A new method to generate anisotropic synthetic turbulence for LES

Abstract

A method to generate anisotropic synthetic turbulence from turbulence statistics provided by a Reynolds stress model is presented. Unlike most methods it does not use the Cholesky decomposition to generate the shear stresses but a new method which is based on probabilities of Fouriermodes to contribute a certain shear stress to the total shear stress. The method is compared to a classical method which uses the Cholesky decomposition. The results show better fit of the new method with the vertical profiles of Reynolds stresses and integral length scales. Also the new method is able to fit the given 1D spectra perfectly. However the performance of the synthetic turbulence can only be judged in further studies where the synthetic turbulence would be fed into a Large-Eddy Simulation.