[adarsh tiwari]

Hi All,

sorry for the new thread. I am analyzing a simple case of flow through a duct with sharp 90 deg. contraction in between.

The geometry is having a 2.78m x 3.28m C/S area at inlet and after length of 10m, a sudden contraction occurs and the C/S area becomes 1.3m x 1.3m and length again runs upto 10m. thus the total length is 10m+10m=20m.

for initial run i am running with the coarse mesh and incompressible steady state solver 'simpleFoam'.

after few iterations, the software says that the solution is converged , but the values I am getting is unrealistic. I also had tried different approaches and boundary conditions, the results are more or less same.

I have also referred to various posts in the forum e.g. http://www.cfd-online.com/Forums/ope...fvschemes.html

http://www.cfd-online.com/Forums/ope...-problems.html

For the convenience, i am also attaching my fvSchemes, fvSolutions and other Files.

Thanks and Regards,
Adarsh Tiwari

[GDTech]

Hi,

Your BCs for U and p are not correct : you cannot fix velocity both at inlet and outlet !
Commonly used BCs are the following :

Inlet :

• U : fixedValue (or flowRateInletVelocity)
• p : zeroGradient

Outlet :

• U : inletOutlet
• p : fixedValue

Have a look at incompressible/simpleFoam/pitzDaily tutorial.


Best regards.

[adarsh tiwari]

Hi GDTech,

Thanks for your quick reply, I have already tried with the boundary conditions which you are mentioning but unfortunately it doesn't work.

If it is simulated using 'simpleFoam' then the solution diverges after few iterations while if we use 'rhoSimpleFoam' it gives some values which are not convincing.

In the results, the velocity values are okay on inlet side, while on the outlet side, the velocity approaches to approx. 2-3 m/s. Practically, it should be around 9-10 m/s.

Hence, for maintaining it, I thought of fixing velocity at both the ends.
I guess the problem may be in 'epsilon'.
Thanks and Regards,
Adarsh Tiwari

[GDTech]

Originally Posted by adarsh tiwari

I have already tried with the boundary conditions which you are mentioning but unfortunately it doesn't work.

Your turbulence modeling is also wrong : nutURoughWallFunction cannot be used for k and epsilon !!

I really suggest you to have a look at incompressible/simpleFoam/pitzDaily tutorial !!!