Since flow doesn't model combustion, everything is an approximation. The best way to test an approximation is with a known solution. Do you have that info? Have you done that?
The reason I recommend that is because it will help you isolate the cause of your issues.
If you don't, you'll have some experimentation to do on your model. Some requisite questions are:
What method or method have you tried?
How does the system respond to changes vs looking at absolutes?
You mentioned some values were ok and some weren't. That is counter intuitive. Could you be more specific?
And could you elaborate on what your stack definition is and how you came up with the estimate you are comparing against?
Thanks Jared, I was out of the office for awhile so sorry for the late reply. I have tried it on a known competitor's unit, that has passed the CSA criteria for radiant coefficient, which is a combination of the of several factors the exhaust temperature being one of them. This is the standard used on a radiant convection type heater. This a heat exchanger type construction heater, the exhaust temperature of the stack is a direct result of the burner flame coming into the chamber. The temperature is ar at the outlet of this stack, the heated air going into the building for instance, is the air heated by the tubular exchanger, this has also been measured. On our prototype and the competitors. What I see from the measured to the simulation, is the heated air discharged into the building is about 9% off while the stack temperature is around 30% off. The method of heat sources I have tried are several:
1. One single small source behind the i of the gas independent of the burner airflow which passes around. About 50%.
2. Several different longer cylinders mimicking the flame path at different values to simulate how a flame changes as it passes. The 33% off.
3. One long cylinder of one value(setup by the instructor) The same 33% off.
4. Another pattern of cylinders at different values, About 35%.
From a black body radiation standpoint on the heat sources seems strange as the diameter change doesn't seem to effect the heat change, although I haven't radically changed to 3 or 4 times larger diameter. Seems from a convection standpoint the heat to the hair around the elements isn't convecting as it should. I know that is a lot to digest, but thanks for any insight you may have. Also is there away to change the convergence criteria? I didn't set that I have a question in to the instructor who may have. Or is the default the way to go.
Convergence criteria can be set in calc control options.
Beyond that, I am not sure what your question is. Some questions were answered, some weren't. Some don't have enough detail.
If you have a known solution, iterate on the possibilities until your results match. Then use that method in your analysis.
This type of question isn't an easy one. You need to spend time on testing approximations. Unlikely a software issue here.
Jared thanks for the reply. I understand it is not easy to understand the questions I have without seeing what has already been done, or the whole design that is being tested. The heat balance in the simulation is not correct. With the known parameters going into the unit this should be alot closer. Latent heat, I now it can't do, but sensible heat and radiant heat it should be able to somewhat close. What questions do you want me to expand on? In your experience where is the best area to place the radiant heat source? I know visually how far the flame spreads. The scenerio that seems to work the best is cylinders arranged the length of that flame, however the temperature of the air should be much warmer than shown in the simulation. The heat source at the orifice, a small disk didn't not work. Is there a specific shape that you have used in the past for a radiant heat source that has given you the best results?