Learning Forum

[Anthony Bowers]

Hello all,

I am seeking guidance in my simulation.

Context: I am simulating a gas-liquid (countercurrent) in a complex geometry. I have a mass transfer term that occurs at the interface. The magnitude of the volume fraction gradient dictates which cells enable mass transfer. My only concern is the mass transfer coefficient. In literature, authors use a theoretical contact time that uses a geometric length of the complex geometry and Nusslet laminar effective velocity; see below. 



While these relations work great for air/water systems, I am simulating a liquid metal with a contact angle and surface tensions 2x and 7x higher than air/water, respectively. Experimentally, the liquid mass transfer is on the order of 10e-5, but with this theoretical method, I am getting a 10-3 value. 

Therefore, I'd like to compare the actual contact time through Ansys fluent and see if there is a difference. 

I understand that the phases share a velocity profile in a VOF simulation. Still, I am curious if it is possible to extract the profile for alpha_< 0.5 and volume average for an average velocity. Then to determine the contact length I would create an iso-surface of the phase volume fraction (alpha_L > 0.5) and determine this length. However, I am uncertain how to take on these tasks in a UDF. 

-Thanks

[Rob]

The example DEFINE_LINEARIZED_MASS_TRANSFER looks to cover the volume fraction (alpha) term. The coefficients are then defined or calculated within the UDF code. 

[Anthony Bowers]

 

I am using DEFINE_MASS_TRANSFER. Not the linearized version. 
How would the linearized method get me liquid mass transfer coefficient ?

[Rob]

Have a look at the definitions in the example. You define the coefficient based on "stuff", Fluent then finds the overall rate. 

[Anthony Bowers]

 

Hmm, I just read it and found an example.. I don’t quite understand how this is different from MASS_TRANSFER udf . 

If you have time and possible could you explain ? 

Also the solubility it highly uncertain in my simulation. So, we are attempting to find Kla to evaluate what the solubility value is experimental. 

 

[Rob]

Basically adding in the extra terms aids in solve stability. We aren't able to cover UDFs in the same detail so we're limited to the documentation; albeit we may know where to look. 

That sounds like you're going to have problems find the rates, but.... And treading carefully to remain using public domain knowledge. Mass transfer coeficients generally come down to solubility, and concentration gradients. If you very gently mix both phases in a tank/bottle can you find the rate to then back out a number for Fluent to use in the complex system? 

[Anthony Bowers]

Okay I understand. 

Is there a way, with either method, for me to obtain an interface length using a UDF?