[ash-khan]

Hi all,

i need some help regarding modeling mutiphase flows in fluent.

first of all i am new to fluent (three months experience). i have a task to model a bubbly flow inside a closed duct entering from two concentric inlets. what i have done uptill now is
1. geometry in gambit
2. incompressible model
a. unsteady, laminar, [pri phase (water), sec phase (air) and vice versa ], non-iterative time stepping, geo-reconstruct

3. Compressible case
a. unsteady, laminar, energy eq, [pri phase (water), sec phase (air-ideal gas) and vice versa ], non-iterative time stepping, geo-reconstruct

in all the cases i tried, my solution divege ( global courant number > 250). i lowered the time step but good. i played around with velocities and pressures but again no good.

i think i am doing something wrong somewhere, which i can not figured out with my limited knowledge. i have consulted the fluent user manual for compressible and VOF portion but cant get any help from there as well.

thanking u in anticipation for ur help

[CFDtoy]

Hi there,
could you post a snapshot of the geometry and let us know what you want to model. If you want to capture sharp interfaces, VOF is applicable. however, you want to predict mere mixing process, eulerian multifluid model may be applicable. However, having said that, case of divergence with VOF on most occasions is due to bad settings. So, lets investigate that part.

Thanks,

CFDtoy

Originally Posted by ash-khan

Hi all,

i need some help regarding modeling mutiphase flows in fluent.

first of all i am new to fluent (three months experience). i have a task to model a bubbly flow inside a closed duct entering from two concentric inlets. what i have done uptill now is
1. geometry in gambit
2. incompressible model
a. unsteady, laminar, [pri phase (water), sec phase (air) and vice versa ], non-iterative time stepping, geo-reconstruct

3. Compressible case
a. unsteady, laminar, energy eq, [pri phase (water), sec phase (air-ideal gas) and vice versa ], non-iterative time stepping, geo-reconstruct

in all the cases i tried, my solution divege ( global courant number > 250). i lowered the time step but good. i played around with velocities and pressures but again no good.

i think i am doing something wrong somewhere, which i can not figured out with my limited knowledge. i have consulted the fluent user manual for compressible and VOF portion but cant get any help from there as well.

thanking u in anticipation for ur help

[eers]

I guess you have to decrease the time step size further..Depending upon the the problem you can as low as the 7th order.

Make sure that your courant number satisfies the equation:


Cu= U*delta x/delta t

regards

eers

[ash-khan]

Hi,
thanks for ur replies.

the snapshot of geometry is attached. Its a simple geometry, rectangular closed channel with air coming in the center of inlet and water entering from outer side of inlet.

I have decreased the time step to as small as 1e-16, but still diverge.
Can grid be a problem or initialization of the domain may be a problematic area. as i have initialized the domain with velocity of water.

[CFDtoy]

are you injecting water and air in a coaxial manner into the system? did i get the problem correct? Actually from the picture, i cant get much. When we say post your problem as a picture, use a schematic ...and avoid grids since they blur the actual inlet outlets and not clear. Anyways, if the problem is VOF with water and air injected - post your solution controls and setup here. Like implicit solver ...surfacetension etc...the way you set up the model.

1. by the way, just try injecting pure water / air into the system and see if it runs ok . If there is any grid related problem, it should be clear from here.

2. Also, what is the domain resolution and length scales involved here?

3. Are you running turbulence or laminar?

4. if (1) works, then run a multiphase model with eulerian two fluid approach to see if it works ok. if so, then we can conclude its the settings in VOF that is causing u the problem.

Thanks,

CFDtoy

Originally Posted by ash-khan

Hi,
thanks for ur replies.

the snapshot of geometry is attached. Its a simple geometry, rectangular closed channel with air coming in the center of inlet and water entering from outer side of inlet.

I have decreased the time step to as small as 1e-16, but still diverge.
Can grid be a problem or initialization of the domain may be a problematic area. as i have initialized the domain with velocity of water.