Learning hyperelastic anisotropy from data via a tensor basis neural network

TL;DR

This paper introduces a tensor basis neural network that learns hyperelastic anisotropy from data, accurately modeling stress responses and identifying material symmetry and orientation.

Contribution

It develops a novel neural network architecture combining classical representation theory with regularization to discover anisotropy type and orientation from stress-strain data.

Findings

Successfully models hyperelastic materials with various anisotropies

Can identify symmetry type and orientation of anisotropic materials

Establishes polyconvex potential ensuring solution existence

Abstract

Anisotropy in the mechanical response of materials with microstructure is common and yet is difficult to assess and model. To construct accurate response models given only stress-strain data, we employ classical representation theory, novel neural network layers, and L1 regularization. The proposed tensor-basis neural network can discover both the type and orientation of the anisotropy and provide an accurate model of the stress response. The method is demonstrated with data from hyperelastic materials with off-axis transverse isotropy and orthotropy, as well as materials with less well-defined symmetries induced by fibers or spherical inclusions. Both plain feed-forward neural networks and input-convex neural network formulations are developed and tested. Using the latter, a polyconvex potential can be established, which, by satisfying the growth condition can guarantee the existence of boundary value problem solutions.