[siw]

Hi,

The most efficient fluid domain for flight vehicle external aerodynamics is a sphere with the vehicle at the sphere centre. Fluent has the pressure far-field boundary condition (help/flu_ug/flu_ug_sec_bc_pfar.html) specifically for flight vehicle compressible aerodynamic flows and one boundary condition is applied to the entire sphere surface - nice and simple (particularly for changes to angle of attack and side slip angle). Unfortunately, CFX does not have this boundary condition (why waste ANSYS engineers time making useful features when they can make junk like AIM and Discovery Live - end of rant).

The closest CFX option is the velocity inlet with cartesian components and pressure outlet but I have noticed that at the boundary interface between the inlet and outlet (basically the interface where the inlet Velocity u is tangential to the sphere surface and the outlet pressure acts) there are anomolies and discontinuites. But as the sphere boundary is far away from the vehicle is does not affect the local flow their. But it would just be neater just to have Fluent's far-farfield boco in CFX.

[ghorrocks]

The most efficient fluid domain for flight vehicle external aerodynamics is a sphere with the vehicle at the sphere centre.

That is a big statement and I am not sure it is correct. In fact, doesn't your question lead to the conclusion that for CFX a sphere is not the optimum domain shape?

What about C grids? Or cylindrical shapes?

Also - what about the inherent non-symmetry of external aerodynamics - that is upstream nothing much happens but downstream there is a wake and shock waves (if supersonic)? Doesn't this suggest you could have a shorter upstream boundary but need a longer downstream boundary?

Yes, I am aware of the problem you refer to (convergence difficulties at tangential flow on the outlet boundary) - why is why I usually try to make the exit boundary normal to the flow. This means I usually use a C grid or a cylinder for external aerodynamics.