[rocketLauncher]

Hi all,

As part of a study comparing Ansys Fluent to open source CFD tools, I am starting with the simple case of laminar, incompressible flow in a mixing elbow. The elbow has two inlets at 1 and 3 m/s respectively and the outlet is set at 0 gauge pressure. I constructed a reference case in Ansys Fluent using the inviscid solver. The simulation converges nicely both in terms of residuals and the results visible in the velocity and pressure field contours.

However, I encountered problems when I exported the mesh from Ansys and used it in openfoam. I tried to set up the OF solver as close as possible to ansys but my velocity contours are very bizzare. Indeed, while the simulation converges in terms of residuals, the actual results are not good.

What could cause this difference of results? I am attaching a zip file that shows the contours to which I referred, the ansys mesh, and the OF case. Any ideas are welcome, this has been very frustrating.

Thanks in advance.


https://www.dropbox.com/scl/fi/i7fvg...=hdmnfavz&dl=0

[dlahaye]

I am unable to view your results on Dropbox my the limited patience I have. My apologies.

Questions:

1/ does checkMesh yield particular information?

2/ have you tried smaller relaxation factor?

[wkernkamp]

I could see the images. There is a pressure singularity in the pressure picture. It looks like that is for the reference (ansys) run. The OpenFOAM velocities show a separation at that point, which is more fluid like. It appears to me that you are not doing the exect same solution type (Euler) on both. There are quite a few CFD validation cases that you could look at. Typically, these validation cases have already been run for all major solvers. Example: https://doc.openfoam.com/2306/exampl...lent/naca0012/

[rocketLauncher]

Thank you for the quick replies. I have solved the issue. It was a rather silly mistake actually, just an issue with the solvers. The openfoam model was taking into account viscosity, but in ansys the solver was inviscid.