Hi!

I am carrying out the computation of a swirling combustion chamber:

1- 3D, (nearly) isotropic unstructured grid

2- SIMPLE, stationary version (no time-dependent terms)

3- k-epsilon turbulence model

4- energy and species equations

The problem is that I can get a "stationary" solution only if I use a first-order upwind scheme for the convective terms. Higher order schemes (CDS, QUICK, LUDS, ecc...) do not yield to a stationary solution even with a limiter function. On the other hand, the cold run (no chemical reactions, no flame, only small T gradients due to high wall temperature) converges with all these schemes.

Moreover: I have tried to perform some iterations keeping all the variables but velocity and pressure constant (under relaxation factor = 0) and even in this case V and p oscillate without end.

Very generic question: what can I do to improve the situation?

Thanks Max

Do you expect the flow to be steady? Are you sure that a 'stationary' solution exists? Turn off all convection stabilisation & see what happens. Plot the output data at various times (say a cutplane), put into Powerpoint & run as a movie. What do you observe? Are there any patterns to what you see?

diaw...