Flow through porous media: permeability issue
I'm modeling a two-phase (gas-water) flow through porous media (laminar flow). I calculated the permeability with the Blake-Kozeny equation.
First there is only gas in the porous media which should be replaced by water, so far so good.
The problem is: I have to consider two different permeabilities don't I? One for the gas, the other for water. Is there a way to connect the permeability with for example the volume fraction of water? There is no further explanation in the theory guide concerning the permeability in two-phase flow.
Thank you for reading and (hopefully) answering! :)
From what I see of that permeability equation it has factors for the fluid properties so it should work fine for any fluid. This is providing the flow in the gas phase is in the correct flow regime, which for this equation appears to be stokes flow. So it is quite possible that the gas phase is not in this flow regime. In that case you probably need a different permeability model which is accurate enough over all the flow regimes you will be getting.
Ok, my idea was to use a relative permeability like using the Van Genuchten model or a similar one. How does CFX work there in generel? Does it get the information for the viscosity of the respective fluid? (I'm refering to eq. 1.185 in the theory guide, release 14.5)
I'd like to work with the Surface Tension Model. But I think I've made a mistake or it doesn't work very well. First I tested it just for a fluid domain (with an contact angle of 45 and 0 degrees, just in order to see what happens), no porous domain, but the air doesn't get replaced by the water at all.
I read here
a comment from you that previous versions of cfx don't do surface tension very well.
Do you have experience with newer versions?
Thank you for your reply and have nice holidays!
Nothing has changed about my opinion of CFX with surface tension modelling. The problem is not one of accuracy but rather of run time. You have to use diabolically small time steps to get CFX to converge accurately with surface tension so simulations take forever to run. The results are pretty good for accuracy and fidelity, the problem is just run time. Other codes such as Fluent run far faster for the same accuracy.
|All times are GMT -4. The time now is 17:10.|