OCR Text |
Show Finally, this work compares the N O emissions obtained from measurements with a variety of numerical models, in addition to the AL model presented above. Figure 10 presents the N O emission at the flue gas exit as resulted from measurements, AL model, Fluent NOx post-processor and P W model, for burner B in both the water-cooled and refractory-lined furnaces and for burner C in the refractory-lined furnace. It is noted that for the P W model the results!1?] are only partial, limited to the medium-momentum burner (cases B-C and B-R). The results show that the AL model is systematically closer to the experimental results, when compared to the other numerical simulations. For all cases, the AL model is of the same order of magnitude with the experimental measurements. For the low- and medium-momentum burners the AL model slightly over-predicts the N O emission in the flue gases. For all cases the Fluent model over-predicts the N O formation by up to one order of magnitude. The PW model for the medium-momentum burner over-predicts the N O emission, with results between the AL model and the Fluent model. It is noted that P W model has been validated for air combustion,t3l and therefore it is not fine-tuned for oxy-flames. IFRF is currently working on an improved NOx model for oxy-flames. Additional calculations using the thermodynamic equilibrium assumption have been performed using the Fluent NO-postprocessor for the cases investigated in Fig. 10. While for cases B_C and C_R the thermodynamic equilibrium assumption leads to slightly better results when compared to the Fluent model results presented above, the case B_R leads to a more significant over-prediction than the Fluent model. The results show that the assumption of thermodynamic equilibrium is very sensitive to the furnace and peak flame temperatures, leading to a dramatic overestimation of N O emission for the hot furnace. 19 |