Channel flow Re = 395

Inlet data and the new DFSEM

Before proceeding to the results let us have a look at the application of the new method to the DNS inlet data set:

These first two pictures give a shot of the points to be represented inside the lumley triangle. As we discussed earlier the method is able to reproduce very high anysotropic states but suffers some problems in case of 2D turbulent structures. Anyway, we can see that turbulence is very well reproduced if $ y^+ > 10 $.

As we had discussed earlier, the new DFSEM reproduces elipsoids structures (by extending the length scale in one direction and keeping constant the others) and then rotate these structures, in order to allign them to the principal reference system of the local reynolds stress. At this stage is worth a while having a closer look at the length scale and at the rotation angle!

It is important to notice that some structure may end up out of the domain (as the following picture highlights)!!

The main effect on the output Reynolds stresses is sketched into the following plot:

As one can see, the stresses are accurately reproduced for $ \frac{x}{\delta}=0.5, y^+=200 $, then (since right at that point structures start not to be fully contained within the domain) a rougher approximation of the stresses begins. What is very important (I think) is their valye very close to the wall: it does not go to 0!! Can a limitation of these structure help? let's have a look at it later on!

Results

Results (at the moment) are not so encouraging but I must say that I noticed a minor bug in the simulation which may have affected them!

The errors I made are:

  1. Under-extimation of the maximum sigma -> eddy box smaller than what it has to be! (the eddy are not fully contained within it!!)
  2. Minor error for those points outside the reproducible area

I personally think that the first error actually affects the simulation, as part of the eddy is not considered (thus resulting in a error in the reynolds stresses). A new simulation is now running!

And here you are the development of the shear stress along the channel (new DFSEM vs previous version and SEM) :

Final and corrected results

Results are very very positive!!! The friction coefficient variation downstream the channel is reduced with respect to any simulation tryed so far! This results finds confirmation in the development of the Reynolds stresses downstream the channel (see followfing pics - comparison New DFSEM vs old DFSEM


Current Tags:
create new tag
, view all tags
Topic attachments
I Attachment Action Size Date Who Comment
pngpng CfcomparisonDNS.png manage 83.6 K 2012-02-13 - 09:16 RuggeroPoletto  
Topic revision: r7 - 2012-02-13 - 09:16:37 - RuggeroPoletto
Main Web
25 Sep 2018

Site

Manchester CfdTm
Code_Saturne

Ongoing Projects

ATAAC
KNOO

Previous Projects

DESider
FLOMANIA

Useful Links:

User Directory
Photo Wall
Upcoming Events
Add Event
 

Computational Fluid Dynamics and Turbulence Mechanics
@ the University of Manchester
Copyright © by the contributing authors. Unless noted otherwise, all material on this web site is the property of the contributing authors.