Showing papers by "Nikolaus A. Adams published in 1999"
TL;DR: An alternative approach to large-eddy simulation based on approximate deconvolution (ADM) is developed in this article, where the main ingredient is an approximation of the nonfiltered field by truncated series expansion of the inverse filter operator.
Abstract: An alternative approach to large-eddy simulation based on approximate deconvolution (ADM) is developed. The main ingredient is an approximation of the nonfiltered field by truncated series expansion of the inverse filter operator. A posteriori tests for decaying compressible isotropic turbulence show excellent agreement with direct numerical simulation. The computational overhead of ADM is similar to that of a scale-similarity model and considerably less than for dynamic models.
735 citations
01 Jan 1999
TL;DR: In this paper, the authors developed and compared two approaches for the proper simulation of flow discontinuities, which are based on approximate deconvolution of the filtered solution to obtain a sufficiently accurate representation of the smoothed nonlinear combination of discontinuous fields, e.g. convection term.
Abstract: We develop and compare two approaches for the proper simulation of flow discontinuities. For the filtered evolution equations the solution is smooth and can be solved for by standard difference schemes without special considerations of discontinuities. Both approaches are based on approximate deconvolution of the filtered solution to obtain a sufficiently accurate representation of the smoothed nonlinear combination of discontinuous fields, e.g., the convection term. One approach is based on a regularized direct deconvolution and requires for stable integration a secondary filter. The other approach is based on an expansion of the filtered field in terms of filter width, which is modified in the vicinity of discontinuities in order to exactly recover the filtered profile. We detail both approaches and investigate their feasibility for the inviscid Burgers equation, isothermal and full Euler equations.
14 citations
01 Jan 1999
10 citations
01 Jan 1999
TL;DR: The approximate deconvolution model (ADM) is applied to LES of incompressible channel flow and shows a significant improvement over classical subgrid scale models such as the standard or the dynamic Smagorinsky model.
Abstract: The approximate deconvolution model (ADM) is applied to LES of incompressible channel flow. With this approach an approximation to the unfiltered data is achieved by repeated filtering. This approximation can be used to compute the non-linear terms in the filtered NavierStokes equations directly without explicit subgrid scale terms. A priori tests show an excellent agreement with direct numerical simulation data. A posteriori tests are performed for incompressible channel flow at two different Reynolds numbers. Both simulations compare well with DNS data and show a significant improvement over classical subgrid scale models such as the standard or the dynamic Smagorinsky model. The computational overhead of the ADM is similar to that of the scale-similarity model and is considerably less than that of dynamic models or the velocity estimation model.
7 citations
01 Jan 1999
TL;DR: The approximate deconvolution model (ADM) is applied to LES of incompressible channel fiow and an approximation to the unfiltered data is achieved by repeated filtering that can be used to compute the non-linear terms in the filtered Navier Stokes equations directly without explicit subgrid scale terms.
Abstract: The approximate deconvolution model (ADM) is applied to LES of incompressible channel fiow. With this approach an approximation to the unfiltered data is achieved by repeated filtering. This approximation can be used to compute the non-linear terms in the filtered Navier Stokes equations directly without explicit subgrid scale terms. A priori tests show an excellent agreement with direct numerical simulation data. Aposteriori tests are performed for incompress ible channel fiow at two different Reynolds numbers. Both simulations compare well with DNS data and show a significant improvement over c1assical subgrid scale models such as the standard or the dynamic Smagorinsky model. The computational overhead of the ADM is similar to that of the scale-similarity model and is considerably less than that of dynamic mode1s or the velocity estimation model.
TL;DR: In this paper, the authors consider isotropic turbulence interacting with a normal shock and show that explicit compressibility effects such as pressure dilatation and dilatations dissipation are non-negligible in the interaction region.
Abstract: We outline results of numerical and theoretical investigation of shock-turbulence interaction. Following the strategy of increasing complexity we first consider isotropic turbulence interacting with a normal shock. In the presence of mean-shear and shock-obliqueness, turbulence-structure becomes dependent on additional parameters such as shock-angle and strain-rate. Presently, the most complex situation achievable by direct numerical simulation is the interaction of a turbulent boundary layer with a shock along a compression ramp. Based on recent results for a low-Reynolds number direct simulation it is found that the simplified cases predict correct trends for turbulent structure. The DNS results indicate that explicit compressibility effects such as pressure dilatation and dilatation dissipation are non-negligible in the interaction region.