I am modeling coal fired boiler. For gas phase radiation I have specified P1 with cell based absorption coefficient and then changed it to domain based.

I am getting completely different answers for both cases.

1) For domain based I got peak temperature as high as 3000K where as for cell based the peak temp is 2000K which is right.

2) The difference in heat flux to water wall is as good as 100MW.

3) The values of absorption coefficient in domain based is 0.01 to 0.1 m-1 where as in case of cell based it is 0.3 to 2 m-1.

I am not able to understand this and why two approaches are giving completely different answers.

Anyone has encountered the same problem. Pl send your suggestions.

Thanks

We have always used the cell based method. Early on, when Fluent incorporated the WSGG model it seemed that the cell based method would better account for differences in H2O and CO2 concentrations in the furnace, so we tried it first.

In simulating hog-fuel and other waste fuel boilers, we have gotten decent agreement between measured and modeled furnace exit gas temperatures for most of the 50 or 60 boilers we have data for. We generally use a value of 0.3 to 0.4 for the wall emissivity with the P1 model, which is much lower than for steel but it could be argued to be representive of wall deposits.

It would seem that if you have large gradients in grey gases, then the cell based method would be better. The P1 model is likely non-specific enough that either way could work. That is, it only generally represents radiative heat transfer.