Autodesk CFD - Problem with Boundary Conditions and Unexpected Flow Results

[stellazimmerman82]

Hello everyone,

I’m currently running a CFD simulation in Autodesk CFD 2025 for a simple air duct model, but I’m having trouble getting realistic flow results. My setup includes one inlet and one outlet, with the inlet defined by a velocity boundary condition (2 m/s) and the outlet as pressure = 0 Pa (ambient).

However, after running the simulation, I noticed that the flow direction and pressure distribution don’t look right — in some regions, the velocity vectors even point back toward the inlet, which doesn’t make physical sense.

I’ve already checked for:

• Mesh quality (seems fine, average aspect ratio < 4)

• Correct fluid properties (air, standard conditions)

• No overlapping geometry or reversed faces

👉 My question:
What are the most common causes of reversed flow or unstable velocity fields in Autodesk CFD simulations, especially when using simple inlet/outlet boundary conditions?
Is there a recommended way to stabilize the results — maybe by adjusting solver settings, mesh density, or turbulence model?

Any insight or tips from experienced users would be really helpful. I just want to understand why the flow behaves this way and how to get more realistic results.

Thank you so much!

[maciej.jaskiernik]

Hi @stellazimmerman82 

Flow circulation in the outlet area is quite common problem in the CFD. 

It is caused by the fact that you assume 0 Pa on the whole surface which is not realistic but no better solution is known. 

I guess in your case it is cased by the geometry in the outlet area. If you want you can send the screenshot showing cross section so we could discuss possible geometry fix solutions.

If your goal is to simply get rid of this problem from CFD simulation, you can extend the outlet (as if you would put a long pipe at the end of your system).

That should make the flow more stable and (what's more important IMO) you'll be able to tell if the circulation is caused by the geometry or the boundary conditions simplification. 

Best regards,

Maciej