Convergence of Density-based solutions
 

Dear all,

I would like to discuss about the convergence problems with density-based (ideal gas)solutions using Ansys Fluent 14.5.

Whenever we tried a solution with Mach>0.3 (high subsonic, tansonic, supersoni and even hypersonic), with a high quality mesh, it is almost impossible to converge the solution either the solution become stiff or diverged after a few iterations.

Using hybrid initialization with external aerodynamic favorable settings make the solution diverged after few iterations whereas normal initialization from farfield etc. diverged after more iterations than hybrid initialization but showing temperature/turb. viscosity/pressure messages (problem with both type of initialization).

Is there any best practice for a density-based solution or solution with high Mach number to get converged without any problem.

Comparing Fluent with CFD++, same mesh, with similar boundary conditions and solver settings (density-based), it converged even in a few hundreds iterations without any efforts.
Is there anyone who is familiar with both the software and can comment on the differences what they have?

Have you tried the NACA0012 tutorial from FLUENT. That is a case of M=0.72 or something similar, that have the settings for a transonic simulation. I hope this helps you. Regards
Gonzalo

I tried the tutorial, it worked but only for that tutorial. When I try to adopt the same technique/procedure in some complex problem and using some other turbulence model like SST k-w; solution diverges.

How can I solve the problem.

Try to work with solution steering adapting the CFL