Natural convection w/radiation problem with simple card cage enclosed in large air volume. Quarter symmetry (slip b/c on 2 surfaces). Open (0 pressure) top, external wall film coefficient b/c's (20C ref temp), external wall radiation b/c's (20C ref temp), volumetric heat on dummy PCBs.
Analysis auto converges w/in 300 iterations, but the summary file values are way off. Energy balance does not jive (fluid or solid), and the outlet bulk temperature value for Outlet 1 b/c is ~2X the highest temperature shown in the results window.
Tried doubling the number of elements (~350K to ~700K) but doesn't resolve the issues. Results appear legitimate, but would like the summary file to support it.
Without digging into your model it is difficult to make a comment. Here are some common comments I give to people that might help?
1)Try using a fixed temperature on the far walls of the large air domain instead of film coefficients. You shouldn't need the radiation BCs on these 4 walls either.
2) Radiation is only performed on the external air domain surfaces that are walls. This implies that anything setup with a pressure or flow rate boundary conditions will not participate in the radiation calculation. Hence, if you have a heated block in a large air space with all 6 sides setup with a pressure BC, then the heat will not emit to those 6 surfaces and only convection wil be solved for. Simplified example, but something to think about. I don't think this is the case here, but figured I would toss this out.
3) Try using ADV5. This should give a better energy balance in most cases.
You have gold support I believe, so we could always work this in a case if you would like.
I'm putting this out here for my reference, but feel free to comment.
Contacted tech support and was given some ideas to try:
1. Initially I used film coefficient b/c's on bottom and vertical walls (1-2 W/m^2K @ ref temp), with open top, and symmetry planes on 2 surfaces. Placed radiation b/c's on far walls/bottom (1 @ ref temp). Using Advection 5. The summary file showed temperatures on the open top which did not make sense for the analysis.
2. Was told that it is better to use fixed temperatures on far walls and bottom (bucket analysis), with a 0 pressure top AND a film coefficient on open top surface (2 W/m^2K at ref temp) to try to stabiize the top surface temp (since it was not fixed in #1 above, it floated to some bad value). The radiation b/c on the far walls does not need to be included.
3. although both setups converge, #2 results in higher average temperatures (~16%) and lower velocities (why?). The summary file still shows incorrect values for outlet temperature. Radiaton contributes ~1/4 of total heat transfer and convection takes the rest.