CFD Online Discussion Forums

CFD Online Discussion Forums (
-   FLUENT (
-   -   Mixture model problem - could someone please advise? (

matlab_monkey July 25, 2012 12:11

Mixture model problem - could someone please advise?
Hi Guys,

I'm trying to put together a two phase model using the mixture model, but have run into some difficulties. I've gone right back to basics and constructed a simpllified problem, I thought this may help me work out where I'm going wrong.

Unfortunately, I'm getting nowhere fast, could someone please offer some advice?

My simplified problem consists of a 2D room. The room 10m long and 2.5m tall. On the left wall there is an inlet on the centre of the wall (flush with the wall) which is 0.2m in height. On the right hand wall I have an outlet which is 0.5m in height. The outlet is 0.75m away from the top edge of the domain.

My mesh appears good, I've done the quality checks etc and don't see any problems there.

For this initial problem I have effectively switched off the secondary phase (setting inlet volume fraction to 0). When I have a satisfactory flow established with a single phase then I'll add in the second phase.

I am using the k-epsilon turbulence model, with default parameters for now.

My inlet (mixture) conditions are a turbulent intensity of 5% and hydraulic diameter of 0.2 (inlet diameter). The air phase enters the room at 1m/s. The outlet condition is a pressure outlet. I have specified a turbulent intensity of 5% and hydraulic diameter of 0.5 to account for any backflow.

When setting my operating conditions I activate gravity (y=-9.81) and (as done in the Ansys multi-phase tutorial) set the operating density to be 0.

When I run the calculation I get reversed flow at the outlet - which doesn't appear physical. In an attempt to avoid this I modified the domain, to add a 'pipe' leading to the outlet. I thought this may give the flow some time to develop before reaching the outlet edge, without altering the flow within the room. Unfortunately this doesn't fix the problem.

If I plot contours of static pressure then I find (almost) horizontal stripes, as expected the pressure is higher at the floor than the ceiling (i.e. hydrostatic).

If I switch off gravity (or the specified density) then I lose the reversed flow but my pressure field is no longer hydrostatic.

My intention is to use the mixing model to simulate heavy particles dropping out of an air stream. But first I need to get this simple problem fixed.

I've only been using Fluent for a few weeks and must admit that I'm a little bit confused over the way that Fluent handles pressures and densities. I suspect that this is somehow responsible for the reversed flow I'm seeing at my pressure outlet.

I should point out that the flow seems to converge well. There doesn't seem to be any issues with the solver settings.

Could someone please advise me how to fix this problem?



Marion July 26, 2012 07:27

How long have you run the case?
When you say it is converged, do you wait until the residuals are very low and stable (as opposed to reached 10-5 or something like that)?
Sometimes you can have reversed flow at the outlet at the bieginning of the run and then it disappears after more iterations.
It is strange that the reverse outlet flow doesn't disappear with the pipe at the outlet.. except if it is not long enough, or if the model has not really converged.

matlab_monkey July 26, 2012 08:20

Hi Marion,

Thanks for your response.

The calculation is converging in the sense that the residuals for continuity are small and all of the residuals are very stable (give or take a small wiggle). The backflow is still present at the outlet after many iterations.

I think it must be something to do with the operating conditions. If I uncheck the specified density option then the reverse flow disappears (after just a few iterations) but I lose the stripey (hydrostatic) static-pressure field.



All times are GMT -4. The time now is 15:01.