Dear all,

I'm humbly entering this interesting circle of CFX users with a simple question.

I'm using CFX5 with K-eps in order to study an industrial coupled heat transfer problem (solid conduction + fluid convection). Boundary conditions are quite trivial: one inlet, one outlet, two openings, two hot walls and some interface walls. On the other hand, the model is rotating and geometry is VERY complex: that's way I chose the unstructured CFX5 approach, rather than CFX4's structured.

I've been working for the last eight months with CFX5.4, obtaining good results in conforting accordance with experimental data (inside K-eps confidence levels obviously ;-) ).

The point is: all benchmarks I'm conducting using the NEW CFX5.4.1 solver turn out to be very disapponting.

The worst come from Eddy Viscosity and and Turb Kinetic Energy: in other words, all turbulence (which should have been re-written in its diffusive terms, I was told) seems to have come to a big nothing, i.e. severe under-estimation and false diffusion.

Does anyone know something about this?

Regards, Sebastian

Sebastien

Contact your local CFX office for support. They will be happy to help you.

Gavin