A Machine Learning Approach to Unresolved-Scale Modeling for Burgers’ Equation

This thesis is part of a greater effort to use machine learning for the development of flexible and universal unresolved-scale models in large eddy simulation (LES). The novelty in the current work is that a neural network learns to predict the integral forms of the unresolved-scale terms directly without a priori assumptions on the underlying...

A numerical investigation of the boundary commutation error in large eddy simulation

The boundary commutation error in large eddy simulation is studied numerically. It is shown that an accurate representation of the boundary commutation error has an important role in a reliable large eddy simulation of wall bounded flows. The numerical experiments in this paper are conducted for a 2D channel flow.

Space-time multiscale methods for Large Eddy Simulation

The Variational Multiscale (VMS) method has appeared as a promising new approach to the Large Eddy Simulation (LES) of turbulent flows. The key advantage of the VMS approach is that it allows different subgrid-scale (SGS) modeling assumptions to be made at different ranges of the resolved scales. Typically, in the VMS method, SGS modeling is...