**5. Conclusions**

We performed nongray radiation calculations of two radiation problems in homogeneous isothermal media. The first medium is typical of wet-recycle oxy-fuel combustion environment, with a molar composition of 65% CO2 and 35% H2O; whereas the second approximates a dry-recycle environment, with a molar composition of 90% CO2 and 10% H2O. The domain was a 12×12×40 m rectangular enclosure at 1 500 K. For each environment, we generated reference solutions using the box model based on the exponential wide band approach. We also calculated solutions using five (recent) oxy-fuel and one (older) air-fuel weighted-sum-of-gray-gases models that were proposed in the literature. Comparing


Table 8. Wall radiative heat transfer for the oxy-fuel environment with 90%CO2

different qualitative and quantitative radiative characteristics from the obtained solutions, we see that significant improvements in predictive capability can be obtained using an oxy-WSGGM. Using the air-fuel model would result in appreciable underprediction of the local and area-integrated radiative heat flux to the wall, and in an overprediction of temperatures due to the underprediction of the heat loss due to radiation. The errors become more pronounced for the high-CO2-concentration case, which is relevant to dry-recycle oxy-fuel combustion. The radiative heat flux was much more sensitive to the gas composition than the radiative source term. For the oxy-fuel WSGG models, no particular model was clearly superior. This suggests that the model used for a particular combustion problem should be selected based on the simplicity of the model and the consistency between the operating regime of the target system and the regime of the training data.
