Exploring the 3D architectures of deep material network in data-driven multiscale mechanics

This paper extends the deep material network (DMN) proposed by Liu et al. (2019) to tackle general 3-dimensional (3D) problems with arbitrary material and geometric nonlinearities. It discovers a new way of describing multiscale heterogeneous materials by a multi-layer network structure and mechanis...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:Journal of the mechanics and physics of solids Ročník 127; s. 20 - 46
Hlavní autoři: Liu, Zeliang, Wu, C.T.
Médium: Journal Article
Jazyk:angličtina
Vydáno: London Elsevier Ltd 01.06.2019
Elsevier BV
Témata:
3D building-block
Algorithms
Carbon fiber reinforced plastics
CFRP composites
Computer simulation
Crystal plasticity
Elasticity
Equilibrium conditions
Exact solutions
Extrapolation
Fiber composites
Fiber reinforced polymers
Formulations
Hyperelasticity
Machine learning
Mathematical models
Mathematical morphology
Model accuracy
Multilayers
Particulate composites
Plastic properties
Polymer matrix composites
Rubber
Three-scale homogenization
3D building-block
Crystal plasticity
CFRP composites
Three-scale homogenization
Machine learning
Hyperelasticity
ISSN:0022-5096, 1873-4782
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:This paper extends the deep material network (DMN) proposed by Liu et al. (2019) to tackle general 3-dimensional (3D) problems with arbitrary material and geometric nonlinearities. It discovers a new way of describing multiscale heterogeneous materials by a multi-layer network structure and mechanistic building blocks. The data-driven framework of DMN is discussed in detail about the offline training and online extrapolation stages. Analytical solutions of the 3D building block with a two-layer structure in both small- and finite-strain formulations are derived based on interfacial equilibrium conditions and kinematic constraints. With linear elastic data generated by direct numerical simulations on a representative volume element (RVE), the network can be effectively trained in the offline stage using stochastic gradient descent and advanced model compression algorithms. Efficiency and accuracy of DMN on addressing the long-standing 3D RVE challenges with complex morphologies and material laws are validated through numerical experiments, including 1) hyperelastic particle-reinforced rubber composite with Mullins effect; 2) polycrystalline materials with rate-dependent crystal plasticity; 3) carbon fiber reinforced polymer (CFRP) composites with fiber anisotropic elasticity and matrix plasticity. In particular, we demonstrate a three-scale homogenization procedure of CFRP system by concatenating the microscale and mesoscale material networks. The complete learning and extrapolation procedures of DMN establish a reliable data-driven framework for multiscale material modeling and design.
Bibliografie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:0022-5096
1873-4782
DOI:10.1016/j.jmps.2019.03.004