Implicit subgrid‐scale modeling by adaptive local deconvolution

A class of implicit Subgrid‐Scale (SGS) models for Large‐Eddy Simulation (LES) is obtained from a new approach for the finite‐volume discretization of hyperbolic conservation laws. The extension of a standard deconvolution operator and the choice of a suitable numerical flux function result in a truncation error which can be forced to have physical significance. As determined by a modified‐differential‐equation analysis, the free parameters of this implicit SGS model can be adjusted to approximate a known explicit SGS model. Computational results for the viscous Burgers equation show that the model with parameters identified by evolutionary optimization gives significantly better predictions than other models. (© 2004 WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim)