[Roland R]

Hello,

I have a complicated geometry which contains some narrow cross sections. In my tutorial simulation (see below) there is a simple choke. It is a 2D axisymmetrical geometry, the pressure ratio is 3.
The main settings:
- steady state
- total pressure inflow/static pressure outflow
- air ideal gas
- SST turb.modell
- high resolution

The results:

http://img405.imageshack.us/i/chokedens.png/

http://img686.imageshack.us/i/chokemach.png/

http://img697.imageshack.us/i/chokemesh.png/

http://img683.imageshack.us/i/chokepress.png/

http://img413.imageshack.us/i/chokeyplus.png/

Based on the convergence history the simulation is converged. I don't understand the cause of the too high Mach number. The geometry doesn't contain Laval tube, this is a simple choke.

I have tried to set the expert parameter/high speed models but the result is unchanged.

Can the material model cause the problem?
...perhabs other ideas?

Thanks in advance;
Roland

[ghorrocks]

Looks like a standard expansion shock to me. What don't you like about it?

If you want to see an expansion shock in real life have a look at a fighter jet engine exhaust, especially with afterburners on. You can see it clearly then.

[Roland R]

Hello Glenn,

As far as I know, the Mach number can not be larger than 1. Although a laval nozzle develops from the flow separations after the choke, but Ma=2.5??? In my opinion this is unrealistic...

I have some measurements to the original investigated geometry. The total temperature was measured before and after the choke. The total temp. is in the case of my simulation ~24C (after the choke). This value was ~15C in case of the measurement...

Roland

[ghorrocks]

In my opinion this is unrealistic...

Looks like you need to do some further research into compressible flow then. In the nozzle the flow does not exceed Ma=1. In the expanding region afterwards it can - and in your case it is. Look at expansion shocks in some textbooks.

The total temp. is in the case of my simulation ~24C (after the choke). This value was ~15C in case of the measurement...

This is a question about simulation accuracy - have you done the basics and checked mesh sensitivity, convergence tolerance and that the physics is correct (material properties is a key one here, properties vary considerably over this temperature/pressure range)