| Large
Eddy Simulation of Turbulent Channel Flows
Large eddy simulation (LES) is one of the most successful
techniques in the numerical simulation of turbulent flows.
Unlike direct numerical simulation (DNS), which tries to capture
all the scales in the flow, LES aims at resolving only the
large-scale flow features as defined by a filtering operation.
One of the challenges in LES is modeling the subgrid-scale
stresses, and a wide variety of models have been developed
for this purpose.
 |
|
 |
|
Figure
2 Normalized mean streamwise velocity,
Re
= 180. The almost perfect overlapping of the results is
interpreted as a measure of success in enforcing a constant
mass flux through the channel. |
|
Iliescu and Fischer applied the rational LES model (RLES)
to numerical simulations of incompressible channel flows at
Reynolds numbers based on the friction velocity and the channel
half-width Re
= 180 and Re
= 395. RLES is an approximate deconvolution model based on
a rational (Padé) approximation to the Fourier transform
of the Gaussian filter and is proposed as an alternative to
the gradient model. The authors compared the RLES results
with those from the gradient model, the Smagorinsky model,
and a coarse DNS with no LES model; all of these were benchmarked
against the fine DNS calculations of Moser et al.
The RLES model yielded the best results for the Re
= 180 case and showed much better numerical stability than
the gradient model (Figure 2). For the Re
= 395 case, the RLES model and the gradient model yielded
comparable results, and the Smagorinsky model performed the
best. The next step will be to develop a mixed model, consisting
of RLES supplemented by a Smagorinsky model.
INVESTIGATORS
P. Fischer, T. Iliescu, G. Leaf, and M. Minkoff, Argonne National
Laboratory; J. S. Mullen, Worcester Polytechnic Institute.
PUBLICATION
T. Iliescu and P. Fischer, “Large eddy simulation of
turbulent channel flows by the rational LES model,”
Phys. Fluids (submitted), preprint math.SC/0205264 (2002).
URL
http://www-unix.mcs.anl.gov/appliedmath/Flow/cfd.html |
