Energy & Fuels, Vol.26, No.6, 3349-3356, 2012
Nongray-Gas Effects in Modeling of Large-Scale Oxy-Fuel Combustion Processes
Studies have been conducted to implement oxy-fuel combustion with flue gas recycle in conventional utility boilers as an effective effort of carbon capture and storage. However, combustion under oxy-fuel conditions is significantly different from conventional air fuel firing, and radiative heat transfer under oxy-fuel conditions is one of the fundamental issues. This paper demonstrates the nongray-gas effects in modeling of large-scale oxy-fuel combustion processes. Oxy-fuel combustion of natural gas in a large-scale utility boiler is numerically investigated, in which investigation a recently refined weighted-sum-of-gray-gases model (WSGGM) applicable to oxy-fuel conditions is used to perform nongray and gray calculations, respectively, and a widely used air fuel WSGGM is also employed for gray calculation. This makes the only difference among the different computational cases. The simulation results show that the gray and nongray calculations of the same oxy-fuel WSGGM make distinctly different predictions in the wall radiative heat transfer, incident radiative flux, radiative source, gas temperature, and species profiles. Relative to the nongray implementation, the gray calculation of the oxy-fuel WSGGM remarkably overpredicts the radiative heat transfer to the furnace walls and underpredicts the gas temperature at the furnace exit plane, which also result in a higher incomplete combustion in the gray calculation. Moreover, the gray and nongray calculations of the same WSGGM make much more pronounced difference in the results than the gray implementation of different WSGGMs does (i.e., the oxy-fuel and air fuel WSGGMs). Even though particle radiation also has an important impact and will compromise the demonstrated nongray-gas effects to some extent in large-scale oxy-coal combustion, the nongray formulation of an oxy-fuel WSGGM is still highly recommended for a reliable oxy-fuel combustion modeling.