You may already have a method if Aspen EDR has this into account.
When you go to older methods (like the Lobo and Evans in Kern) you see that they consider all the radiation reaching the refractory as re-emmited, but they make no difference on whether it is by reflection or absorption-radiation.
Aspen EDR may have this calculation sophistication already built in (that is, that the gas is more transparent to absorbed-re-radiated radiation than to reflected one).
The way to know is to look at the radiant section results, or to simulate a furnace without convective section.
The total gas needed may be kept constant by the simulator either:
a-) because Aspen EDR does not considers the effect of wall emmissivity (then you have no way to calculate with it) or
b-) because it does, but the specs are such that in the second case you have more heat transferred in the radiant section and less in the convective but overall the same.
In the case I was involved, we relied on the supplier calculation. They were accurate. There was no fuel saving, by the way. The main benefit for us was lower tube metal temperature, that was a bottleneck. The heat in the radiant section of the reformer increased by 4% and the flue gas temperature leaving the radiant section was reduced initially by about 40°C. Then we had to change the firing profile to return the flue gas to the previous temperature (we needed the heating in the convective section), but this resulted in a reduction in the peak tube metal temperature.
Edited by Saml, 27 October 2016 - 06:42 PM.