanics.

[42]  Kun Wang,et al.  A multiscale multi-permeability poroplasticity model linked by recursive homogenizations and deep learning , 2018, Computer Methods in Applied Mechanics and Engineering.

[43]  C. Horgan,et al.  Simple Shearing of Incompressible and Slightly Compressible Isotropic Nonlinearly Elastic Materials , 2010 .

[44]  Geoffrey E. Hinton,et al.  Deep Learning , 2015, Nature.