Learning Forum

[simonhenn]

Hello,

I am simulating a compressible airflow through a test stand with the RANS method (steady, SIMPLE/Coupled). I use a pressure inlet and two pressure outlets (red and green).

This is a picture of the geometry:

[Rob]

I'd start here /forum/discussion/21916/numerical-instabilities-in-simulation-of-an-airflow-through-a-diffusor#latestn

[simonhenn]

Thanks for your answer! However, I did in fact look up the Coanda effect but didn't really understand how it applies to my problem. Could you explain what you meant ?n

[Rob]

You have a jet moving into a volume, that tends to be unstable due to the pressure field if it doesn't attach fully to a surface. n

[simonhenn]

Ok thanks. What would you recommend to change on my setup? Use a different model or method? If so, which one? Also, can you think of a reason why the simulation is relatively stable with one particular mesh (6.3 mm connector dim)? n

[Rob]

Not really, it's likely a true result of the flow. Mesh size can stabilise the flow as if it's too coarse the numerical diffusion masks the jet motion. It's why some mesh independence studies never reach a conclusion: each time the mesh is refined the solver picks up something new (and usually transient) in the solution. n