Grid adaption
 

Hi

I am modeling a drop in an immiscible liquid-liquid system using VOF.

I want to adapt the grid just around the interface, but doesn't want to use the predefine function in the menu (Is hard to control the exact position using gradient and using region the cell number become to big).

Can a write an own UDF to mark cells in area just around the interface?

//Johan

Re: Grid adaption
 

hi,

I am trying a liqiud drop falling in gas. What I did is I took a small domain and made quite a fine grid there using periodic boundary conditions on both coordinates. Isn't your problem unsteady? If so,then this you may try, or maybe try to use dynamic mesh, which I do not have much idea.

A question: Do you expect your drop to stay spherical? Are your velocity fields are as expected?

Because my drops are getting deformed, in the regions where I expect them to stay spherical from experiments and literature.

Re: Grid adaption
 

Hi

Yes, my problem is unsteady. I use an udf to calculate the drop velocity; this velocity is used as boundary conditions, so the drop stands still. The problems is that I can't make a fine grid in the whole domain, the number of cells becomes to may = to long simulation time.

The drop is deformed as expected. Both the shape and the velocity are close to experimental values.

Re: Grid adaption
 

Can you explain a bit more, how you write UDF to calculate the drop velocity? It is not so easy for a system with moving interphase, I guess. Also, a bit more about your mesh size & drop size, and your boundary conditions, turbulence model,..? Are your velocity fields in and out of the drop are satisfying? (I guess you should also have some circulations inside your drop, as in my case).

An idea for your case: Have you tried to adap your interphase with isovalue of density= (density phase1 + density phase 2)/2, so that just the interphase region is adapted?

Re: Grid adaption
 

or better to adapt isovalue of phases-volume fraction like min:0.0001 & max:0.9999 for anyone of the two phases

Re: Grid adaption
 

The model is laminar, drop radius around 1.5 mm. Around 300 cells inside the drop in 2D before adaption.

The velocity fields are satisfying, internal circulation with stagnant zones in the front and end of the drop and a stagnant zone behind the drop.

The udf use a loop over the dispersed face, calculating the volume-weighted mean-velocity for the drop which is used as boundary condition.

Thanks for the idea using isovalue, I have looked a little bit at it but have to look more.