CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLOW-3D (http://www.cfd-online.com/Forums/flow-3d/)
-   -   Can we use Dynamic Mesh in VoF (http://www.cfd-online.com/Forums/flow-3d/70868-can-we-use-dynamic-mesh-vof.html)

Tygrys December 8, 2009 13:07

Can we use Dynamic Mesh in VoF
 
Hi everyone,
I try to solve problem with droplet break-up.
To solve this case I used volume of fluid methods. When I was using 3D case I had problem with to big grid. I would like to ask if it possible to concentrate grid point near the droplet. It is possible to use dynamic mesh or reference frame to VoF method? Or maybe could You recommend another method?

Thank you for any advices.
Best regards
Lukasz

jemyungcha December 12, 2009 08:59

Re
 
Lukasz

sorry I don't fully understand what you point is.

:-)


it needs a number of grid for droplet breakup analysis using VOF.

(flow module + VOF = unsteady problem. and grid quality and the number affect solution)


You told about "concentration near the droplet".......right?

But think about it. Behavior of droplet not static but dynamic.

Using the dynamics mesh hardly be applied to this kind of problem.

"Dynamic" means that mesh should moves with droplet's motion. I am not sure whether it affect the solution of this problem.

If i were you, i would not use this function.


The only solution for good analysis result is (My own thought)

high performance computer and parallel processing.


I am using another commercial CFD code (CFD-ACE+) and quad core (intel i5) pc.

but i am not satisfied its performance. too time consuming work.

about 80,000 cells brought me 3~4 days per single analysis (3D analysis of droplet breakup in the microchannel).

I want to hear other researcher's experience about CFD.

Jemyung

Tygrys December 14, 2009 11:36

Sorry for my pure English language.
You had correctly understood my problem. Recently I use parallel processing (4 CPU), but my case have almost 800 000 cells. So I need 2-3 weeks to solve it. That is why I try to find less consumption time method.
Lukasz

ck5285 December 16, 2009 05:34

All
I did some benchmark before for Fluent & CFX and Flow-3D in droplet simulation. Flow-3D is much faster in simulation time due to its True-VOF approach vs typical VOF approach that solve both air and fluid in momentum equation. Hope this help. :)

MuxaB December 18, 2009 20:12

Quote:

Originally Posted by Tygrys (Post 239949)
Sorry for my pure English language.
You had correctly understood my problem. Recently I use parallel processing (4 CPU), but my case have almost 800 000 cells. So I need 2-3 weeks to solve it. That is why I try to find less consumption time method.
Lukasz

Lukasz,

Why is it so long? What is the time step size and the total simulation time?

Michael

Tygrys December 20, 2009 19:14

The time step size is 10^-8. Flow time which is needed to droplet break-up it's about 0.002 [s] (for We=20 and grid size about 10^6 cells). Real time to solve this problem is about 3-4 weeks.

Regards
Lukasz

MuxaB December 21, 2009 20:49

Quote:

Originally Posted by Tygrys (Post 240580)
The time step size is 10^-8. Flow time which is needed to droplet break-up it's about 0.002 [s] (for We=20 and grid size about 10^6 cells). Real time to solve this problem is about 3-4 weeks.

Regards
Lukasz

Is the time step size kept down by the surface tension or viscous stability limits? Can you try using implicit approximations for those forces to relax the stability limits on the time step size?

Tygrys December 23, 2009 14:17

I using small time step to capture the oscillation of the droplet interface and the associated high velocity(recommended by Fluent Tutorial Guide). I using implicit approximation for this calculation.

MuxaB December 23, 2009 17:22

Quote:

Originally Posted by Tygrys (Post 240837)
I using small time step to capture the oscillation of the droplet interface and the associated high velocity(recommended by Fluent Tutorial Guide). I using implicit approximation for this calculation.

I am not sure Fluent's advice applies to FLOW-3D simulations, does it?


All times are GMT -4. The time now is 10:14.