Learning Forum

[usama]

Dear Ansys Community,

I am working on two-phase flow. 

I am facing the problem of incomplete filling. here I attached some screenshots of the results. the length of the channel is 40mm and the height is 20mm and 50 micrometers respectively. when I reduced height I get an incomplete filling. can you please guide why it is happening and what should I do for complete filling. looking for your suggestions. thanks in advance.

Regards

Usama

but if I reduced height like h=50 micrometer and length=40 mm I am getting incomplete filling as shown in the attached figure.

[DrAmine]

Is it longer?
Pleaae make a good screenshot.

[usama]

Dear Amine i post the sketch of the problem above
I used vof model with laminar flow also
My question is if i reduced the height i am getting incompelet filling
First one height and length is 20mm and 40 mm
Second case height is 50 micrometer and length is 40 mm.
For the second case i am getting incompelet filling.
Regards
Usama

[usama]

The length of the channel can not be captured fully thats why i added the area which is incomplete filled.
Regards
Usama

[DrAmine]

So the larger one is still not filled? Perhaps it takes longer. .

[usama]

What longer if the length is 40mm and speed is 0.02m/s=20mm/s then it takes 2 second to fill the length of the channel. But it is showing incompelet filling the second case.
Regards
Usama

[usama]

The second one on the hand written geometry..

[DrAmine]

As I said your screenshots are not really ideal. Why making pictures from monitor...

Anyway: post here all model and bc settings

[usama]

here I described the case-1 and case-2. with these settings in fluent for both cases. I also attached the archive file if you need any further detail you may check it.

now I show you the results of case 1 and case-2 respective

 

case- 1 shows full-filling but In the second case with the same settings, it does not fill the air area. as shown here.

I also increase the no of iteration per time step and also use different models like the k-epsilon model with standard wall functions and scalable wall functions. but it does not work for me. hope this time you get good screenshots.

I need your suggestion to fix this issue and show complete filling.

 

 

[DrAmine]

No surface tension? Is phase 1 liquid? Make phase 1 air and 2 liquid and use compressible ideal gas for air and turn off energy equation

[usama]

Dear Abenhadj,

I attached the archive file the above post for further understanding. used the above screenshots settings and run the simulations. in order to attach the file,I clear all data in the above-attached file.

Regards

Usama

 

[usama]

I apply surface tension between two phases also it's a normal thing. one should have to set the surface tension between two phases. which is 0.072. i turned off the energy equation as you can see in the above post.

I am asking about settings that I need to put in fluent to get the accurate filling in the channel. 

and does it matter to make phase 1 air and phase 2 liquid?

because while patching i maked air as 0 in the liquid region.

Regards

Usama

[DrAmine]

As ANSYS Employee, I do not download any attachments

[DrAmine]

Try running with surface tension force and without wall adhesion.

 

[usama]

No surface tension? Is phase 1 liquid? "Make phase 1 air and 2 liquid and use compressible ideal gas for air and turn off energy equation"
Dear Abhenhadj in your reply.
1. Does it matter to make phase 1 air and phase 2 liquid?
because while patching i maked air as 0 in the liquid region.
2. I apply surface tension force with out wall adhesion force between two phases.
I think you are not getting my point.
Reply after reading above post as well.
Regards
Usama

[DrAmine]

Be a bit polite and cooperative. I am helping here on voluntary basis. I assuming you are using high quality quadrilateral grid for this simple geometry.

Now (how I would debug this or model filling):

 

1/I recommend to make phase-1 air and phase-2 liquid

2/By doing 1/ you will need to adjust the boundary conditions and the way you patch the case

3/Why 1/ because it is better to make the compressible phase the primary phase. Also use ideal gas for air. In order to avoid solving energy equation one needs to turn it off (turned on automatically after you select ideal gas for air). We want just to have baratropic air

Now run this system

Another point is to turn the system without adhesion or surface tension effects: you asked for help so please help us to help you otherwise that would be my last comment wherever I see messages from you.

[usama]

Dear Abenhadj I was not rude with you i said to you are not getting my point means the things you are suggesting me i already applied after your first comment. Yeah i added all screenshots except the surface tension thats why you point out on that.
I said that is obvious.
May be it seems you rude from my side so sorry for this i just stuck in that. And its my last work of my studies and i have to complete it as well after several tries i put this thing here to get some suggestions which may be i am missing.
Regards
Usama

[usama]

Dear Amine Now come to the point
After your suggestions
1) I picked
air as phase-1
water as phase-2.
VOF model implicit with body force.
Fluid is laminar.
inlet velcoity for mixture =0.025m/s.

Inlet volume fraction for phase-2 =1.

Outlet for mixture as atmospheric pressure.

Outlet for phase-2 : Back flow Volume fraction=0.

Surface tension interaction b/w two phases =0.072.

Specfied operating density =1.225 kg/m3.

Standard Intiallize the problem by keeping the phase-2 volume Fraction=0.

Patching the water region in phase-2 in hexadron =1.

These settings i used. And no filling still.

2. As you pointed "May be you are using high quality grid".
In Reponnse to this:
i used in Meshing body sizing, method-->triangles, Face meshing--> triangles-split.
And mesh them.
3) so you suggest me phase-1 take as air and i did.
4) you also suggest me "avoid surface tension effect and wall adhesion"
So not to click on surface tension effects b/w two phases Right?
I will try the 4th one and update you about this.
I write all settings above if these is some mistake than correct me.
And some more suggestions needed from your side.
Please stay tuned untill i get my results.
Hope you shall oblige me
Kind Regards
Usama

[usama]

Dear Amini I avoid surface tension effects filling increase but not fully filled.
Suggest me about that.
Regards
Usama

[DrAmine]

Can you screenshot the mesh and the residuals of your run?

[usama]

Dear Amine I attched the screen shots above.
And please tell me the settings i used are write?
Regards
Usama

[DrAmine]

We do not download attachments

[usama]

[usama]

Dear Amine,

i attached screenshots may be they are not good but please try to read them.

one more question i want to ask why to ignore surface tension effects in the narrow long channel?

Regards

Usama

[usama]

residual with k-epsilon model with standard wall functions

i just change the model for fluid and here are the residual but same results.

 

 

[DrAmine]

The idea of ignoring surface tension is just for debugging purpose. Your grid is poor. Make quads and increase the number of cells in the height to at least 10 or more.

[usama]

Dear Amine,

i refined this grid more by using refinement-2 also and run the simulations but it is not working. i already done this. any how i try it again and update you.

Regards

Usama

[usama]

@peteroznewman and @rwoolhou can you please suggest to me for complete filling after checking this post.

Regards

Usama 

[Rob]

Amine's on holiday, I finish today too

 

Looking at the set up it's fine. Your problem is the mesh. With VOF you resolve the interface between the two phases. Insufficient mesh means you will fail to do this, and (probably) lose mass which is why you don't consistently see the fluid trapped in the pocket (assuming it's the lighter phase. 

To add, you may need a lot more hex mesh than you think: aim for absolute minimum 6-10 cells in the depth of the pocket and aspect ratio of 1. Time step may also be critical. 

[usama]

I added the gravity in X-direction is it fine because the flow is in X-direction?

and i need to add the inflation layer under the depth of the pocket? with aspect ratio 1? 

[usama]

Dear Peter,

I also added the archive file of the model in the above post.

regards

Usama

[peteroznewman]

Dear Usama,

I am not a CFD expert. The CFD experts are both on holiday until next year.

Kind regards,
Peter