The approximate deconvolution is implemented to reconstruct the instantaneous velocities from the filtered velocities before using them in the momentum equations of particles in the large-eddy simulation (LES) of particle-laden turbulent flows. It is shown that the various statistics of particles obtained through deconvolution are in good agreement with those obtained by the direct numerical simulation (DNS) by conducting a priori and a posteriori tests in a particle-laden homogeneous shear turbulent flow. On the other hand, the neglect of the effects of subgrid scales on the particles results in discrepancies between DNS and LES results.
Most of the studies on large-eddy simulation (LES) of particle-laden flows assume that the effect of subgrid scales on the particle motion is negligible. This assumption may break down, particularly when particles have a small time constant and/or the filtered energy is significant. In this work, a stochastic model is proposed for the particle motion in LES while considering the effect of subgrid fluctuations. The model assumes that the fluid particle seen by the heavy particle evolves based on a diffusion stochastic process. For model assessment, both a priori and a posteriori tests are conducted for a particle-laden decaying isotropic turbulence. In the a priori test the filtered velocity field, obtained via filtering the DNS velocities, and in the a posteriori test the LES velocity field, obtained via the dynamic Smagorinsky model, are applied to particles through the stochastic model. The small-particle statistics obtained through the stochastic model very well match those obtained through DNS in the a priori test, once the effects of the initial conditions have decayed. There is also a good agreement with DNS results in the a posteriori test. It is shown that the neglect of subgrid fluctuations is not acceptable for these cases. For large particle time constants, the model needs to be adjusted.
1 January 2003 in International Journal of Heat and Mass Transfer
Direct numerical simulation is conducted to study particle concentration in particle-laden homogeneous shear turbulence. The equilibrium assumption in which the velocity of particles is approximated in terms of the velocity and acceleration of the fluid phase is assessed for this flow by analyzing the radial distribution function (RDF), the probability density function (PDF) of the number density of particles, the autocorrelation of particle Lagrangian velocity, and the particle turbulent diffusivity for various particle time constants. A good agreement between exact and equilibrium Lagrangian particles is observed for these statistics when nondimensional particle time constant . In this work, we also introduce a new methodology for measuring the anisotropy of the Lagrangian particle concentration through calculating two-dimensional RDFs. It is shown that particles are the most and least accumulated in the streamwise and cross-stream directions, respectively. Also, a methodology is introduced to extract a spatially smooth number density of particles from discrete Lagrangian particles by employing a locally averaged formulation. In this work, the equilibrium Eulerian equations are reformulated for the case of homogeneous shear turbulence as well and the concentration of particles are simulated in a Eulerian framework by solving the corresponding transport equation. A good agreement between the exact Lagrangian and equilibrium Eulerian results is observed. The governing equations of the fluid phase along with the concentration transport equation of the particle phase are solved by a pseudospectral method.
The article is concerned with the propagation of uncertainties in the values of turbulence model coefficients and parameters in turbulent flows. These coefficients and parameters are obtained through experiments performed on elementary flows, and they are subject to uncertainty. In this work, the widely used turbulence model is considered. It consists of model transport equations for the turbulence kinetic energy and the rate of turbulent dissipation. Both equations involve various model coefficients about which adequate knowledge is assumed known in the form of probability density functions. The study is carried out for a flow over a 2D backward-facing step configuration. The Latin Hypercube Sampling method is employed for the uncertainty quantification purposes as it requires a smaller number of samples compared to the conventional Monte Carlo method. The mean values are reported for the flow output parameters of interest along with their associated uncertainties. The results show that model coefficient variability has significant effects on the streamwise mean velocity in the recirculation region near the reattachment point and turbulence intensity along the free shear layer. The reattachment point location, pressure, and wall shear are also significantly influenced by the uncertainties of the coefficients.