My thermal solution for a problem using the Radiosity Method looks physically plausible, but certain "hot spots" are not in the expected location. What ways exist to refine the solution?

1- Although the radiosity method resolves the non-linearity in one substep, it is sometimes of benefit to force a radiosity problem to do two substeps.

2 - The radiosity equations are solved with the Gauss-Seidel method by default, and the default convergence (TOLER on the RADOPT command) is 0.1. This tolerance is applied to the rate of change of the solution; however, sometimes the method only very slowly approaches the solution and terminates too quickly. Changing this value to .0001 will force more GS iterations. One can replace the GS method with the direct solver, but this is not advisable for 3D problems or for very-large 2D problems.

Show Form
No comments yet. Be the first to add a comment!