CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   OpenFOAM Running, Solving & CFD (http://www.cfd-online.com/Forums/openfoam-solving/)
-   -   Basic boundary condition question (http://www.cfd-online.com/Forums/openfoam-solving/58474-basic-boundary-condition-question.html)

msrinath80 September 17, 2008 15:04

Hi all, Given that the outw
 
Hi all,

Given that the outward drawn unit normal to inlet and outlet boundaries point parallel to the x-direction, are the following the only conditions that need to be satisfied at these boundaries?


For a fixed velocity inlet,

u = fixed value
v = fixed value
(dp/dx) = 0.0


For a fixed pressure outlet,

p = fixed value
du/dx = 0.0
dv/dx = 0.0


Where: In 2D: U = [u,v] (u and v are the two components of velocity vector U) and assume that 'd' represents the partial derivative symbol.

I would appreciate if someone can point out if I am incorrect or missing anything here.

Thanks for your help.

msrinath80 September 18, 2008 13:02

Anyone? Even a simple Yes or N
 
Anyone? Even a simple Yes or No will suffice.

liu September 18, 2008 13:15

For incompressible laminar flo
 
For incompressible laminar flow, Yes

msrinath80 September 19, 2008 10:59

Thanks Xiaofeng.
 
Thanks Xiaofeng.

msrinath80 October 1, 2008 17:05

For a plain channel or pipe fl
 
For a plain channel or pipe flow, dp/dx = 0 at the inlet boundary does not seem to make sense. The pressure should be allowed to change in the X-direction. Can anyone comment on that?

ngj October 1, 2008 17:55

Hi Srinath Yes, I would lik
 
Hi Srinath

Yes, I would like to give my two cents:

Ideally you would say, that the only thing you know for sure is the pressure at the inlet and at the outlet, i.e. the energy gradient. Thus the intuitive way to set the boundary conditions would be to apply the pressure at both ends, which is equivalent of applying a body force. This solution converges asymptotically, thus far to slow to be interesting if you start with initial field 0 (unfortunately I cannot recall the reference on the asymptotic part).

This would give a correct pressure gradient, i.e. constant, along your uniform channel. Another method is to apply the velocity at the inlet and a non-zero pressure gradient, but that implies that you _know_ the solution to the pressure field in advance, thus people have settled with the zero pressure gradient. I have read another article (again, cannot recall the exact reference), which did compare the results between a zeroGradient and the correct gradient solution, and the differences was small, at least from an engineer-point-of-view.

Best regards,

Niels

msrinath80 October 1, 2008 19:56

Thanks very much for that resp
 
Thanks very much for that response Niels. You've brought some interesting facts to my attention. I'll try and dig up more information on this.

mahendra October 14, 2008 00:25

Hello People, I am have a s
 
Hello People,

I am have a small doubt regarding pressure boundary condition. In Fluent i have specified -1500 Pa gauge (negative) boundary condition at outlet. In OpenFOAM what boundary condition should i give.

like this or something different?

outlet
{
type fixedValue;
value -1225; // (-1500/1.225) pressure/density since units of pressure are [0 2 -2 0 0 0 0].
}

--


All times are GMT -4. The time now is 07:01.