Large-eddy simulation of heat transfer downstream of a backward-facing step

Keating, Anthony; Piomelli, Ugo; Bremhorst, K.; Nešić, Srdjan

doi:10.1088/1468-5248/5/1/020

Cited by 77 publications

(56 citation statements)

References 18 publications

(43 reference statements)

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…For the backward-facing step a Fourier expansion is performed in the spanwise direction only and for each wave number a pentadiagonal matrix is inverted directly using cyclic reduction [15]. The code is parallelized using a message-passage interface and has been validated extensively for a variety of turbulent flows [16][17][18]. To represent the step in the backward-facing step simulation, an immersed boundary method with direct forcing was used [19].…”

Section: A Governing Equationsmentioning

confidence: 99%

“…At the outlet, a convective outflow boundary condition [29] was used. The length of the expanded duct is L out = 20h s , which was also used in previous LES simulations [17,30]. To ensure the adequacy of L out , a longer domain with L out = 30h s was also used; the flow statistics at x = 19h s showed very little difference.…”

Section: B Backward-facing Stepmentioning

confidence: 99%

“…Figure 11 highlights the clustering of grid points at the step edge and near the walls. The models were tested on three grids, summarized in Table III: was adopted in previous studies of this problem [17,30]. Here we also consider a coarser and a finer level for a more extensive comparison.…”

Section: B Backward-facing Stepmentioning

confidence: 99%

See 2 more Smart Citations

Dynamic subfilter-scale stress model for large-eddy simulations

2016

Self Cite

View full text Add to dashboard Cite

We present a modification of the integral length-scale approximation (ILSA) model originally proposed by Piomelli et al. [Piomelli et al., J. Fluid Mech. 766, 499 (2015)] and apply it to plane channel flow and a backward-facing step. In the ILSA models the length scale is expressed in terms of the integral length scale of turbulence and is determined by the flow characteristics, decoupled from the simulation grid. In the original formulation the model coefficient was constant, determined by requiring a desired global contribution of the unresolved subfilter scales (SFSs) to the dissipation rate, known as SFS activity; its value was found by a set of coarse-grid calculations. Here we develop two modifications. We define a measure of SFS activity (based on turbulent stresses), which adds to the robustness of the model, particularly at high Reynolds numbers, and removes the need for the prior coarse-grid calculations: The model coefficient can be computed dynamically and adapt to large-scale unsteadiness. Furthermore, the desired level of SFS activity is now enforced locally (and not integrated over the entire volume, as in the original model), providing better control over model activity and also improving the near-wall behavior of the model. Application of the local ILSA to channel flow and a backward-facing step and comparison with the original ILSA and with the dynamic model of Germano et al. [Germano et al., Phys. Fluids A 3, 1760(1991 show better control over the model contribution in the local ILSA, while the positive properties of the original formulation (including its higher accuracy compared to the dynamic model on coarse grids) are maintained. The backward-facing step also highlights the advantage of the decoupling of the model length scale from the mesh.

show abstract

Section: A Governing Equationsmentioning

confidence: 99%

Section: B Backward-facing Stepmentioning

confidence: 99%

See 1 more Smart Citation

Dynamic subfilter-scale stress model for large-eddy simulations

2016

Self Cite

View full text Add to dashboard Cite

show abstract

“…A well-validated finite-difference code (Keating et al 24 ) based on a staggered grid was used to solve the discretized equations. The governing equations are integrated in time using a fractionalstep method.…”

Section: Problem Formulationmentioning

confidence: 99%

Numerical investigation of pulsatile flow in endovascular stents

2013

Self Cite

View full text Add to dashboard Cite

The flow in a plane channel with two idealized stents (one -shaped, the other X-shaped) is studied numerically. A periodic pressure gradient corresponding to one measured in the left anterior descending coronary artery was used to drive the flow. Two Reynolds numbers were examined, one (Re = 80) corresponding to resting conditions, the other (Re = 200) to exercise. The stents were implemented by an immersed boundary method. The formation and migration of vortices that had been observed experimentally was also seen here. In the previous studies, the compliance mismatch between stent and vessel was conjectured to be the reason for this phenomenon. However, in the present study we demonstrate that the vortices form despite the fact that the walls were rigid. Flow visualization and quantitative analysis lead us to conclude that this process is due to the stent wires that generate small localized recirculation regions that, when they interact with the near-wall flow reversal, result in the formation of these vortical structures. The recirculation regions grow and merge when the imposed waveform produces near-wall flow reversal, forming coherent quasi-spanwise vortices, that migrate away from the wall. The flow behavior due to the stents was compared with an unstented channel. The geometric characteristics of the -stent caused less deviation of the flow from an unstented channel than the X-stent. Investigating the role of advection and diffusion indicated that at Re = 80 advection has negligible contribution in the transport mechanism. Advection plays a role in the generation of streamwise vortices created for both stents at both Reynolds numbers. The effect of these vortices on the near-wall flow behavior is more significant for the -stent compared to the X-stent and at Re = 200 with respect to Re = 80. Finally, it was observed that increasing the Reynolds number leads to early vortex formation and the creation of the vortex in a stented channel is coincident with the near wall flow reversal in an unstented one. C 2013 AIP Publishing LLC. [http://dx

show abstract

“…• The scale similarity hypothesis can be used to obtain others models which are combined with the Smagorinsky model to give a mixed model in the following form ( [21][22][23][24]):…”

Section: Subgrid Modelsmentioning

confidence: 99%

The Symmetry Group of the Non-Isothermal Navier–Stokes Equations and Turbulence Modelling

et al. 2010

View full text Add to dashboard Cite

In this work, the non-isothermal Navier-Stokes equations are studied from the group theory point of view. The symmetry group of the equations is presented and discussed. Some standard turbulence models are analyzed with the symmetries of the equations. A class of turbulence models which preserve the physical properties contained in the symmetry group is built. The proposed turbulence models are applied to an illustrative example of natural convection in a differentially heated cavity, and the results are presented.

show abstract

Large-eddy simulation of heat transfer downstream of a backward-facing step

Cited by 77 publications

References 18 publications

Dynamic subfilter-scale stress model for large-eddy simulations

Dynamic subfilter-scale stress model for large-eddy simulations

Numerical investigation of pulsatile flow in endovascular stents

The Symmetry Group of the Non-Isothermal Navier–Stokes Equations and Turbulence Modelling

Contact Info

Product

Resources

About