[NoradFirst2]

Hello Foamer,

I am simulating a 2D-system which is quite simple using buoyantPimpleFoam

It's a square. I blow air (fixed T and U) coming from the middle of the top side onto the bottom side which is a wall (fixed T). The right and left side plus the rest of the top side are outlets (fixed P0 - BC on T is zeroGradient).

It worked.

Then I removed the outlet parts of the top side to replace them by suction holes (fixed P<P0). These holes are sucking air from my system but also from the outlet (left and right sides) which is normal. The thing is, this air coming from outlet parts is at 0K (because OF doesnt know its temperature I guess) which causes some problems...

Is there a BC which say "I don't know the temperature on this side but if you need to suck air from this side take 298K or internalvalue or current value on the border or whatever but not 0K...". A kind of inletOutlet BC which is fixedValue 298K or zeroGradient depending on the direction of U?

I'll post a drawing for helping you understanding my geometry.

Thanks

[NoradFirst2]

Here is a picture.

I've runned a few time step with fixedValue for T at the outlet for you to see the picture but I'd like to change this BC.

[jherb]

Does inletOutlet for the temperature work?
See http://www.cfd-online.com/Forums/ope...tml#post261573
It should also work for scalar fields.

[NoradFirst2]

I tryed for T :

outlet
{
type inletOutlet;
value $internalField;
inletValue $internalField;
}

It may work, U is now creating problem after a few time step, I have huge velocity coming from the outlet and going to suction.

So I am now trying to create the suction thanks to fixedFlowRate rather than given pressure drop.

Working on it!

[NoradFirst2]

I turned off the suction from the outter holes.

I changed the suction based on pressure drop by suction based on flow rate using FlowRateInletVelocity.

I still have problem because velocity coming from the left/right has huge increase sometimes and it makes the solver stop. I am expecting velocity to be around 30m/s in the suction hole (Bernouilli + Exp data)

Here is a screenshot of the last timestep I have computed.

Do you have any idea of how to prevent this problem?

Increase the size of the domain maybe?

BTW : I have two suction holes here but only one BC for these two holes. I should consider the total flowRate for my BC (FlowRateInletVelocity.) and not the flowRate per hole right?

[pici]

Originally Posted by NoradFirst2

I turned off the suction from the outter holes.

I changed the suction based on pressure drop by suction based on flow rate using FlowRateInletVelocity.

I still have problem because velocity coming from the left/right has huge increase sometimes and it makes the solver stop. I am expecting velocity to be around 30m/s in the suction hole (Bernouilli + Exp data)
Here is a screenshot of the last timestep I have computed.
Do you have any idea of how to prevent this problem?
Increase the size of the domain maybe?
BTW : I have two suction holes here but only one BC for these two holes. I should consider the total flowRate for my BC (FlowRateInletVelocity.) and not the flowRate per hole right?

Hi James,
I have exactly the same problem trying to solve my case, except that T for me is neglected.
I did a lot of tentative with suction based on flow rate using FlowRateInletVelocity, but no results; more, I have same strange U and P images like you, in geometrical zones with apparently no problems.
My project is a medical device that work with natural inhalation of air, so pressure based suction should be the best, because I know the depression that a patient (usually an old women/man) can apply.
But, seeing to your results, I'm not so confident of this method.

To understand, I tried to use the case tutorial/icoFoam/elbow and naïvely I changed the sign of inlet speed, leaving all other 0/p and 0/U unchanged: solution is terrible and software stops before to finish. Patches names too are unchanged, so patch named inlet is actually a physical outlet.
The small inlet seams to be 'closed up' (as in images attached): so it seems that boundary conditions are globally wrong (or openFOAM does it have some problem in suction cases?)
I really will appreciate your (or any other) help to find right BC
Bye

Piero

[NoradFirst2]

I changed my geometry to prevent the problem from happening.

In the end I used the BCs for EACH of my suction hole, no common BC :

for U :
{
type flowRateInletVelocity;
flowRate constant -1200;
value uniform (0 0 0);
}

for p : zeroGradient

Maybe try those BC as you can't use my trick for the geometry.

But I think you should be able to use velocity inlet and pressure outlet for your system.

[pici]

Hi James,
I have tried your suggestion, and I did some short test using tutorial/icoFoam/elbow modified, and now it works.
Truly, I still have some problem, at first because the elbow mesh is poor, and also because in fvSchemes I'm not sure which is the best method:

divSchemes
{
    default         none;
    div(phi,U)      Gauss upwind;//linear;limitedLinearV 1;
}

upwind seems to work better then others, but I'm not sure.
Now, I will export this solution method to my true case.
And I will cross my fingers
Thank you a lot

Piero