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/)
-   -   Steady pipe flow mean velocity higher than inlet velocity (http://www.cfd-online.com/Forums/openfoam-solving/59066-steady-pipe-flow-mean-velocity-higher-than-inlet-velocity.html)

anita March 6, 2008 03:59

Hi, I am trying to simulate
 
Hi,

I am trying to simulate a turbulant pipe flow with simpleFoam k-epsilon (with settings from pitzDaily tutorial).

My pipe has a diameter of 43.1 mm. At the inlet is a flow of 10 m/s.

In 2D:
If I take the velocity profile after a pipe length of 20 * diameter, there is a flat "parabolic" profile with mean velocity of 9.7 and a max velocity of 11.1 m/s. This is a factor 0.88.
My book says the factor should be around 0.8.

In 3D:
There is a mean velocity of 10.8 m/s and a max velocity of 12.5 m/s.

My problem is that the mean velocity is higher than the velocity at the inlet.
Furthermore the profile is developed already near the inlet and doesn't change along the pipe. In 2D I can see how the flow develops slowly.

Is this possible ?

Anita

lillberg March 6, 2008 04:40

Hi Anita, Could you post so
 
Hi Anita,

Could you post some pictures describing your problem. Also, what boundary conditions are you using on the inlet and outlet planes for U,P,k and epsilon.

Regards

/Eric

anita March 6, 2008 06:29

Hi Eric, boundary conditions
 
Hi Eric,
boundary conditions from 3D case

Inlet
U fixedValue, uniform(0 10 0)
k fixedValue, uniform 0.375
p zeroGradient
epsilon fixedValue, uniform 8.7549

Outlet
U zeroGradient
k zeroGradient
p fixedValue, uniform 0
epsilon zeroGradient

profile development in 2D:
http://www.cfd-online.com/OpenFOAM_D...ges/1/6917.jpg
Cut through pipe:
http://www.cfd-online.com/OpenFOAM_D...ges/1/6915.jpg
Comparing of 2D and 3D results:
http://www.cfd-online.com/OpenFOAM_D...ges/1/6916.jpg

ngj March 6, 2008 08:34

Hi Anita, The 3D case, what
 
Hi Anita,

The 3D case, what is the cross section like? Any way, the difference in that maximum velocity is dependent on the flow geometry.
What on the other hand is worrying me is that your mean velocity is not the same as your input mean velocity. Could you please try to evaluate it along the axis and see if this deviation is varying along the axis or not? This might tell you if you have done something wrong when evaluating U_mean.

Best regards,
Niels

anita March 6, 2008 11:11

Hi Niels, thanks for your h
 
Hi Niels,

thanks for your help. The cross section is a circle.

When I evaluated U_mean again, I found out I was my fault. I calculated U_mean over a line on the cross section (like in 2D) instead over the area of cross section.

Now I get a U_mean of 9.82 m/s.

My second fault was that I defined inlet velocity in the wrong direction. So the flow was coming in through the outlet and going out though inlet.

I will calculate again with the right inlet conditions and evaluate U_mean along axis.

Anita

anita March 7, 2008 08:58

Hi, after rerunning the 3D
 
Hi,

after rerunning the 3D case U_mean is between 9.94 and 9.96 m/s along the axis. 10 m/s is the inlet velocity. I think now this simulation should be ok.

Regards
Anita

Javed April 12, 2010 06:29

Hi Anita...
R u using axisymmetric geometry?

Thanks
Javed

ashgun September 25, 2012 05:35

can u tell me how u calculated mean velocity?
I calculate the velocity value by mass weighted average value over cross section and that value is very less. close to 0.7 times Umax


All times are GMT -4. The time now is 02:22.