OCR Text |
Show 4~----------------------------~-------------' 2 o -2 ~ ~ -4 ~ -6 ~ -8 x ~-10 Cl .3-12 -14 -16 -18 Birmingham Natural Gas 0% Excess Air 21 % Oxygen in Air _20L-----~-------L------~------~------L-----~ 3.1 3.3 3.5 3.7 Log (temperature) (F)) FIGURE 11. RATE OF FORMATION OF NO VS TEMPERATURE determine the temperature-time distribution in a furnace, one would require to formulate a numerical finite element model. Since this was outside the scope of this work, it was decided that the furnace temperature used to determine the residence time in an actual furnace, would be assumed to be the average between the peak temperature and the furnace gases exit temperature (furnace temperature). The actual residence time in a specific furnace would be thus approximated by the specific furnace volume divided by the actual flow rate at the furnace temperature. thus, and T ave. ACFM (Tp eak + T furnace)/2 SCFM x Ta ve. /537 The actual NO produced in the furnace could now be calculated ~y the product of the actual residence time and the rate of NO formation given by equation(8) , at the peakXtemperature in the flame (see Figure 11). Thus, ~O The accuracy of the above outlined methodology will therefore depend on how accurately the peak temperature and the actual residence time are predicted. The NOx expressed in mole fractions cannot be easily compared with published information which is usually expressed as ppm of NO based on a dry basis and 0% oxygen in the exh~ust products. The NSPS are expressed in pounds of N02 per million BTU fired. The method used to convert the ppm of NO to al~ry basis and 0% oxygen was reported in KV~ Inc. and is given below: 194 ppmNOO%dry ppmNO meas. ,dry x % °2Comb.air %0 - %0 o 2Comb.air 2Coillb.Pnrl.dry For ease of calculation the conversion of mole fraction of NO to ppm ~O on a dry basis and 0% oxygen is pres~nted graphIcally in figure 12. As an approximation and for comparison purposes, the NSPS for steam generators of 0.2 lbs NO/ million BTU converts to about 300 ppm NO on a dry basis and 0% oxygen for the combustion of natural gas with stoichiometric amounts of regular (21% oxygen) air. In order tOl~erify the above methodology, NO measurements on an experimental furnace we~e obtained from GTE Products Corp., ~n Towanda, Pa. The GTE furnace of 160 ft (5'x4'x 8') fired with natural gas and regular air was used for burner tests. The preheated air to this burner was obtained from a separate furnace. The GTE measurements are presented in Table 1 together with predicted values using this methodology, details of which are presented in Table 2. Comparison of the NOx values in Table 1 indicate that the predicted values are in the general range of the measured values. Note that runs 5 through 8 represent operation of the burner at high fire at the burner rating. The predicted values at high fire are consistently lower than the measured values. This may be because at high fire, the turbulent intensity in the flame results in actual peak temperatures higher than assumed in this work. Furthermore, the flame surface area could be higher resulting in more NOx actually formed. The measured NOx values indicate that for the stoichiometric combustion of natural gas and regular air in the GTE experimental furnace, preheat temperatures of 1000°F would already produce more NO than permitted by the NSPS for an industrialXsteam generator. It would appear that the air preheat limit for this specific furnace to avoid the 0.2 lb/million BTU would be about 750°F. For purposes of illustration, the above methodology was used to determine what values of NOx would result from using oxygen enriched air(25% oxygen) in the stoichiometric combustion with natural gas in the GTE experimental furnace. The results are presented in Table 3 for runs # 2 and 6 in Table 1. Note that the NOx values have considerably increased. The adiabatic flame temperature has increased from 3750°F to 4200°F while the peak temperatures increased from 2780°F to 2950°F. Note also that a considerable less flow rate is required because the increased oxygen content results in higher values of the available heat which results in considerable fuel savings. The residence time of the gases in the furnace would therefore increase by about 27% and 33% for runs n 2 and 6 respectively. The additional residence time and increased peak temperatures both contribute to |