🤖 AI Summary
Predicting effective constitutive tensors (e.g., elasticity, thermal conductivity, permeability) for heterogeneous materials faces key bottlenecks: high computational cost, non-differentiability, and violations of fundamental physical bounds—namely the Voigt (upper) and Reuss (lower) limits. To address these, this paper proposes a physics-constrained neural network incorporating spectral normalization. At the architectural level, the model intrinsically enforces Loewner order constraints, rigorously guaranteeing output tensor symmetry and adherence to Voigt–Reuss bounds—achieving feature-agnostic generality. By synergistically integrating Voigt–Reuss homogenization theory with surrogate modeling, the method attains superior accuracy and strong robustness on large-scale datasets, significantly outperforming conventional neural networks. Moreover, its fully differentiable formulation enables gradient-based microstructure inverse design—a capability absent in standard black-box surrogates.
📝 Abstract
Heterogeneous materials are crucial to producing lightweight components, functional components, and structures composed of them. A crucial step in the design process is the rapid evaluation of their effective mechanical, thermal, or, in general, constitutive properties. The established procedure is to use forward models that accept microstructure geometry and local constitutive properties as inputs. The classical simulation-based approach, which uses, e.g., finite elements and FFT-based solvers, can require substantial computational resources. At the same time, simulation-based models struggle to provide gradients with respect to the microstructure and the constitutive parameters. Such gradients are, however, of paramount importance for microstructure design and for inverting the microstructure-property mapping. Machine learning surrogates can excel in these situations. However, they can lead to unphysical predictions that violate essential bounds on the constitutive response, such as the upper (Voigt-like) or the lower (Reuss-like) bound in linear elasticity. Therefore, we propose a novel spectral normalization scheme that a priori enforces these bounds. The approach is fully agnostic with respect to the chosen microstructural features and the utilized surrogate model. All of these will automatically and strictly predict outputs that obey the upper and lower bounds by construction. The technique can be used for any constitutive tensor that is symmetric and where upper and lower bounds (in the L""owner sense) exist, i.e., for permeability, thermal conductivity, linear elasticity, and many more. We demonstrate the use of spectral normalization in the Voigt-Reuss net using a simple neural network. Numerical examples on truly extensive datasets illustrate the improved accuracy, robustness, and independence of the type of input features in comparison to much-used neural networks.