[evcelica]

I am performing a quasi-2D axis-symmetric model and would like to include radiation if possible. The model is an enclosed cylinder and I an modeling it as a 1 degree 2D wedge (but I have removed the central prism element and used a symmetry condition there as well to increase mesh quality) The sides of the enclosure are steel, and it is filled with a monatomic gas which is completely transparent to radiation. I thought Discrete transfer model would be my best option, but I read in the documentation it uses a number of rays at different angles (8 by default) to compute view factors to different elements. With the thin model I'm guessing pretty much all the rays will contact the symmetry boundaries before another surface. Can I control the angles of the rays somehow? How does the discrete transfer model treat symmetry conditions? Will the rays just reflect perfectly off the boundaries until they meet other walls? Or will they be lost? Is any other radiation model more appropriate? The documentation describes nothing about this radiation model and symmetry. I'm really hoping CFX didn't blow it with their 3D only code again.

Thanks in advance

[ghorrocks]

The CFX documentation is a bit thin on describing the details of the various radiation modelling. But my understanding is the rays used to calculate view factors in the discrete transfer model reflect off symmetry planes, so you should be OK. SO I think the discrete transfer model is the best one to use for you.

But I would do a little benchmark model to check it is functioning properly - maybe a fixed temperature wall inside a 2D tube and check the incident radiation is the correct analytical result.

[evcelica]

Very good advice Glenn, I simple benchmark is defiantly a good idea, Thanks, I'll do that!