[Pauel]

Hey,

I'm facing some issues with the non-reflective boundary condition at an outlet/opening, I tried both.

I'm running a simulation to get the pressure fluctuations at an outlet. It is a transient analysis with a moving mesh to compute a compression. At a given moment in time there is another mesh motion which closes an interface to get the flow out of the compression domain into a collector duct. To figure out the pressure fluctuations at the outlet of the collector duct, I need a non-reflective boundary there.

At the outflow area I set the pressure boundary. When using a non-reflective opening I won't get fluctuations around the target opening pressure of 130 bar. The average over time is at 128 bar. It never reaches the set pressure.
Is there anything I can do to nail the set value in average, with the non-reflective opening boundary?

Because it couldn't reach the defined pressure, I switched to non-reflective outlet boundary. Which worked, but with a partial wall which has been placed there during the run. This didn't make sense to me, because there shouldn't be any backflow.
Now I enlarged the collector duct, same element sizes, just a few more and same timestep, but this doesn't work at all with non-reflective outlet. I can't find any problematic areas in the domain, just at the boundary. After a view timesteps there is a wall at portions without any recirculation downstream or at the outlet, but I can see escalating velocities and get a fatal overflow.
Can anybody give me a hint what might be the problem there?

Thanks in advance,
Paul

[ghorrocks]

There is no such thing as a non-reflective boundary. It should be renamed a "less reflective than the others" boundary condition.

If you want to model conditions at the location you currently have the boundary at, then you need to move the boundary away from this location so the location of interest is in the middle of a flow domain and the boundary condition is far enough away that reflections and other problems do not cause problems at your area of interest. In other words, important features in your simulation should be in the middle of the model, not the boundaries.