Learning Forum

[BillT]

Hello,

I try to model the gas-liquid flow in structured packing according to the model of Soulaine. The superficial velocity of each phase is estimated using a phase Darcy's law and the pressure gradient is related to this velocity by a permeability tensor K (see below). I can't figure out how to implement this tensor or rather the equation in Fluent. I wanted to write UDF source terms for the momentum balances, but there is no inverse matrix for K of the liquid phases, so i can't get an equation for delta_p.   

Any suggestions how to model/implement these equations?

Thank you in advance!

 

 

[Amine Ben Hadj Ali]

The greek Letters are referring to directions. Which ones?

[BillT]

The angle in the cosine and sinus functions is related to the liquid film flow along an inclined plane.

[Amine Ben Hadj Ali]

I am talking about gamma and the betas?

[Amine Ben Hadj Ali]

As your angle is constant just use the porous media model in Fluent and accommodate the direction vectors to the ones corresponding to your case. Or rotate to have everything aligned with Cartesian coordinate and just adjust the gravity direction.

[BillT]

the betas specify the two separate liquid films in their preferential direction. Thank you for your advice. I will try to change the direction vectors tomorrow.

 

[Keyur Kanade]

Yes, the best way is to rotate model in cartesian coordinate. In Fluent, you can use, Setting up domain - mesh - transform - rotate. 

[BillT]

If I want no flow in one direction (e.g. z-direction), can I model this with the porous media conditions or should I set the velocity in this direction to zero with fixed values?

[Amine Ben Hadj Ali]

Both methods are okay but I prefer imposing a resistance then assuming a fixed velocity of zero. If you zero velocity in one direction why not using 2d solver?

[BillT]

 I solve it as 3D because I want to simulate multiple packings which are rotated around the y-axis by 90 degrees. I tried to set the direction vectors for the porous media conditions and noticed that my vectors are not normal to each other. Am I missing something here?

[Amine Ben Hadj Ali]

The vectors are your input. We recommended to rotate your model so that you can just give the axis as vectors.