Nitrogen oxide (NOX) tests were performed on various gas burners at the Flynn Burner facility in New Rochelle. The purpose was to determine the NOX levels and to correlate the NOXemission levels against variables such as excess O2 level and firing intensity. The results are presented in this report.
2.0 SUMMARY OF RESULTS
The test program yielded NOX concentration from 20 to 180 ppm (dry, at 3% O2) for the four burners tested and over a range of excess air levels. The levels of excess air were a function of the premix ratio of the fuel/air mixture and the amount of fuel flowrate. The major finding from the testing was that the NOX emissions can be controlled to very low levels by controlling the excess air levels.
The excess air levels are formed by two sources; the fuel/air premix ratio and the amount of air infiltration. While the premix can be easily controlled, the air infiltration is a function of the design of the combustion system and the specific operating conditions. The more open the system, the more air infiltration will occur. The exiting gas temperature is a good indicator of excess air level. At high infiltration, the exit temperatures will be low. This is due to the quenching affect of the ambient air as it mixes with the combustion gases. Therefore, low temperature systems are indicative of low NOX levels.
The following specific results were attained from the testing:
- For the 4 burners tested, the NOX emissions were strongly correlated to excess air level in the combustor. The higher the excess air, the lower
- The NOX level varied from 180 ppm (dry, at 3% O2) at 100% excess air to 20 ppm at 300% excess air.
- The NOX levels increased negligibly or very slightly with firing intensity for a given plasma index.
- The excess air changed as a function of the premix ratio (plasma index) and the firing intensity (gas flowrate). Because the chamber was open, a great amount of air infiltrated and mixed with the combustion products. The infiltrating air caused the exiting excess air to be much higher than the original premix. For example, a premix equivalent to a stoichiometric combustion ratio (0% excess air or a plasma index of 50) could produce an exiting excess air of 100 to 250% according to firing intensity.
- The exiting combustion gas temperature was correlated to the excess air level. Thus, gas temperature may be a good indicator of NOX. Generally, gas temperatures in the 1200 to 1500°F range exhibited NOX levels in the 120 to 180 ppm range whereas temperatures in the 900 to 1100°F range exhibited NOX levels in the 20 to 80 ppm range.
3.0 CONCLUSIONS AND RECOMMENDATIONS
- The amount of NOX production is a strong function of excess air level. High excess air levels produce low NOX levels. Therefore, high excess air levels are recommended for NOX control.
- High excess air levels are formed when the system is open to ambient. A negative draft is created due to the high temperatures of the combustion gases. This draft allows ambient air to infiltrate and diffuse into the combustion gases. The diffusion rate will control the rate of NOxproduction because it lowers the peak flame temperature.
- The exiting gas temperature is a good indicator of the excess air level. When the ambient air mixes with the combustion gases, the gas temperature is reduced. This is the main method of temperature change in the system.
- In systems where the exiting gas temperatures are low, below 900°F, it would be expected that the NOX levels would be low, less than 50 ppm. However, each burner site would have to be tested to verify this. It would be expected that if several sites with similar configurations were tested, the data could be used to show correlation on a generic or similarity basis. It would be expected that this might be sufficient to show compliance with regulatory requirements.
4.0 DISCUSSION OF RESULTS
4.1 Test Setup
The testing was performed in a cylindrical chamber of dimensions approximately 20 inches in diameter by 4 feet in length. The burners were rod shaped and installed either in the center or along the bottom of the cylinder. The sides of the cylinder were left open except for the band burner where the ends were loosely capped. The flue gas samples were extracted from a 4 inch flue pipe at the top of the cylinder. The emissions were measured in real time with the use of an ETEC emissions laboratory van.
4.2 Emissions Measurements
The emissions van was equipped with on-line instruments to measure NOX, CO and O2. The flue gas was pumped, filtered and dried prior to being admitted to the instruments. NOx was measured by a Thermal Electron Model 10A vacuum type chemiluminescent sensor, CO was measured with a ANARAD non-dispersive infrared sensor while O2 was measured with a Thermox zirconium oxide sensor. The instruments were calibrated with the use of certified bottled gases. Pure nitrogen was used as a zero gas. A mixture of 350 ppm CO and 10% CO2in nitrogen was used for the CO calibration gas. The CO2 is used to calibrate out interference effects of CO2 in the instrument.
NOX is measured dry in parts per million by volume. In order to correct the NOx concentrations for different air dilution levels. The measured NOx is reduced to a standard excess O2 level. The standard in the industry is 3% O2 on a dry basis. This makes the NO2 level comparable to any other reported NOX level. The NOX emission level at reduced to 3% O2 is directly convertible to a rate in lbs of NOX (as NO2) per million Btu of heat input. For natural gas, 245 ppm at 3% O2 is equivalent to 0.3 lbs/million Btu.
NOX is measured dry in parts per million by volume. In order to correct the NOXconcentrations for different air dilution levels. The measured NOX is reduced to a standard excess O2 level. The standard in the industry is 3% O2 on a dry basis. This makes the NO2level comparable to any other reported NOX level. The NOX emission level at reduced to 3% O2 is directly convertible to a rate in lbs of NOX (as NO2) per million Btu of heat input. For natural gas, 245 ppm at 3% O2 is equivalent to 0.3 lbs/million Btu.
The gas is premixed with air prior to entering the burner. The amount of premix is measured by Flynn with an instrument which indicates the "plasma index." A very small portion of the premixed gas is sent into the instrument which measures the fuel to air ratio. The following table presents the best estimates of equivalent excess O2 versus plasma index.
4.3 Test Results
Four types of burners were tested in the laboratory furnace. The designations were as follows:
The Band Burner
1.25" Pipe Burner
HC511 with one slot
HC511 with three slots
Tests were run under various premixed ratios and various firing rates. For all tests, the CO levels were negligible (less than 10 ppm) indicating completion combustion. The raw data are presented in the Appendix.
The best correlation of NOX was shown to be with excess air level (or O2) regardless of the burner type. Figure 1 presents a plot of NOX versus Excess air level. As can be seen from the plot there is a definite trend which shows the NOX emissions to decrease as the excess air is increased. This is attributed to a flame cooling effect from the surrounding air as the gas premixed fuel burns. The reason for this is a follows. While the premix determines the initial excess O2 level (or excess air level), the surrounding air quickly diffuses into the initial flame to increase the excess O2. The high combustion temperature causes a negative draft which in turn causes the air to quickly infiltrate into the flame section. This dilution decreases the peak flame temperature. The NOX does not have a chance to be formed because the flame is immediately cooled. This process is borne out by observing the high measured excess air levels even when the premix levels were at stoichiometric ratios.
Figure 2 is a plot of NOX versus heat input for each of the burners. The plot seems to indicate that the NOX is correlated with firing intensity. However, in order to change the stoichiometry of the premix, the gas flowrate was changed while the air flowrate remained essentially constant. Therefore, the firing intensity increased as the fuel to air ratio was changed. Noted from the plot, the NOX increases with plasma index for each of the burners. To correlate the NOX with firing intensity, the premix needs to be constant. In Figure 3, the N0x levels are plotted against heat input for each burner at constant plasma index numbers. As can be seen, there is a positive correlation of NOX with firing intensity but with much less of a slope. It should be noted that the heat input is in terms of heat per time per inch of burner. This variable gives an indication of heat flux per area of burner length. As a group there appears to be a correlation between heat flux and NOX for the various burners. The three burners at a plasma index of 50 show this trend while the two burners at a plasma index of 30 show the same trend but at different N0x levels.
The effect of firing intensity, however, may be even less since at the higher firing intensities, the total excess air level was lower. This effect is seen in Figure 4 which is a plot of excess air against heat input. Note that as the heat input increases the total excess air decreases. Since NOX was shown to be an inverse function of excess air level, that function may be the overriding factor. Thus, firing intensity is considered to be a weak factor in predicting NOX.
Figure 5 presents a plot of exit gas temperature versus excess air level. As can be seen from the figure, the temperature is somewhat correlated with the excess air level. The higher the excess air the lower the temperature. The two points for the band burner appear to be outlets or due to the configuration change when the end caps were installed. The significance of the correlation is that gas temperature which is easily measured, may be a good indicator of excess air and thus, NOX level. This may be used in establishing correlations in which NOXlevels may not have to be tested at each burner site if regulatory limitations are imposed.
5.0 THEORETICAL ASPECTS
As indicated previously, the NOX levels were shown to decrease with increasing excess air levels. The formation of NOX in premixed flames was studied at varying equivalence ratios by Sarofim and Pohl (Reference 1). Figure 6 illustrates theoretical curves of the rate of NOXproduction as a function of fuel to air equivalence ratio for two fuels, hydrogen and methane. Equivalence ratio is defined as the ratio of actual fuel to air to the stoichiometric fuel to air. Thus, the lower the equivalence ratio, the higher the excess air level.
For the testing at Flynn, the equivalence ratios were calculated from the exit gas excess air levels. Referring back to Figure 1, a mean curve was put through the data points. Excess air points were taken off the mean curve and converted to equivalence ratio points. In Figure 7, the NOX levels are plotted against the equivalence ratios. As can be seen from the plot, there is a trend similar to the theoretical curve.
- Sarofim, A.F. and Pohl, J.H. "Kinetics of Nitric Oxide Formation in Premixed Laminar Flames", The Combustion Institute,