Progress in the Development of Stochastic Coherent Adaptive LES Methodology

Abstract

Stochastic Coherent Adaptive Large Eddy Simulation (SCALES) methodology, originally proposed by Goldstein and Vasilyev [1], is an extension of the Large Eddy Simulation (LES) approach that uses a wavelet filter-based dynamic grid adaptation strategy to solve for the most energetic coherent structures in a turbulent flow field, while modeling the effect of the less energetic eddies. Despite some similarities, SCALES is drastically different from classical LES, mainly in its ability to couple grid evolution and flow physics. This direct coupling is achieved by resolving and “tracking” on a space–time adaptive mesh physically important flow structures, whose evolution is influenced by a subgrid scale (SGS) model. This coupling provides a unique feedback mechanism that allows numerics to compensate for the inadequacies of the SGS model: mesh is automatically refined in flow regions, where the SGS model does not provide adequate dissipation, and coarsened in regions, where the model is over-dissipative.