[Sripadaraja]

I am currently working on 50um channel where oil and water are passed at different velocity to form water droplets. I am supposed to simulate it for 30 seconds. But, When I enter time step beyond 1ms, solution seems to be diverging. I get the error message : Global Courant Number Greater than 250. Channel dimensions are 50um and 500um in length with 1um element size.


-Sripad

[cfdnewer]

time step too large

[Sripadaraja]

So, what should be the time step in that case?

[raghulvr]

Reduce the courant number and check the solution for convergence.
Can you tell me the values of under-relaxation factors, type of solution scheme you are using?

[Sripadaraja]

Hi Rahul,

Courant number is 0.06. I am using PISO scheme and under relaxation factors are

Pressure - 0.3
Density - 1
Body Forces - 1
Momentum - 0.7


-Sripad

[raghulvr]

Try Coupled scheme and run the simulation with first order accurate for initial 250 iterations and then switch to second order accurate for turbulence and momentum

[Sripadaraja]

Thank you for your quick reply. Turbulence? Laminar flow is expected in micro channels isn't it?

Also, while adopting your recommendations, what is the time step size you recommend for 250 iterations.

[Sripadaraja]

Thank you Rahul for the suggestion. Error Courant number 250 is no longer seen. But solutions seems diverging as the water phase scatters everywhere in the channel.

I Played around with time step and number of time steps and iterations. from time step 0.01 to 60 sec with no of steps 100 - 1000000 having 250 iterations in all cases seems diverging.

Originally Posted by raghulvr

Try Coupled scheme and run the simulation with first order accurate for initial 250 iterations and then switch to second order accurate for turbulence and momentum

[raghulvr]

keep the time step size to 0.005 and set max iterations/time step equivalent to 25.
I'm not too sure about laminar or turbulent flow....it varies from geometry to geometry...are you following a technical paper?, if it is specified as laminar/turbulent, then it's okay.
Another thing what mode of initialization are you using?
Does the solution diverges at the very beginning of your simulation?
Are there any messages on the fluent console that shows "some parameter is limited?"

[Sripadaraja]

Rahul - No luck ya. I am using standard initialization and initializing with oil inlet channel. Yes my simulation diverges in the beginning with no error message in the consol (but I see reverse flow 34 faces on pressure outlet)which I did not see using PISO scheme.

[raghulvr]

can you send me the mesh file and solver settings....my mail id is raghulvr@gmail.com