CFD Online Discussion Forums

CFD Online Discussion Forums (
-   Main CFD Forum (
-   -   CFD simulation of Three phase separator (

tushar393 December 18, 2009 07:12

CFD simulation of Three phase separator
Dear all

I am working on CFD simulation of Three phase separator using FLUENT, to study the effect of various internals on separation efficiency. I used Multiphase eulerian model for phase modelling, gives mass flow inlet boundary condition for inlet and pressure outlet boundary condition for all three outlets (oil , water and gas) The pressure valus are calculated as hydrostatic gauge pressure. I initiate the solution by patching the domain with varuious phase volume fraction with time step of 0.001 sec.

Now after 120 sec run and time step value of 0.005sec the solution shows continuous fluctuations of outlet mass flow rate at both oil and water outlets from positive to negative values and repeats the profile after about 5000 iterations or 20 sec run time. and solution not stabilizing.
I need to have a stabilise solution for run time of 8 min.

Kindly suggest the ways to approach the stabilised and acceptable convergence solution.

looking forward for response


acampbell December 22, 2009 17:54

I would suggest starting with a smaller model (and maybe even just 2 phases) to ensure that you are achieving adequate convergence around the physics of your phases. Also note that the morphology of the phases (droplets, continuous, sizes etc.) is also important.

I suspect you may even need a slightly smaller timestep/timescale also. Also ensure that you have had enough time for the flow to traverse the system - and what you see is the instabilities working their way through the system.

tushar393 December 22, 2009 23:17

Thanks for the Reply.
I already achieve the converged two phase solution. Here the mass balance getting disturb thats the worry, why is it so i try to figure out.

acampbell December 22, 2009 23:41

I would be testing a smaller timescale, checking the multiphase setup in terms of droplet sizes etc.
Again I would be personally checking the separation on a small model to ensure this is working first.

All times are GMT -4. The time now is 15:17.