Lossless multi-scale constitutive elastic relations with artificial intelligence

The elastic properties of materials derive from their electronic and atomic nature. However, simulating bulk materials fully at these scales is not feasible, so that typically homogenized continuum descriptions are used instead. A seamless and lossless transition of the constitutive description of the elastic response of materials between these two scales has been so far elusive. Here we show how this problem can be overcome by using Artificial Intelligence (AI). A Convolutional Neural Network (CNN) model is trained, by taking the structure image of a nanoporous material as input and the corresponding elasticity tensor, calculated from Molecular Statics (MS), as output. Trained with the atomistic data, the CNN model captures the sizeand poredependency of the material’s elastic properties which, on the physics side, can stem from surfaces and non-local effects. Such effects are often ignored in upscaling from atomistic to classical continuum theory. To demonstrate the accuracy and the efficiency of the trained CNN model, a Finite Element Method (FEM) based result of an elastically deformed nanoporous beam equipped with the CNN as constitutive law is compared with that by a full atomistic simulation. The good agreement between the atomistic simulations and the FEM-AI combination for a system with size and surface effects establishes a new lossless scale bridging approach to such problems. The trained CNN model deviates from the atomistic result by 9.6% for porosity scenarios of up to 90% but it is about 230 times faster than the MS calculation and does not require to change simulation methods between different scales. The efficiency of the CNN evaluation together with the preservation of important atomistic effects makes the trained model an effective atomistically-informed constitutive model for macroscopic simulations of nanoporous materials, optimization of nano-structures, and solving of inverse problems. Keywords— Atomistic simulations, Molecular statics, Aluminium, Machine learning, Porosity, 1 ar X iv :2 10 8. 02 83 7v 1 [ co nd -m at .m tr lsc i] 5 A ug 2 02 1 Artificial intelligence, Anisotropic elasticity tensor.

[1]  Yung C. Shin,et al.  In-Process monitoring of porosity during laser additive manufacturing process , 2019, Additive Manufacturing.

[2]  Yoshitaka Adachi,et al.  Microstructure Recognition by Deep Learning , 2016 .

[3]  F. Pan,et al.  Topological representations of crystalline compounds for the machine-learning prediction of materials properties , 2021, npj Computational Materials.

[4]  Gabor Csanyi,et al.  Achieving DFT accuracy with a machine-learning interatomic potential: thermomechanics and defects in bcc ferromagnetic iron , 2017, 1706.10229.

[5]  Ramin Bostanabad,et al.  Train Once and Use Forever: Solving Boundary Value Problems in Unseen Domains with Pre-trained Deep Learning Models , 2021, ArXiv.

[6]  X. Chu,et al.  A plane stress model of bond-based Cosserat peridynamics and the effects of material parameters on crack patterns , 2021 .

[7]  Jimmy Ba,et al.  Adam: A Method for Stochastic Optimization , 2014, ICLR.

[8]  R. Lakes,et al.  Experimental Study of Elastic Constants of a Dense Foam with Weak Cosserat Coupling , 2019, Journal of Elasticity.

[9]  Yang Liu,et al.  Three-dimensional convolutional neural network (3D-CNN) for heterogeneous material homogenization , 2020, Computational Materials Science.

[10]  Shaohua Chen,et al.  Elastic Theory of Nanomaterials Based on Surface-Energy Density , 2014 .

[11]  A. McBride,et al.  A unified computational framework for bulk and surface elasticity theory: a curvilinear-coordinate-based finite element methodology , 2014 .

[12]  Stefanie Reese,et al.  Application of artificial neural networks for the prediction of interface mechanics: a study on grain boundary constitutive behavior , 2020, Adv. Model. Simul. Eng. Sci..

[13]  S. Nikolov,et al.  DAMASK – The Düsseldorf Advanced Material Simulation Kit for modeling multi-physics crystal plasticity, thermal, and damage phenomena from the single crystal up to the component scale , 2018, Computational Materials Science.

[14]  David A. Winkler,et al.  Machine learning property prediction for organic photovoltaic devices , 2020, npj Computational Materials.

[15]  Paris Perdikaris,et al.  Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations , 2019, J. Comput. Phys..

[16]  R. Lakes,et al.  Experimental Cosserat elasticity in open-cell polymer foam , 2016 .

[17]  Mgd Marc Geers,et al.  Multi-scale first-order and second-order computational homogenization of microstructures towards continua , 2003 .

[18]  Thomas Brox,et al.  U-Net: Convolutional Networks for Biomedical Image Segmentation , 2015, MICCAI.

[19]  Yuksel C. Yabansu,et al.  Material structure-property linkages using three-dimensional convolutional neural networks , 2018 .

[20]  D. Kondo,et al.  Effects of particles size on the overall strength of nanocomposites: Molecular Dynamics simulations and theoretical modeling , 2021 .

[21]  Kurt Kremer,et al.  Research Update: Computational materials discovery in soft matter , 2016 .

[22]  M. Marques,et al.  Recent advances and applications of machine learning in solid-state materials science , 2019, npj Computational Materials.

[23]  Olga Wodo,et al.  Microstructural informatics for accelerating the discovery of processing–microstructure–property relationships , 2016 .

[24]  B. Xu,et al.  Data-driven microstructure sensitivity study of fibrous paper materials , 2021 .

[25]  S. Reese,et al.  Direction-dependent fracture in solids: Atomistically calibrated phase-field and cohesive zone model , 2021 .

[26]  Samuel Forest,et al.  Generalized continua and non‐homogeneous boundary conditions in homogenisation methods , 2011 .

[27]  Sigrid Adriaenssens,et al.  A data-driven computational scheme for the nonlinear mechanical properties of cellular mechanical metamaterials under large deformation. , 2020, Soft matter.

[28]  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.

[29]  Timon Rabczuk,et al.  An Energy Approach to the Solution of Partial Differential Equations in Computational Mechanics via Machine Learning: Concepts, Implementation and Applications , 2019, Computer Methods in Applied Mechanics and Engineering.

[30]  M. Ashby,et al.  Effective properties of the octet-truss lattice material , 2001 .

[31]  Alexander L. Shluger,et al.  Roadmap on multiscale materials modeling , 2020, Modelling and Simulation in Materials Science and Engineering.

[32]  Yan Xu,et al.  Using machine learning and feature engineering to characterize limited material datasets of high-entropy alloys , 2020 .

[33]  B. A. Le,et al.  Computational homogenization of nonlinear elastic materials using neural networks , 2015 .

[34]  Elizabeth A. Holm,et al.  Characterizing powder materials using keypoint-based computer vision methods , 2017 .

[35]  M. Itskov,et al.  Mechanics of Nanostructured Porous Silica Aerogel Resulting from Molecular Dynamics Simulations. , 2017, The journal of physical chemistry. B.

[36]  T. Böhlke,et al.  Anisotropic hyperelastic constitutive models for finite deformations combining material theory and data-driven approaches with application to cubic lattice metamaterials , 2020, Computational Mechanics.

[37]  M. Sadd.  Chapter 11 – Anisotropic Elasticity , 2009 .

[38]  Michele Parrinello,et al.  Generalized neural-network representation of high-dimensional potential-energy surfaces. , 2007, Physical review letters.

[39]  Jure Leskovec,et al.  Learning to Simulate Complex Physics with Graph Networks , 2020, ICML.

[40]  M. Torabi Rad,et al.  On Theory-training Neural Networks to Infer the Solution of Highly Coupled Differential Equations , 2021, ArXiv.

[41]  Pekka Koskinen,et al.  Structural relaxation made simple. , 2006, Physical review letters.

[42]  V. Kouznetsova,et al.  Multi‐scale constitutive modelling of heterogeneous materials with a gradient‐enhanced computational homogenization scheme , 2002 .

[43]  Brian L. DeCost,et al.  Exploring the microstructure manifold: Image texture representations applied to ultrahigh carbon steel microstructures , 2017, 1702.01117.

[44]  Guido Wimmers,et al.  Wood: a construction material for tall buildings , 2017 .

[45]  S. Kalidindi,et al.  Novel microstructure quantification framework for databasing, visualization, and analysis of microstructure data , 2013, Integrating Materials and Manufacturing Innovation.

[46]  C. Cooper,et al.  Osteoporosis: burden, health care provision and opportunities in the EU , 2011, Archives of osteoporosis.

[47]  Frederic E. Bock,et al.  A Review of the Application of Machine Learning and Data Mining Approaches in Continuum Materials Mechanics , 2019, Front. Mater..

[48]  N. Fleck,et al.  High fracture toughness micro-architectured materials , 2020 .

[49]  Daniel W. Davies,et al.  Machine learning for molecular and materials science , 2018, Nature.

[50]  D. Dimiduk,et al.  Perspectives on the Impact of Machine Learning, Deep Learning, and Artificial Intelligence on Materials, Processes, and Structures Engineering , 2018, Integrating Materials and Manufacturing Innovation.

[51]  D. Raabe,et al.  Influence of microstructure and atomic-scale chemistry on the direct reduction of iron ore with hydrogen at 700°C , 2021 .

[52]  P.‐J. Sell,et al.  The Surface Tension of Solids , 1966 .

[53]  Yu Shouwen,et al.  Finite element characterization of the size-dependent mechanical behaviour in nanosystems , 2006, Nanotechnology.

[54]  Liang Feng,et al.  Machine learning–assisted molecular design and efficiency prediction for high-performance organic photovoltaic materials , 2019, Science Advances.

[55]  Paris Perdikaris,et al.  Multiscale Modeling Meets Machine Learning: What Can We Learn? , 2019 .

[56]  Ying Zhao,et al.  Effects of surface tension and electrochemical reactions in Li-ion battery electrode nanoparticles , 2016 .

[57]  Morton E. Gurtin,et al.  A continuum theory of elastic material surfaces , 1975 .

[58]  Stefanie Reese,et al.  A nonlocal method for modeling interfaces: Numerical simulation of decohesion and sliding at grain boundaries , 2020 .

[59]  Jürgen Schmidhuber,et al.  Long Short-Term Memory , 1997, Neural Computation.

[60]  Radu Grosu,et al.  Neural circuit policies enabling auditable autonomy , 2020, Nature Machine Intelligence.

[61]  J. Behler.  Perspective: Machine learning potentials for atomistic simulations. , 2016, The Journal of chemical physics.

[62]  Shyue Ping Ong,et al.  Deep neural networks for accurate predictions of crystal stability , 2017, Nature Communications.

[63]  A. Choudhary,et al.  Perspective: Materials informatics and big data: Realization of the “fourth paradigm” of science in materials science , 2016 .

[64]  Chiho Kim,et al.  Machine learning in materials informatics: recent applications and prospects , 2017, npj Computational Materials.

[65]  M. Buehler,et al.  Deep learning model to predict complex stress and strain fields in hierarchical composites , 2021, Science Advances.

[66]  Hadi Salehi,et al.  Emerging artificial intelligence methods in structural engineering , 2018, Engineering Structures.

[67]  Xiang Li,et al.  Predicting the effective mechanical property of heterogeneous materials by image based modeling and deep learning , 2019, Computer Methods in Applied Mechanics and Engineering.

[68]  Shijie Cheng,et al.  Artificial Intelligence to Power the Future of Materials Science and Engineering , 2020, Adv. Intell. Syst..

[69]  P. Steinmann,et al.  RVE-based studies on the coupled effects of void size and void shape on yield behavior and void growth at micron scales , 2006 .

[71]  Michael Kaliske,et al.  DNN2: A hyper-parameter reinforcement learning game for self-design of neural network based elasto-plastic constitutive descriptions , 2021 .

[72]  A. Stukowski.  Visualization and analysis of atomistic simulation data with OVITO–the Open Visualization Tool , 2009 .

[73]  Y. Estrin,et al.  Architectured Lattice Materials with Tunable Anisotropy: Design and Analysis of the Material Property Space with the Aid of Machine Learning , 2020, Advanced Engineering Materials.

[74]  Liwei Wang,et al.  Data-Driven Topology Optimization with Multiclass Microstructures using Latent Variable Gaussian Process , 2020, Journal of Mechanical Design.

[75]  Elizabeth A. Holm,et al.  A computer vision approach for automated analysis and classification of microstructural image data , 2015 .

[76]  Bhushan Lal Karihaloo,et al.  A scaling law for properties of nano-structured materials , 2006, Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences.

[77]  J. Godet,et al.  A fully molecular dynamics-based method for modeling nanoporous gold , 2019, Computational Materials Science.

[78]  D. Kochmann,et al.  Inverse-designed spinodoid metamaterials , 2020, npj Computational Materials.

[79]  Alexie M. Kolpak,et al.  Discovering charge density functionals and structure-property relationships with PROPhet: A general framework for coupling machine learning and first-principles methods , 2017, Scientific Reports.

[80]  Konstantinos Karapiperis,et al.  Data-Driven multiscale modeling in mechanics , 2020, Journal of the Mechanics and Physics of Solids.

[81]  R. Pokharel,et al.  Machine learning based surrogate modeling approach for mapping crystal deformation in three dimensions , 2021 .

[82]  Michael J. Mehl,et al.  Interatomic potentials for monoatomic metals from experimental data and ab initio calculations , 1999 .

[83]  P. Steinmann,et al.  Homogenization of Composites with Extended General interfaces: Comprehensive Review and Unified Modeling , 2021 .

[84]  J. S. Huang,et al.  Artificial Intelligence in Materials Modeling and Design , 2020, Archives of Computational Methods in Engineering.

[85]  Steve Plimpton,et al.  Fast parallel algorithms for short-range molecular dynamics , 1993 .

[86]  Peyman Mostaghimi,et al.  Machine learning for predicting properties of porous media from 2d X-ray images , 2020 .

[87]  David L. McDowell,et al.  The need for microstructure informatics in process–structure–property relations , 2016 .

[88]  D. Raabe,et al.  Teaching solid mechanics to artificial intelligence—a fast solver for heterogeneous materials , 2021, npj Computational Materials.

[89]  Filippo Masi,et al.  Thermodynamics-based Artificial Neural Networks for constitutive modeling , 2020, Journal of the mechanics and physics of solids.

[90]  Eric Darve,et al.  Learning Constitutive Relations using Symmetric Positive Definite Neural Networks , 2020, J. Comput. Phys..

[91]  T. Bieler,et al.  Overview of constitutive laws, kinematics, homogenization and multiscale methods in crystal plasticity finite-element modeling: Theory, experiments, applications , 2010 .

[92]  Paul Raccuglia,et al.  Machine-learning-assisted materials discovery using failed experiments , 2016, Nature.