[peter969]

I'm a long-time lurker, first-time poster, so I should start by thanking the contributors to this forum for knowledge that I have pilfered in the past!

I am trying to simulate (in steady-state) the flow around a vehicle when it's travelling around a corner. Initially I am simply considering a 2D scenario with a semi-circular annulus (see hastily-drawn picture), and a brick representing the vehicle.

As the picture shows, I want to create the flow field by applying a Domain Rotation about the axis of the annulus, with fresh mass-flow being added at the inlet, and the wake of the brick being removed at the outlet.

The inner/outer radial walls have been defined as free slip walls, but I'm unsure about the inlet and outlet. My (not necessarily correct) logic tells me that I should specify Total Pressure = 1atm at the inlet, and Static Pressure = 0atm at the outlet .The user guide seems to indicate that using 'Opening' boundaries is the best way of doing this.

However, while this gives feasible forces and total pressures, the radial velocity distribution across the annulus is far away from the linear profile one would expect upstream of the brick, and is in some areas an order of magnitude larger!
When I specify the inlet as Static Pressure = 0atm, the velocity distribution is much more valid, but the total pressures (and consequently forces) are wrong; in the wake of the brick, Ptotal is apparently negative!

If anyone is able to shed any light on why I my BCs may not be giving me what I expect, I would be very grateful.

Pete

(I should mention that I'm using CFX within Ansys Workbench V12, on Windows).

[ghorrocks]

Use a rotating frame of reference, probably a frozen rotor approach I guess. Set the rotation velocity to the rotation velocity of the car in the corner. At the exit set an outlet with pressure=0, and the inlet with a normal velocity set at the solid body rotation velocity at that radius.

I think that should do it. It is always a bit tricky to get the frames of reference right in weird models like this.

[peter969]

Thanks Glenn,

You're right about it being tricky! Thanks to your advice - and a certain degree of trail-and-error - I now have a much more stable and sensible solution.

However while the velocity distribution is now spot-on, the Ptotal values are still not realistic (they hover around 0 Pa in the far-field). In the domain settings, my Reference Pressure is 1atm, and my inlet and outlet are as you advised. Might this be just a by-product of the way my model is set up?

Pete

[ghorrocks]

Check what frame of reference you are calculating Ptot in.

[peter969]

Once again, thanks for your help.

Got a few more issues to iron out, but things now working well.