Catálogo de publicaciones - libros

This paper investigates the capabilities of different subgrid scale (SGS) models, including the recent “multiscale” models, for large-eddy simulation (LES), here in a vortex-in-cell (VIC) method, of complex wake vortex dynamics. More specifically, we here consider the multiscale dynamics developing in a counterrotating four-vortex system, that evolves from a simple state to a turbulent state. The various SGS models are tested and compared on this complex and transitional flow. Comparisons are also made with results obtained using a pseudo-spectral method. Energy diagnostics (global and modal) and spectra are provided and used to support the comparisons. A discussion on the applicability of the various models to LES of complex wake vortex flows is made. The multiscale models are seen to be the most appropriate.

LES quality assessment is very important in view of predictive LES applications. Recently Klein [1] proposed to evaluate the numerical as well as the modeling error in a LES using an approach based on Richardson extrapolation, where it is assumed that the modeling error scales like a power law. In order to apply this approach, the scaling exponent for the numerical error with respect to the filter width has to be known in advance. This scaling law will be explored for three different configurations: a channel flow, a plane jet and a swirling recirculating flow. Theoretical argumentation [2, 3] leads to a scaling of = 2/3. The current findings suggest to use Δ for flow configurations operating at moderate Reynolds numbers. The resulting scaling exponent will be used to assess the quality of LES simulations of these configurations.

In this paper, numerical and modelling errors in large eddy simulation (LES) applying the standard Smagorinsky model and a second-order scheme are studied and compared a posteriori in a turbulent plane Poiseuille flow. The gridindependent solution of the LES equations is approached keeping the product of the model coefficient and the length scale constant as the grid is refined. The aim is to clarify the choice of the model parameters via the two error components, and to study the possibility to control the numerical error via the built-in filter of the model.

The purpose of this work was to extend knowledge on the potential of the Linear Eddy Model (LEM) to reproduce scalar statistics that are relevant to the modelling of turbulent combustion. In this context, emphasis was placed on the evolution of the scalar dissipation and its conditional moments. To allow for a systematic and quantitative analysis on the sensitivity of model to its input parameters, LEM has been employed as a stand alone model with prescribed flow field properties rather than being incorporated in an LES code as a subgrid model. Results from twelve parametric LEM computations are compared to experiments and DNS data from non-combusting and reactive turbulent mixing layers. It is shown that the LEM model reproduces a number of important features of scalar fields. However, input parameters are shown to significantly influence the model behaviour so that care is required when employing the model particularly at early times.

The success in the use of lattice-Boltzmann methods for the simulation of laminar .ows has prompted the interest in extending the technique for the simulation of turbulent flows. In this paper, a square cylinder at = 21400 is used to investigate the suitability of lattice-Boltzmann methods to solve turbulent unsteady problems with open boundary conditions. Numerical simulations are performed with a single-relaxation-time lattice-Boltzmann method and with a large-eddy simulation turbulence model. These simulations present several not-fully-resolved issues for lattice-Boltzmann methods, which are briefly discussed in this paper. Specifically, we propose a filtered-density open-boundary condition for fluid-flow simulations with the lattice-Boltzmann equation. This filter modification of inflow boundary condition is shown to improve simulations of unsteady flows at high Reynolds numbers, in terms of calculation stability and speed, and quality of the results.

Cartesian grids with local anisotropic adaptation are combined with an Immersed Boundary (IB) method and used to perform large eddy simulations. The approach used to generate Cartesian grids with desired normal and tangential resolution to three-dimensional surfaces is described. Anisotropic refinement result in a considerable reduction in grid size. An IB treatment based on solution reconstruction is proposed; it has the property of ensuring local mass conservation and its accuracy is investigated for laminar and turbulent flows in channel not aligned with the grid. Mesh modifications that improve the grid quality for LES are proposed. The resulting mesh then requires a fully unstructured discretization and a parallel polyhedral-based finite-volume solver is applied to perform simulations of the flow around several spheres.

Control-volume (cv) and node-based (cell-vertex) finite volume discretizations of the incompressible Navier-Stokes equations are compared in terms of accuracy, efficiency, and stability using the inviscid Taylor vortex problem. An energy estimate is shown to exist for both formulations, and stable convective boundary conditions are formulated using the simultaneous approximation term (SAT) method. Numerical experiments show the node-based formulation to be generally superior on both structured Cartesian and unstructured triangular grids, displaying consistent error levels and nearly second-order rates of velocity error reduction. The cv-formulation, however, out-performs the node-based for the case of Cartesian grids when the Taylor vortices do not cut the boundary.

This paper presents a hybrid finite-difference method for the large-eddy simulation of compressible flows with low-numerical dissipation and structured adaptive mesh refinement (SAMR). A conservative flux-based approach is described. An explicit centered scheme is used in turbulent flow regions while a weighted essentially non-oscillatory (WENO) scheme is employed to capture shocks. Several two- and three-dimensional numerical experiments and validation calculations are presented including homogeneous shock-free turbulence, turbulent jets and the strongly shockdriven mixing of a Richtmyer-Meshkov instability.

We present results from large-eddy simulations (LES) of three-dimensional Richtmyer-Meshkov (RM) instability in a rectangular tube with reshock off the tube endwall. A hybrid numerical method is used that is shock capturing but which reverts to a centered scheme with low numerical viscosity in regions of smooth flow. The subgrid-scale (SGS) model is the stretched-vortex (SV) model [1].

Three dimensional variable density jets at low Mach number conditions are analyzed by means of Large Eddy Simulations. The LES equations were derived starting from a low Mach number approximation of the continuity, Navier- Stokes and energy equations. The numerical method is based on the high-order compact/Fourier pseudospectral schemes. The superposition of axial and flapping (helical) periodic disturbances at the jet outlet, forced the jet to bifurcate at certain excitation frequency. The LES calculations for non-isothermal case revealed that, for the case when the jet density is lower in comparison with the ambient fluid, the excitation frequency needed for bifurcation is higher than for the isothermal jet.