CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   OpenFOAM (https://www.cfd-online.com/Forums/openfoam/)
-   -   Correct boundary condition (https://www.cfd-online.com/Forums/openfoam/92940-correct-boundary-condition.html)

Andrea_85 September 29, 2011 13:04

Correct boundary condition
 
Hi all,
I would like to perform this simulation using interFoam:
1) Inject a fluid in a certain domain (a box for example). My boundary conditon are (i guess standard BC for this case):

alpha1
inlet fixedValue
outlet inletOutlet

U
inlet fixedValue
outlet zeroGradient

p_rgh
inlet zeroGradient
outlet fixedValue

left and right simmetryPlane

2) After a certain period of time i would like to stop the injection and let the system relaxes. Now, i'm quite confused about which boundary condition for alpha1 U and p_rgh is better to use at the inlet. I do not want to put simply a "wall" boundary condition, because i guess it is not physical so, in your opinion, which are the correct boundary conditions to have no fluid that comes in from the inlet?

Then i have a second question. Is possibile with openFoam to switch (during the same simulation) from a boundary condition to an other?


thanks a lot

andrea

nimasam September 29, 2011 15:07

1) why treating the inlet in relaxation time like a wall is not physical?

2) Are you familiar with groovyBC ? to implement ur boundary easily ;)

YJ Lee September 30, 2011 03:57

Hi,
Perhaps you can extend the inlet section to have an inflow pipe/duct of reasonable length. Then during relaxation time, apply wall boundary condition in the pipe/duct which is equivalent to pipe/duct closing?
Then there will be reasonable interaction between the domain of interest and the inlet section during relaxation.

Andrea_85 September 30, 2011 04:06

Hi, and thanks for the answer.
1)Because physically there isn't a wall (if you think for example an experiment). simply i want no flow that comes in. Put a "wall" boundary condition is the only choise i have? Something like:

alpha1: zeroGradient
U: fixedValue (0 0 0)
p_rgh: zeroGradient or buoyantyPressure ??

2)no, i'm not familiar with groovyBC, but if you are saying that you can easily implement a sort of "switching BC", i wiil give a look for sure.

Best

andrea

Andrea_85 September 30, 2011 04:25

Another idea could be to impose a "correct" pressure gradient between the inlet and the outlet and then to simulate until nothing moves (until there is no flow from the inlet). Is possibile with OF?

andrea

nimasam September 30, 2011 06:02

Quote:

Originally Posted by Andrea_85 (Post 326194)
Hi, and thanks for the answer.
1)Because physically there isn't a wall (if you think for example an experiment). simply i want no flow that comes in. Put a "wall" boundary condition is the only choise i have? Something like:

alpha1: zeroGradient
U: fixedValue (0 0 0)
p_rgh: zeroGradient or buoyantyPressure ??

i feel you can consider for U zeroGradient ;) or outletInlet condition

Quote:

2)no, i'm not familiar with groovyBC, but if you are saying that you can easily implement a sort of "switching BC", i wiil give a look for sure.
look in wiki and froum you can find many examples using it ;)

Andrea_85 September 30, 2011 09:34

Hi,
i found the "timeVaryingUniformFixedValue" BC, which is probably what i was looking for (it seems easier than groovyBC). I set
alpha1 inlet inletOutlet, inletValue 1, value 1
outlet inletOutlet , inletValue 0, value 0

p_rgh inlet zeroGradient
outlet fixedValue

U inlet timeVaryingUniformFixedValue
outlet zeroGradient

and then i used the "ramp" file to switch between fixed velocity and zero velocity at the inlet. The inletOutlet BC at the inlet for alpha1 should ensure fixedValue =1 when the velocity is non-zero and zeroGradient when the velocity is zero. It seems to work.
thanks again

andrea


All times are GMT -4. The time now is 05:40.