CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   CFX (http://www.cfd-online.com/Forums/cfx/)
-   -   Hypersonic setings?, i am afraid this doesnt work (http://www.cfd-online.com/Forums/cfx/21384-hypersonic-setings-i-am-afraid-doesnt-work.html)

jorge sancho June 28, 2005 04:38

Hypersonic setings?, i am afraid this doesnt work
 
Hello, I am trying to simulate a simple supersonic compresion corner (~40, max deflexion for M=5.0). For high Mach numbers 10, doesnt converge at all. i tried: 1) diferent phisical timestepings 2) diferent meshes 3) Mesh refinement 4) diferent inicializations, (stoped, full flow, ramps) 5) steady and transient simulations 6) first upwind

I have read the compression corner validation case, but is just mach 2. And the hypersonic intake at mach 10, but is close to the limit i have found.

I am starting to belive that CFX cannot handle high compressible flow. Pressure based code? Thnks

Neale June 28, 2005 17:26

Re: Hypersonic setings?, i am afraid this doesnt w
 
This will have nothing to do with the code being pressure based. CFX has been run on flows up to mach 25. I think there was an AIAA paper on this recently.

There are so many factors which lead to non-converging solutions. Things to carefully check before throwing your hands in the air and running a density based code are:

Where exactly are the max residuals stalling?

What is the flow doing in that region?

Are you running turbulent or laminar flow?

If turbulent are your boundary conditions appropriate?

If turbulent are the turbulence levels and eddy viscosity appropriate at the shock? Sometimes the flow will go laminar, even with a turbulence model, then you basically get the Euler equations which means all sorts of badness can happen.

What does the mesh look like where the max residuals stall?

What advection discretisation scheme are you using? I hope it's not specified blend...

Neale


test June 29, 2005 00:19

Re: Hypersonic setings?, i am afraid this doesnt w
 
Hi,

Not sure where you have generated your mesh from, but assuming that your mesh is good,

Try a steady state case using Local Timescale Factor of about "1" to start. The mesh sizes may be varying a lot in the domain and also the type of flow (supersonic/subsonic). You need to careful about your boundary conditions. For some combination of boundary conditions the solver will be sensitive to the initial guess. Would advice you to start with a uniform and coarse mesh to see whether the convergence issues are due to the mesh.

Regards, test


All times are GMT -4. The time now is 07:53.