CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   CFX (http://www.cfd-online.com/Forums/cfx/)

 az_f August 30, 2009 20:46

Hi All
Could anyone tell me how can i calculate wall shear stress gradient in normal and tangential direction in cfx-post.

Thanks

 ghorrocks August 31, 2009 07:43

Wall shear stress only exists on the wall (ie a surface) so how can you have a normal gradient of it?

 az_f August 31, 2009 09:22

Hi glen,
I 'm not really sure how to answer your question. I found below formula from this paper
"Flow-induced Wall Shear Stress in Abdominal Aortic Aneurysms: Part I – Steady Flow Hemodynamics

Computer Methods in Biomechanics and Biomedical Engineering
, 2002 Vol. 5 (4), pp. 309–318"

"The calculation of the local Wall Shear Stress Gradient (WSSG) is based on the predictor equation proposed by Lei and Kleinstreuer [19] at the cellular level:
WSSG=sqrt((dtu/dt^)^2+(dtu/dn^)^2)
where tˆ and are the local tangential and normal directions to the wall,"

I want to look at rupture of aneurysm so in addition to the magnitude , I need to consider the direction of the WSS as well so I tought this eqation might help me.
Do you have any idea that can help me?
By the way I'm living in sydney too :)
Best Regards,

 ghorrocks August 31, 2009 18:46

The variable tu has not been defined. What is it?

I consider calculating wall shear stress gradient directly in CFD-Post. Have a look in the CEL Expression language documentation (in the reference manual) to see the available functions. It includes gradient functions, but when I try to apply it to a boundary only variable it gives an error.

 az_f August 31, 2009 19:28

Tu is a wall shear stress. I couldn't find the gradient function in CCL. I just found an example that suggest a fortran dll. and itroducing a gradient function as an additional variable. I don't have Fortran to complie the dll. Do you have another suggestion?

thanks

 ghorrocks August 31, 2009 19:42

There are a lot of new gradient calculation functions in V12. Are you using V12?

If tu is the wall shear stress then I go back to my original question - wall shear stress is not defined away from the wall so the gradient is not defined. The formula seems non-physical to me.

 Lance January 28, 2011 05:35

Sorry for bumping an 1½ year old thread, but does anyone have an idea on how the calculate the gradient of the wall shear?
Using the gradient calculator in POST gives me
Internal error detected:

edit: Im using V13

 ghorrocks January 28, 2011 06:23

Yes, that might be tricky in CFX. Wall shear only exists on the walls, and the gradient function only applies to the volume. Not sure what you can do to get the gradient on a wall. You might have to export to another post processor like tecplot, ensight or fieldview (and I am not sure they can do it either).

 Lance January 28, 2011 07:07

Hi,
I also think that post cant do this. I have exported the wall shear values to Matlab, and Im trying to calculate the gradient in various ways, none that seems to work though...

 alinik April 16, 2011 03:06

Hi guys

Did any body succeed in finding a way to calculate WSSG in cfx post? I really need this

 ghorrocks April 16, 2011 07:21

As I said at the time, I suggest you export the wall shear stress and XYZ locations from Post and calculate them using another application. That should be pretty easy.

 alinik April 16, 2011 07:51

I do not have time to find out about a way to introduce WSSG in another software all by myself. Has anybody succeeded in solving this problem?

 ghorrocks April 17, 2011 07:25

? But calculating gradients like this should be easy. You can set up some orthogonal lines in CFD-Post to export the points from so the gradient calculation is very simple. And packages like Tecplot, Fieldview or Ensight probably have in built gizmos to calculate this sort of thing.

 Lance April 18, 2011 13:07

This is what I ended up doing:

Export the WSS and coordinates on the entire surface and compute the gradient (in e.g. Matlab) using the method described in "An implicit upwind algorithm for computing turbulent flows on unstructured grids" Kyle Anderson and Daryl L Bonhaus, Computers & Fluids Volume 23, Issue 1, January 1994, Pages 1-21

It is pretty straightforward, all you need to do is a nearest neighbor search for each node and then implement eqn. 20. It works with unstructured/structured grids.

 ghorrocks April 18, 2011 18:23

Good idea Lance. But my comment about extracting orthogonal lines on the surface is probably easier as then you can do simple linear gradients along the lines as there is no geometric cross terms. But on the other hand my suggestion will reduce resolution.

So there you go, two ways of doing it.

 az_f June 5, 2011 21:42

Hi
I was wondering can we use Poisson equationg in cfx-pre for additional variables for calculationg WSS gradient??

 ghorrocks June 6, 2011 07:41

You can, but note that it will not be accurate as curvature of the surface will distort it. If the surfaces are flat (or at least flat enough) then this is a good idea.

 All times are GMT -4. The time now is 16:55.