OCR Text |
Show Figure 9. Index of N-Reduction along the GTCC. Instantaneous vs Sub-Streamed Mixing: Effect of the Methane in the Syngas. Parameter: [CH4]= 0%, 1%,3%,6%. 1.0 • 1st STAGE 2nd STAGE 3rd STAGE i i i i 1 0.4 ! ///Jt^l ^ ! ! | i /^f[CH4] j j 0.2 Mr _*^^^^^^^^^^™^^^^ inst" i J J _ ^ « « ^ ^ ^ ' ' ' 0.0 <fl^.. p= 10 bar, T= 1300 degC; Dopant: [NH3]= 1000 ppmv. 0 5 10 15 residence time [ms] the more the number of sub-streams the better the Nfix reduction. Again, the oxygen mixing delay results in a greater conversion of the N H 3 to N 2 and to H C N that is later converted also to N 2 (Fig. 7). When delayed mixing, the more the C H 4 in the fuel gas the lower the fixed nitrogen reduction: by rising the C H 4 content in the syngas the formation of H C N is enhanced by the larger availability of hydrocarbons for reactions with NH3 . In the instantaneous cases the methane in the syngas does not affect the Nfix reduction: all the hydrocarbons can be oxidized because of the large oxygen availability. Whatever the methane concentration, provided ignition takes place, the lower the temperature the better the nitrogen conversion. At low temperature traces of H C N are predicted at the outlet of the G T C C . CONCLUSIONS Within the investigated ranges, ignition troubles can be predicted depending on the values of pressure, temperature, methane content in the fuel, and model used (Fig. 9). The Instantaneous Model always predicts lower degrees of Nfix conversion to N 2 as compared to the Delayed Models. The main deviations are located along the first stage (sub-stoichiometric region), whereas along the third stage (over-stoichiometric region) the only differences are predicted in the share between N O and N 0 2 (Fig. 7). With the Sub-Streamed Model, predictions are heavily affected by the adopted number of sub-streams: the more the sub-streams the higher the nitrogen reduction. Comparing 1 to 5 sub-streams models the degree of reduction rises from some 20 % to over 70 % (Fig. 3). N o differences are predicted if mixing completion take place within a residence time in each stage of 2 or 4 ms. With the Combined Model, predictions are heavily affected by the share of residence time between CSTRs and PFRs: the longer the residence time in the CSTRs the higher the Nfix reduction (Fig. 4). As expected, predictions by the Sub-Streamed Model always are intermediates between the ones by the Instantaneous and by the Combined Models (Fig. 5). Therefore the Combined Model can be definitively assumed as an extreme case of the Sub-Streamed Model. Pressure and temperature are the main affecting variables on the prediction for the Instantaneous Model and for both the Delayed Models. Such dependence is much wider in the first case: by varying pressure and temperature the degree of nitrogen reduction ranges here from less than 2 0 % to over 80%, whereas in the delayed cases it ranges from over 5 0 % to over 9 0 % (Fig. 6). Similarly to the Instantaneous Model, with a Delayed Model comparable nitrogen reduction are predicted if N H 3 and/or N O or H C N are the fuel dopants (Fig. 8). Predictions are heavily affected by the model variable of both Delayed Models: increasing the number of sub-streams or arising the residence time in the CSTRs, the degree of nitrogen reduction can go up from less than 2 0 % to over 90%. A somewhat lesser dependence is shown when the dopant is H C N : the fuel dopant to reduce the easiest is a function of the model variable. When a Delayed Model, a decrease in the methane contents of the fuel deals with an increase of the fixed nitrogen reduction; whereas when the Instantaneous Model, the methane amount does not have major affects on the predictions (Fig. 9). Once again the condition for the adopted Model affect heavily the prediction of the nitrogen reduction: by increasing the number of sub-streams, the degree of reduction arises from 2 0 % to 80%. Acknowledgments This work belongs to a larger international project partly funded by European Commission through its Joule Programme under the contract No. JOR-CT97-0157. REFERENCES Alzueta, M.U., Bilbao, R., Millera, A., Glarborg, P., Ostberg, M., and Dam-Johansen, K., 1998, "Modeling Low-Temperature Gas Reburning. N O x Reduction Potential and Effects of Mixing," Energy & Fuels, Vol. 12, pp. 329-338. Bevan, S., et al., 1994, "Development and Testing of L o w Btu Fuel Gas Turbine Combustors," Advances in I G C C and P F B C Rewiev Meeting, DOE/METC-94/1008, D O E , Morgantown, pp. 280-289. |