Survey of machine-learning wall models for large-eddy simulation

Abstract

This survey addresses two primary concerns within the domain of machine learning for wall modeling in large-eddy simulations: the adherence to physical laws and the perceived opacity of the prediction process, often termed as the ``black box'' problem. Three machine-learning wall models are tested against the law of the wall and their predictions are analyzed and explained. This research marks an initial step towards gaining wider acceptance of machine-learning wall models in the scientific computing community.

Topics & Concepts