Generation of synthetic turbulent inflow data for large eddy simulation of spatially evolving wall-bounded flows

Abstract

A method for generating inflow conditions for large eddy simulations (LESs) of spatially developing turbulent boundary layers is presented. It is an adaptation of the synthetic eddy method (SEM) of Jarrin et al. [Int. J. Heat Fluid Flow 27, 585 (2006)], which uses the Cholesky decomposition of the Reynolds stress tensor to enforce second-order moments starting from a normalized stochastic velocity signal, the latter being constructed with a superimposition of turbulent structures with prescribed geometrical shape and random signs and position. The present method modifies the definition of the stochastic signal so that it can be split into several modes, with different time, length and velocity scales and also with different vorticity contents. The idea is to reproduce more realistically the distribution of scales in the wall-normal direction of a turbulent boundary layer flow. The novelty of the proposed modified SEM is that physical information concerning the coherent vortical structures of such flows are extracted from the literature and used in the definition of the modes. It is shown that the specification of realistic modes for the buffer and the logarithmic layers significantly helps to reduce the spatial transient undergone by the synthetic inflow data. The new method is assessed in the framework of LES and compared to the original SEM and to a reference simulation which uses the recycling procedure of Lund et al. [J. Comput. Phys. 140, 233 (1998)]. First- and second-order statistical results, as well as instantaneous behavior of turbulence, are shown to be in excellent agreement with the reference after an adaptation distance of five to six initial boundary layer thicknesses.