CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   2 Phase flow simulation in a T-junction microchannel (http://www.cfd-online.com/Forums/fluent/126280-2-phase-flow-simulation-t-junction-microchannel.html)

prohackrav November 13, 2013 04:11

2 Phase flow simulation in a T-junction microchannel
 
1 Attachment(s)
Hi. I am doing a transient laminar flow simulation to get the slug flow of 2 phase immiscible liquids in a t junction micro channel using VOF model. Its a simple 2d geometry and surface tension modelling is enabled. Whenever my timestep advances I get an error stating that the global COurant number is greater than 250. I reduced the time step to values as low as 1e-5. Still the problem persists. I also reduced the UR factors and increased the iterations per time step. (Pressure-0.2, Density-0.5, Body forces-0.5 Momentum- 0.3 As seen in some forum). The velocity seems to be fine at time 0 but drops to zero as soon as the timestep advances. I have included the snapshot. Can somebody help me?

http://postimg.org/image/k7mjf5k71/
http://postimg.org/image/k7mjf5k71/

flotus1 November 13, 2013 04:36

Did you estimate the time step size necessary to achieve a Courant number below 250?
When you say "micro channel" and the velocity magnitude is around 2.5 m/s, I estimate a time step size in the order of 10^{-6} - 10^{-7} to meet this requirement.

prohackrav November 14, 2013 01:01

Quote:

Originally Posted by flotus1 (Post 461776)
Did you estimate the time step size necessary to achieve a Courant number below 250?
When you say "micro channel" and the velocity magnitude is around 2.5 m/s, I estimate a time step size in the order of 10^{-6} - 10^{-7} to meet this requirement.

Thanks for the reply. Its 0.8mm microchannel t-junction. I have two opposite inlets one of kerosene and other water. Both with velocity ranging from 0.02-0.1 m/sec. So what would be the estimated time step size?

flotus1 November 14, 2013 04:20

The Courant number is defined as

C = \frac{u \, \Delta t}{\Delta x} thus \Delta t = \frac{C \, \Delta x}{u}

where \Delta x is the size of a cell. Just to make sure, decrease the time step size further and see if the error persists.
Maybe the problem arises from the initial conditions. Did you try to initialize each half of the domain with the substance at the inlet?

prohackrav November 21, 2013 01:55

Quote:

Originally Posted by flotus1 (Post 461914)
The Courant number is defined as

C = \frac{u \, \Delta t}{\Delta x} thus \Delta t = \frac{C \, \Delta x}{u}

where \Delta x is the size of a cell. Just to make sure, decrease the time step size further and see if the error persists.
Maybe the problem arises from the initial conditions. Did you try to initialize each half of the domain with the substance at the inlet?

Do you mean to patch both the inlets till the junction with their corresponding inlet fluids? No. I havent tried that. Will coarsening the mesh help for increasing the time step? I did that instead of reducing the time step further from 1e-5.

flotus1 November 21, 2013 09:47

Would it be such a mayor inconvenience to reduce the time step size to lets say 10e-8s just to see if the error is really related to the time step size?

Zaktatir November 21, 2013 17:00

Could you please give more input about your Simulation?

-Reference density
-I think you are using a tracking method (explicit that's why CFL-dependency) Geo-REconstruct
-Are both phases well separated like a normal slug flow simulation or do they interpenetrate each other? Are small bubbles and liquid entrainment expected to occur or not?
-Which interpolation of Pressure are you using?
-Reference Location of the operating Pressure
-Operating Pressure..

amasoudi November 21, 2013 17:53

"Monitoring bubble rising velocity in Fluent"
 
Hi,
Does anybody know how can I monitor the rising velocity of single bubble in fluent?


All times are GMT -4. The time now is 00:23.