A dynamic $ \alpha$ model for the Lagrangian Averaged Navier-Stokes-$ \alpha$ equations

Hongwu Z., K. Mohseni, and J. E. Marsden

Proceedings of IMECE04, IMECE2004, 61591, 2004, 1-9

Abstract:

A dynamic model for calculating the parameter $ \alpha$ in the Lagrangian Averaged Navier-Stokes-$ \alpha$ (LANS-$ \alpha$) equations is derived. The incompressible Navier-Stokes equations are Helmholtz-filtered at the grid and test filter levels. A Germano type identity is derived by comparing the filtered subgrid scale stress terms with those given in the LANS-$ \alpha$ equations. Assuming a constant value of $ \alpha$ and by averaging in the homogenous directions of the flow, a nonlinear equation for the parameter $ \alpha$ is derived, which determine the variation of $ \alpha$ in the non-homogeneous directions or in time. The resulting nonlinear equation is then solved by an iterative technique, in which the parameter $ \alpha$ is calculated during the simulation instead of a pre-defined value. The dynamic LANS-$ \alpha$ model is initially tested for the isotropic homogenous forced and decaying turbulence, where the value of $ \alpha$ is constant over the computational box while its variation in time is allowed. The results of the dynamic LANS-$ \alpha$ simulations are compared with the direct numerical simulations and with the LANS-$ \alpha$ simulations with constant value of $ \alpha$. It is found that the total kinetic energy decay rate and the energy spectra are predicted accurately by the dynamic LANS-$ \alpha$ equations. In order to verify the applicability of the dynamic LANS-$ \alpha$ model in spatially varying turbulent flow an a-priori test in a wall bounded flow is performed. The parameter $ \alpha$ is found, as expected, to change in the wall normal direction where the turbulent scales vary. A correct behavior of the subgrid scale stress in the wall normal direction is observed using the dynamic LANS-$ \alpha$ equations. The current results show the first derivation and application of a dynamic LANS-$ \alpha$ model in spatially varying turbulent flow.

pdf.gif