[cp703]

Hi. In an attempt to determine how well ANSYS/CFX deals with scenarios were surface tension effects play a significant role in fluid flow, my adviser has suggested that I first try to model a large droplet of water on a surface and vary the contact angle. To do this, I have scoured the ANSYS tutorials and the internet (to be honest, most of the useful information has come from the cfd-online.com forums...). I created an air/water fluid domain (in the shape of a rectangular prism) with the walls all having a no slip wall boundary condition. I defined an initialized water volume fraction in the shape of a sphere near the "bottom" (in terms of the direction of gravity g in the -y direction) of the domain. After including all the surface tension effects (including specifying the wall contact angle in the wall BC), there doesn't seem to be any difference in the shape of the droplet when I change the wall contact angle.

After investigating problems that people have had with free surface models and taking into account surface tension, etc. I feel like I've covered my bases: I have the multiphase free surface model turned on and the free surface model turned "on", the surface tension coefficient specified for water in air, the surface tension model set to "continuum surface force", turned on and then off the volume fraction smoothing, changed the curvature under relaxation factor, etc.

I really feel like I am missing something. Is this a silly thing to try in CFX? Am I going about it totally wrong?

[ghorrocks]

I have done this exact same benchmark study with CFX years ago. The wall contact angle will work when the solver is setup properly. I cannot recall the exact settings I used, but I do remember that I systematically tried EVERY multiphase option available to find out what everything did. There was a few non-default options which helped.

One thing I can remember is that steady state simulations didi not work. It really needed to be transient.

But the main thing I remember is that this exact simulation runs faster and more accurately under Fluent, so I moved to fluent. How much faster? About 20 times faster. For this type of simulation CFX really is miles behind.

[cp703]

Wow...I'm not sure if I'm relieved that it's not just me who has struggled with this or embarrassed that it took me so long to just post my problem. If you don't mind, I have a follow-up question for you, ghorrocks. Would you say that the inability of CFX to handle this sort of thing would extend to flow in very small, open channels? That's ultimately what I'm working toward, but if a switch to Fluent would be advantageous I'd rather do it now than later (though I've never used it at all).

Thanks for your insight!

[ghorrocks]

My work in this area was for MEMS (inkjet printers to be precise). But the general principle is applicable to all free surface flows with low Re number. So I suspect your small open channel will qualify for that.

Make sure you use the correct words. CFX can handle this type of simulation. It is just much slower than other simulation codes for equivalent accuracy.

And yes, even though I am a CFX disciple I will recommend you move to Fluent. Also Flow3D has some interesting abilities in this area, that is worth a look too.

Are you looking for high accuracy? Or do you just want a general idea of what the flow does? How complex is the flow you want to model?

[Kartikpatel]

How to do the simulation of this kind of condition? If you have any relevant material then please give me.

[ghorrocks]

Here's a good FAQ to start you off: http://www.cfd-online.com/Wiki/Ansys...ible_answer.3F