[Diger]

Using a Pressure Outlet with a very low Pressure compared to the typical ambient pressure, the velocity at this outlet can become Mach1.

If the pressure is further lowered the exit acts as a nozzle and the exit blocks, which means that the software is no longer capable of lowering the pressure further and also mass flow does not increase anymore.

This is all more or less known theory, but sometimes the velocity does indeed exceed Mach1 and ANSYS Fluent does not complain.

But if a Target Mass Flow is specified and the Machnumber exceeds 98% then Fluent Does complain...

Why?

[LuckyTran]

Well you already hit the nail on the head.

The target mass flow rate boundary condition is a pressure outlet condition with the pressure being adjusted each iteration to steer the mass flow towards the set point. If the exit is supersonic then adjusting the pressure does nothing. So you have told Fluent to adjust the pressure to reach the desired mass-flow, but it knows (and you also know) that adjusting the pressure does nothing.

[Diger]

Yes, but the point is why does it complain when specifying a target mass flow, but not when the boundary is specified as a simple pressure outlet.

Shouldn't it always complain when the velocity exceeds mach1 at a boundary?

[LuckyTran]

The Mach number at a boundary (or any location for that matter) can exceed Mach 1. That is not a problem. With a fixed pressure and somehow the Mach number becomes 1 or greater, you get what you get (and you don't say anything about what the mass-flow needs to be). Here there is also not any problem. It is not a problem because all you have to do is extrapolate the pressure from the interior cells to the boundary and the pressure that you specify is not needed.

With a target mass flow, you say you want a certain mass flow, but there's no longer any knob Fluent can adjust to get you that mass flow. Here there is a problem because how the heck can I guarantee a certain mass flow if the boundary is sonic/supersonic?

[Diger]

Originally Posted by LuckyTran

It is not a problem because all you have to do is extrapolate the pressure from the interior cells to the boundary and the pressure that you specify is not needed.

Extrapolate from the boundary adjacent cell center you mean?
And what do you mean the pressure that I specify is not needed?
How is a boundary pressure not needed, unless I use target mass flow?

[LuckyTran]

When the flow becomes supersonic, the nature of the flow changes from elliptic (in the subsonic case) to hyperbolic. When this happens it is no longer a boundary value problem but is now an initial value problem and you no longer need to provide the downstream BC's.

Originally Posted by Diger

How is a boundary pressure not needed, unless I use target mass flow?

The targeted mass flow is a pressure outlet boundary condition. The difference is Fluent has a very dumb algorithm to change the pressure for you until you get the flow rate that you want. It is a convenience option and does not add any novel capability. You can manually adjust the outlet pressure yourself and achieve the same result (assuming it is even possible). Of course in the supersonic case, it is not possible and hence the warning.

[Diger]

Originally Posted by LuckyTran

The difference is Fluent has a very dumb algorithm to change the pressure for you until you get the flow rate that you want.

Is CFX using the same algorithm?