Physics-informed machine learning in the determination of effective thermomechanical properties

Physics-informed machine learning in the determination of effective thermomechanical properties

SOYARSLAN Celal, PRADAS Marc

Abstract. We determine the effective (macroscopic) thermoelastic properties of two-phase composites computationally. To this end, we use a physics-informed neural network (PINN)-mediated first-order two-scale periodic asymptotic homogenization framework. A diffuse interface formulation is used to remedy the lack of differentiability of property tensors at phase interfaces. Considering the reliance on the standard integral solution for the property tensors on only the gradient of the corresponding solutions, the emerging unit cell problems are solved up to a constant. In view of this and the exact imposition of the periodic boundary conditions, it is merely the corresponding differential equation that contributes to minimizing the loss. This way, the requirement of scaling individual loss contributions of different kinds is abolished. The developed framework is applied to a planar thermoelastic composite with a hexagonal unit cell with a circular inclusion by which we show that PINNs work successfully in the solution of the corresponding thermomechanical cell problems and, hence, the determination of corresponding effective properties.

Keywords
PINNs, Computational Homogenization, Thermomechanics, Effective Properties

Published online 4/19/2023, 10 pages
Copyright © 2023 by the author(s)
Published under license by Materials Research Forum LLC., Millersville PA, USA

Citation: SOYARSLAN Celal, PRADAS Marc, Physics-informed machine learning in the determination of effective thermomechanical properties, Materials Research Proceedings, Vol. 28, pp 1621-1630, 2023

DOI: https://doi.org/10.21741/9781644902479-175

The article was published as article 175 of the book Material Forming

Content from this work may be used under the terms of the Creative Commons Attribution 3.0 license. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.

References
[1] S. Torquato, Random Heterogeneous Materials, Microstructure and Macroscopic Properties, 1st ed. Interdisciplinary Applied Mathematics 16, Springer-Verlag New York, 2002.
[2] D. Lukkassen, L.-E. Persson, P. Wall, Some engineering and mathematical aspects on the homogenization method, Compos. Eng. 5 (1995) 519–531. https://doi.org/10.1016/0961-9526(95)00025-I
[3] C. Soyarslan, S. Batgmann, M. Pradas, J. Weissmuller, 3D stochastic bicontinuous microstructures: Generation, topology and elasticity, Acta Mater. 149 (2018) 326-340. https://doi.org/10.1016/j.actamat.2018.01.005
[4] C. Soyarslan, M. Pradas, S. Bargmann, Effective elastic properties of 3D stochastic bi-continuous composites, Mech. Mater. 137 (2019) 103098. https://doi.org/10.1016/j.mechmat.2019.103098
[5] C. Soyarslan, J. Havinga, L. Abelmann, T. van den Boogaard, Asymptotic homogenization in the determination of effective intrinsic magnetic properties of composites, Result. Phys. 44 (2023) 106188. https://doi.org/10.1016/j.rinp.2022.106188
[6] I. Lagaris, A. Likas, D. Fotiadis, Artificial neural networks for solving ordinary and partial differential equations, IEEE Transactions on Neural Networks 9 (1998) pp. 987-1000. https://doi.org/10.1109/72.712178
[7] G. Cybenko, Approximation by superpositions of a sigmoidal function, Mathematics of Control, Signals and Systems 2 (1989) 303–314. https://doi.org/10.1007/BF02551274
[8] K. Hornik, M. Stinchcombe, H. White, Multilayer feedforward networks are universal approximators, Neural Networks 2 (1989) 359–366. https://doi.org/10.1016/0893-6080(89)90020-8
[9] K. Hornik, M. Stinchcombe, H. White, Universal approximation of an unknown mapping and its derivatives using multilayer feedforward networks, Neural Networks 3 (1990) 551-560. https://doi.org/10.1016/0893-6080(90)90005-6
[10] L. Lu, X. Meng, Z. Mao, G.E. Karniadakis, DeepXDE: A Deep Learning Library for Solving Differential Equations, SIAM Review 63 (2021) 208–228. https://doi.org/10.1137/19M1274067
[11] I. Lagaris, A. Likas, D. Papageorgiou. Neural-network methods for boundary value problems with irregular boundaries, IEEE Transactions on Neural Networks 11 (2000) 1041-1049. https://doi.org/10.1109/72.870037
[12] J. Sirignano K. Spiliopoulos, DGM: A deep learning algorithm for solving partial differential equations, J. Computat. Phys. 375 (2018) 1339–1364. https://doi.org/10.1016/j.jcp.2018.08.029
[13] E. Haghighat, M. Raissi, A. Moure, H. Gomez, R. Juanes, A physics-informed deep learning framework for inversion and surrogate modeling in solid mechanics, Comp. Meth. Appl. Mech. Eng. 379 (2021) 113741.
[14] S. Cuomo, V. Schiano di Cola, F. Giampaolo, G. Rozza, M. Raissi, F. Piccialli, Scientific Machine Learning through Physics-Informed Neural Networks: Where we are and What’s next, arXiv preprint (2022), arXiv: 2201.05624
[15] C. Soyarslan, M. Pradas, Physics-informed machine learning in asymptotic homogenization of elliptic equations. arXiv preprint (2023), arXiv: 2301.09725
[16] W. Nowacki, Dynamic Problems of Thermoelasticity, Polish Scientific Publishers Warszawa, Poland, 1975.
[17] R. Guinovart-Díaz, R. Rodriguez-Ramos, J. Bravo-Castilero, F.J. Sabina, G.A. Maugin, Closed-Form Thermoelastic Moduli of a Periodic Three-Phase Fiber-Reinforced Composite, J. Therm. Stress. 28 (2005) 1067–1093. https://doi.org/10.1080/014957390967730
[18] F.E. Neumann. Vorlesungen über mathematische Physik: Vorlesungen über die Theorie der Elasticität der festen Körper und des Lichtäthers. B. G. Teubner, Leipzig, 1885.
[19] J.F. Nye, Physical properties of crystals: their representation by tensors and matrices, Oxford University Press, 1985.