OCR Text |
Show The NOx emissions were estimated using a simple model. One, the model assumed that NOx emission were made at the maximum flame temperature and only small amounts of NOx were produced at lower temperatures. This is only approximately true as the NOx production rate decreases by a factor of two for each 100 degree decrease in temperature. Two, the model assumed that the time temperature history of the furnace is of secondary importance. This assumption appears to be true as only at temperatures above 1500°C is any appreciable NOx produced at long times and in this case NOx emissions is increased only by a factor of 2. Three, the oxygen concentration is of secondary importance for lean flames. This assumption appears to be generally true as increasing the O2 from 0 to 2 percent only increased the NOx emissions by a factor of two. In general, then, the simple model used should approximate the trends in the measured NOx emissions. The estimated NOx emission agreed with the measure NOx emissions within a factor of two to three at oxygen concentrations above 3 percent. However, some aspects of the kinetics model (6) need to be re-evaluated. The estimated levels of prompt NOx increase strongly with increased temperature and decreased weakly with decreased 02. The level of prompt NOx predicted at 1327°C is 0.5 ppm and at 1627 is 5 ppm (d, 3% 02). Data from the large premixed flame indicate prompt NOx might be a minimum of 12 ppm (d, 3% 02) and is not strongly temperature dependent. CONCLUSIONS A 10 X 106 Btu/hr premixed burner was designed, fabricated, and tested. Results show that 12.5 ppm (d, 3% 02) can be achieved by diluting the flame with air or flue gas to lower the maximum flame temperature. The maximum flame temperature has a larger influence than 02 concentration on the NOx emission from these premixed flames. The limit of NOx emissions for these flames appears to be 12.5 ppm (d, 3% 02) and is relatively temperature insensitive. The level of estimated prompt NOx seems too low and to be too temperature sensitive. The prompt NOx portion of the kinetic model should be re-examined. 16 WJSlOO5A.PAOj AFRC-90-25 |