When you don't apply a boundary condition to a boundary, it will automatically be defined as a wall (boundary 0 - Wall). Did anybody notice that if you explicitly define those boundaries as walls, that the convergence of the solution is changing drastically? Why is that?

Nijso

You can alter the definition of the default wall boundaries by redefining region 0 and you can also change the wall properties of any defined wall by defining the appropriate region number (RDEFINE command). Are you absolutely sure that the walls you defined have exactly the same properties as the default walls?

I have a convergence problem, but I have not noticed this. What is your version? Is it for pc or unix? I'll try some cases.

Giovanni

I'm running on Sun/solaris, version 3.101 It seems that it does not do this in general, I tried some other simple configurations. I'm doing a simulation with 2-D axisymmetric flow with swirling inlet conditions, the inlet is about 1/2 the radius of the pipe.With this configuration, it's convergence is different. I have not isolated the exact source of the problem yet... Nijso