problem 3D case with narrow outlets

A subforum specific to the development of the OpenFoam-based workbenches ( Cfd and CfdOF )

Moderator: oliveroxtoby

Posts: 40
Joined: Sun Nov 25, 2018 8:57 pm

problem 3D case with narrow outlets

Postby julieng » Sat Jan 19, 2019 9:30 am


I want to calculate the pressure drop of a duct with narrow outlets (duct diameter 220 mm, inlet velocity 13 m/s, air at 100°C). I use stationnary calculation with incompressible kw SST model.

I have import the step file of the geometry in freeCAD.

When I mesh the geometry with cfMesh I cannot visualize the narrow outlets in Paraview. I have refine the outlets boundaries but no success.
With a first coarse mesh, I was able to compute the case but it seems that no output was taken into accout...

I joint the case

Best regards

(27.58 KiB) Downloaded 27 times
Posts: 1708
Joined: Sat May 20, 2017 12:06 pm
Location: Germany

Re: problem 3D case with narrow outlets

Postby thschrader » Sat Jan 19, 2019 9:12 pm

You can see the outlets when using a max element size = 10 mm instead of 30.
The outlet area is ten times smaller than the inlet. You get an average outlet
speed of 130 m/s (inlet 13 m/s) when using incompressible fluid.
In the openfoam data field your write intervall is set to 100. When terminating
the calculation before 100 iterations there is no output written.
Maybe you should use the Hisa solver for compressible fluid
(I got 170 m/s outlet-speed maximum).
U.JPG (26.09 KiB) Viewed 345 times
p.JPG (25.01 KiB) Viewed 345 times
Posts: 40
Joined: Sun Nov 25, 2018 8:57 pm

Re: problem 3D case with narrow outlets

Postby julieng » Sun Jan 20, 2019 11:27 pm

Thank you thschrader,

I reduce to 10 the write interval and I see curves and everything works. The case converges with error <0.0001 after 750 iterations so why 100 is too high?
I was obliged to stop the computation even if error was lower than 0.0001.
In stationary cases only the number of iterations (1000 in my case) is taken into account to finish the calculation?

I have done the same computation (with a coarser mesh) with a commercial software and I found with a compressible RANS ke solver a pressure close to 24000 Pa. So it seems to be correct. (With FreeCAD close to 20000 Pa)

I have not yet try the Hisa solver (my PC has only 8 GB of RAM, with a low processor)

Thank you again

Best regards