Approximate deconvolution models of turbulence. Analysis, phenomenology and numerical analysis. (Q642466)

This monograph presents a mathematical approach to turbulence modeling and is aimed at graduate students and researchers in the field of computational fluid dynamics. The starting point are the incompressible Navier-Stokes equations which describe the flow evolution, even in the turbulent regime, deterministically. In the regime of fully developed turbulence, however, the range of dynamically active scales becomes prohibitive and modeling is unavoidable for simulating turbulent flows. The start is the large eddy simulation approach which is presented. The range of active scale is significantly reduced as only the large scales of the turbulent flow are computed deterministically while the influence of small scales onto the large ones is modeled. The large scale contributions are obtained by applying a low pass filter to the Navier-Stokes equations. The idea of approximate deconvolution models is then to seek for an approximate solution of an inverse problem, i.e., the filter is inverted in suitable way, using e.g. regularization techniques, and the small scale contributions of the flow are reconstructed at least approximately. This allows the construction of improved turbulence models for the small scales which are not resolved in the numerical simulation. The book presents the governing Navier-Stokes equations and the basics of large eddy simulation from a mathematical perspective without going into details of the flow physics. Different approximate deconvolution operators and regularization techniques are reviewed, and their phenomenology is described. Time relaxation of the truncated scales is also discussed. Difficulties of the models in the vicinity of walls with no-slip boundary conditions are mentioned, and numerical examples of different flows illustrate some properties of the approximate deconvolution models. Some appendices and a list of references conclude this volume.