An adaptive local deconvolution method for implicit LES with optimized spectral eddy viscosity

Abstract

The Adaptive Local Deconvolution Method (ALDM) is a new nonlinear discretization scheme designed for Implicit Large-Eddy Simulation (ILES). Free parameters of a solution-adaptive deconvolution operator allow to adjust the numerical truncation error which acts as implicit subgrid-scale (SGS) model. To assign physically appropriate values to these parameters the spectral numerical viscosity is analyzed. By an automatic optimization employing an evolutionary algorithm a set of parameters was found which yields an excellent spectral match for the numerical viscosity with theoretical predictions. Computational results for forced and decaying three-dimensional homogeneous isotropic turbulence show an excellent agreement with theory and experimental data.