[camposrf]

I am trying to make a conditional expression with "if" command in CFX-10.0

I read these articles:
Conditional Expression in CFX Pre

"if" command in CFX-13 is not really an "if" command!!!!

But I do not know how to do it in CFX pre 10.0

I saw in youtube that "if" expression works on CFX 14, but I do not have it (my PC is old)
https://www.youtube.com/watch?v=GJXQNhmqyhE
(11:25)

The string that I want to use is for define the volume fraction in a water tank:
WaterVolFrac =if(y<WaterHt,1,0)*if(y>-0.1[ft],1,0)

This is the error message:
__________________________________________

ERROR
Error in setting: "WaterVolFrac" via the expression:
if(y<WaterHt,1,0)*if(y>-0.1[ft],1,0)
Error found at the start of the line, at item: if
unrecognised name
__________________________________________


I read the _CFXTutorial.book.pdf, Tutorial_FreeSurfaceBump.pdf but they don't use the if command, they use this string:
DownVFAir = step((y-DownH)/1[m])

[ghorrocks]

The if statement was introduced in a "recent" version of CFX. Not sure which one, maybe CFX 12, but it looks like CFX 10 did not have it. Sounds like a good time to upgrade to the current version of CFX to me.

[camposrf]

I just adapted the original string to my case "WaterVolFrac =if(y<WaterHt,1,0)*if(y>-0.1[ft],1,0)". The function plot works and makes sense.



But I still do not knowing how "the step commad" works. On the other hand, the results in the post environment says that the tank is full of water until the top!...



I was waiting something like this:

[ghorrocks]

The step command is very simple, and is explained in the documentation (CFX Reference manual). If the argument is positive, it returns +1, if negative it returns 0, if exactly zero it returns 0.5.

[camposrf]

Originally Posted by ghorrocks

(...) If the argument is positive, it returns +1, if negative it returns 0, if exactly zero it returns 0.5. (...)

Thx ghorrocks,

Your comment was very useful, now I can adapt this string for make a synthetic "if" command if "step" returns +1 0 or 0.5



The problem with my simulation was not the step command. The mistake was in the vertical definition (Expressions: UpH, UpVFair) as you can see in my second comment in this post; the vertical variable was "Y" but the correct one is "Z".

Here you can see the results with the corrections applied.

Water Volume Fraction profile:




Water Volume fraction surface

Water Volume Fraction profile and surface:

[ghorrocks]

Please do not post multiple versions of the same post on different threads.

Your surface looks very lumpy and rough. I assume it is not meant to be like this.
* You probably have too coarse a mesh. Have you done a mesh sensitivity study?
* For the initial condition you should blur the free surface over a small distance so irregularities in the mesh do not distort the surface. Have a look in the CFX documentation under best practices guide for multiphase for tips on this.

[camposrf]

Originally Posted by ghorrocks

Please do not post multiple versions of the same post on different threads.(...)

I do not want to do that, I was just trying to help

Originally Posted by camposrf

I had the same issue with the interface. I defined the volume fraction (...)

Originally Posted by ghorrocks

(...) Your surface looks very lumpy and rough. I assume it is not meant to be like this. (...)

Yes, you are right.

Originally Posted by ghorrocks

(...)* You probably have too coarse a mesh. (...)

Yes, right again, it is a coarsed one. I have to do approximation cases first before run a detailed model “if I want the results this year” (joke). Further more, it has a worst residuals convergence (0.01). The good news is that the time step is ok because it is a steady simulation.
I have to do this previous work because I do not have a powerful PC. I accept donations (is not a joke).

Originally Posted by ghorrocks

(...) Have you done a mesh sensitivity study? (...)

No, I have not done. I think that my PC can not handle that. I do not know how I can do that. I did not find the FAQ with the instructions.

Originally Posted by ghorrocks

(...) Anyone - we really need to write a FAQ on the basics of checking mesh, timestep and residuals convergence and sensitivity for these type of basic questions. 90% of the "my simulation is not accurate" questions do not do the basics here - any takers?

Glenn Horrocks

Originally Posted by ghorrocks

(...)* For the initial condition you should blur the free surface over a small distance so irregularities in the mesh do not distort the surface. Have a look in the CFX documentation under best practices guide for multiphase for tips on this.

Thanks for your comment, I will check it.

[camposrf]

This is the mesh that I used, is not a detailed one but works for preliminary runs.

[camposrf]

Originally Posted by camposrf

(...) I did not find the FAQ with the instructions. (...)

Finally I found the FAQ, but searching from an external search engine.

Originally Posted by ghorrocks

There is an FAQ about linear solver overflow now: http://www.cfd-online.com/Wiki/Ansys...do_about_it.3F

https://www.cfd-online.com/Wiki/Ansy..._inaccurate.3F