Hi !

I am trying to calculate the average streamwise velocity in my flow. CFX guys adviced me to write an expression in CFX-Build e.g.

uaver=volumeAve(u)@fsub (fsub is my subdomain name)

The problem with this is, that the return values are something else then I expected and they are not consistent with the expression

uavar=volumeAve(Velocity u)@fsub

in CFX-Post. Does anyone has a idea what is wrong and how should I get my streamwise velocity average.

Thanks, Andrej

Are you working in conservative values?

Hi Holidays !

For faster communication I would rather have your e-mail address.

Till now I haven't discover conservative/non-conservative switch in CFX5 solver.

I think that incompressible calculation was selected somewhere, but I cannot find where.

Nevertheless, this should not effect expression volumeAve()@fsub. For example volumeAve(v)@fsub calculates 10000 time higher value than volumeAve(Velocity v)@fsub in CFX-Post; volumeAve(u)@fsub returns 48 per cent of volumeAve(velocity u)@fsub value.

Andrej

Hi Andrej,

It will affect your calculation in Post. Hybrid values use the value at the wall for the wall node instead of the near wall control volume (which you get if you choose "conservative").

Robin

Hi !

Hybrid/Conservative switch in CFX-post changes values for max 10 per cent, but not for factor of 10000.

Andrej

With such a discrepancy you must be able to suspect one or the other to be the faulty expression. Which one is it?

Hi !

Apparently I discover a new bug; Is there a bug list for CFX programs ? I don't like to spend much time on mistakes of others.

CFX 5.5 is not able to calculate correct volumetric average for velocity components - something has to be wrong with integration. Expressions:

volumeAve(u)@domain and volumeInt(u)@domain do not produce right values. The same is for v and w.

I did't have any problem calculating pressure (p) volumetric average (volumeAve) or areal average (areaAve) for velocity components (u,v,w).

It is just velocity volumetric average or integration which I found WRONG.

Andrej

P.S. Maybe CFX 5.5.1 doesn't have this problem any more.

"P.S. Maybe CFX 5.5.1 doesn't have this problem any more. "

...or 5.6 Beta?

So, to get this straight. You calculated an expression in the solver and the same expression in Post and the result is different?

How were you reporting the result of the expression from the solver?

Also, what units is the solver running in and what units are you evaluating to in Post?

Robin

I don't think this is in the solver, Andrej is saying that volumeAve(Velocity <u|v|w>)@location is giving different results that volumeAve(<u|v|w>)@location, in CFX-Post. Perhaps whoever works on CFX-Post should be informed of this, if it's true and can be reproduced.

Neale.

Yes, I can see that from he previous post. What I would like to know is where he is calling this in the solver and how it is being reported.

Robin