An equivariant neural operator for developing nonlocal tensorial constitutive models

Jiequn Han, Xu-Hui Zhou, Heng Xiao

Abstract: Highlights•Proposed a neural operator for learning nonlocal tensorial constitutive models.•Preserved invariance to coordinate translation and equivariance to coordinate rotation.•Suitable for unstructured meshes with any number of arbitrarily arranged grid points.•Demonstrated predictive capability for emulating the Reynolds stress transport equations.•Demonstrated capability for predicting the Reynolds stress in the DNS dataset.