Problems with a compressible shear flow
 

Hi Foamers,

In my research group we are having some problems running the flow field proposed in "Fourth NASA-CAA Workshop", 2004, Category 4 Problem 2 proposed by S.K. Lele.

It is a 2D compressible mixing layer flow formed by a splitter plate with a blunt trailing edge. On the upper stream there is a freestream at Mach1 = 0.1, while the lower one is at Mach2 = 0.6.

The Re number is 250, based on the the momentum thickness of the boundary layer at the inlet of the domain at the lower surface. So we expect a laminar steady solution.

We are trying to solve the flowfield with rhoSimpleFoam, laminar turbulence model, first order upwind schemes for div.

As bc we use:

Inlet (same condition for up and down, with different values):
p - zeroGradient
T - totalTemperature fixed
U - a Von Karman boundary layer profile that matches the freeStream Mach in the far field and that has the correct momentum thickness, using groovyBc.

Outlet:
p - waveTransmissive in order to prevent reflections, with a fixed value of 103682 Pa, consistent with the far field thermodynamics.
T - zeroGradient
U - zeroGradient

Up and down:
p - zeroGradient
T - zeroGradient
U - slip

wall:
p - zeroGradient
T - fixedValue, 300k (required)
U - fixedValue (0 0 0)

The calculation blows up after a few of iterations, so we tryed to smoothly past from a quiescent state to the final using a time-variable inlet velocity that gives a linear ramp from 0 to Mach1, and then another ramp that rises the lower inlet from Mach1 to Mach2, but the calculations is not capable to reach a converged state. It blows up prior to reach the Mach1 conditions.

Anyone can help? Is there any error in the setup, or are there some better solver or bc to use?

Thanks in advance,
Ivan

Update:

Not even capable to obtain a stable solution, with rhoSimpleFoam, rhoPisoFoam and rhoPimpleFoam, even with a full-hex mesh.
I tried to run this case with star-ccm+, but no solutions.
I tried to run with simpleFoam, and I had a stable solution, but clearly wrong, as the inlet Mach in the lower flow is 0.6...

Any suggestions?
Thanks, Ivan

HI Ivan,

i am trying to simulate a jet-flame and came up with some problems that might be similar.

I once used a time dependent inlet to raise velocity...but you have to do it REALY slow!
Did you try with a lower constant speed? (try to lower it until it works)

You might want to check your overall massflow over time!

I have the problem that in my simple jetflame-case (which should converge to some kind of steady state), despite of the constant BC, I have a time dependent massflow as soon as i use a compressible solver !
[see picture, really ugly: (m_{out,max}-m_{out,min})/m_{in} ~~80% ]

regards&good luck with your work!