Radiation Emissions from Turbulent Diffusion Flames Burning


Radiation Emissions from Turbulent Diffusion Flames Burning...

1 downloads 82 Views 3MB Size

Article Cite This: Energy Fuels XXXX, XXX, XXX−XXX

pubs.acs.org/EF

Radiation Emissions from Turbulent Diffusion Flames Burning Vaporized Jet and Jet-like Fuels Eric D. Zeuthen† and David L. Blunck* School of Mechanical, Industrial, and Manufacturing Engineering, Oregon State University, 314 Rogers Hall, Corvallis, Oregon 97331, United States ABSTRACT: This study measured and compared the size and radiation emissions from turbulent diffusion flames burning large hydrocarbon fuels, including jet fuels. This effort is motivated by the need to understand how burning conventional or alternative jet-like fuels changes radiation emissions. Radiative heat transfer is significant because it is can alter pollutant formation and the lifetime of combustion devices. Eleven fuels were evaluated, consisting of traditional and alternative aviation fuels (e.g., Jet-A) and 1−3 component fuels. The flames were burned on a reduced-scale piloted Sydney turbulent diffusion burner with Reynolds numbers (Re) ranging from 7500 to 45 000. The vaporized fuel exited the burner near 300 °C. The radiative heat flux and the radiation intensity were measured using a radiometer (Medtherm, Model 64-0.2-20) and a mid-infrared camera (FLIR, Model SC6700), respectively. The radiant fraction (χR) is reported for the different fuels. χR typically varied by 20 000. This trend is attributed to competing effects of flame stretch, entrainment, and surface-to-volume ratio on radiation emissions. The peak radiation intensity occurs near the location of the peak heat flux, which corresponds to the visible flame height. Radiation emitted from soot and unburned hydrocarbons is predominantly observed from within the flame sheet; significant variations in radiation emissions was observed between the different fuels. In contrast, emissions from CO2 for all the fuels varied by 80

>80

>80

140.3 Fuel: A4 106.2 170.3 Fuel: C1 170.3 224.3 Fuel: C2 78.1 198.4 198.4 198.4 198.4 198.4 Fuel: C3 78.1 170.3

140.3 136 Fuel: C4 78.1 160 170.3 160 224.4 160 Fuel: C5 78.1 110 142.3 110 142.3 110 142.3 110 Fuel: C6 170.3 170 Fuel: Jet-A similar to fuel A2

× × × × × ×

× × × ×

using a bubble Gilibrator. Air to the pilot flame was metered using a rotameter downstream of an anhydrous filter. The burner was surrounded by a coflow of air with a cross section of ∼30 cm × 25 cm. The air flow was provided by two fans mounted 30 cm below the tip of the burner at 45°, relative to the coflow exit. The air was conditioned by a series of screens, with the final screen being ∼5 cm below the nozzle exit. The coflow typically exited near the burner tip at an average velocity equal to 0.7 ± 0.2 m/s, depending on the location. These values are based on measurements performed using a wind anemometer (Mastech, Model MS6252B). A subset of testing was performed without coflow or with reduced coflow. Flames without coflow were noticed to be less stable, compared to similar flames with coflow and susceptible to airflow disturbances in the laboratory. The visible flame height without coflow was reduced by

was monitored using a set of multimeters and controlled using a dimming circuit. The temperature of the fuel at the exit of the burner was estimated to be (300 ± 15 °C), based on measurements acquired using a Type K thermocouple. The thermocouple was inserted into the central fuel tube and measured the temperature 60 cm upstream of the burner exit. It is possible that cracking of the fuel occurred within the vaporizer and heater system. The residence time of the vaporized fuel within the vaporizer and tubing was between 0.5 and 4 s, depending upon the flow rate. A pilot flame of premixed ethylene and air at an equivalence ratio equal to 0.9 was used to anchor the central diffusion flame. Heat release from the pilot flame was adjusted to be less than 2% of the heat release from the central flame for each Reynolds number. Ethylene was metered through a thermal flow controller (MKS, 647C) calibrated C

DOI: 10.1021/acs.energyfuels.7b02261 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels 95% purity), was evaluated for comparison of radiation emissions from fuels that contained aromatics. The naming of the fuels was consistent with the National Jet Fuels Combustion Program (NJFCP), with the exception of C6. It is noteworthy that the lower heat values of all of the fuels agreed within 5% while the hydrogen/carbon ratio changed by as much as 20%, because of variation in the hydrocarbon classes and molecular sizes. An 11th liquid fuel (Jet-A) was acquired from a local municipal airport and used to evaluate the radiant fraction measurement techniques. This fuel was assumed to have similar molecular constituents and enthalpy of combustion as conventional jet fuel (A2). The viscosity of the vaporized fuel, which was used to determine the Reynolds number (Re) of the flame, was assumed to be the values of surrogates. The use of surrogates was required because it was not viable to measure the vaporized viscosity for the fuels. The viscosity of the surrogate mixtures was calculated using the Wilke equation21 with reported vapor properties established from the literature.22 The surrogate contained up to three constituents, representative of the most common aromatics, alkanes, and cycloparaffins in the fuel. Variation in the viscosity as a function of temperature was 1% aromatics (i.e., C1 and C4) had a peak χR value of ∼0.32. This finding is significant, because only an ∼15% decrease in the χR is observed when fuels that contain no aromatics are burned, indicating that the aromatic content of these fuels is not the strongest indicator of overall radiation emissions. The lack of sensitivity to aromatic content is attributed to two reasons. First, radiation emissions from large hydrocarbon fuels are roughly linearly correlated to the percent mass of hydrogen in the fuel, not necessarily the aromatic content. Yang and co-workers34 showed that the lower the percentage of hydrogen in large hydrocarbon fuels, the greater the tendency to produce soot and emit radiation when burned in a combustor. Consistent with these observations, in this study, the C5 and A3 fuels have ∼14% hydrogen (by mass), while C1 and C4 fuels have ∼15%. The second reason why aromatic content is not a strong indicator of radiation emissions

radii. The value of γ decreased by roughly a factor of 2 over the range of Re evaluated. As γ decreases, the relative heat losses from convection and conduction28 decreases and the temperature increases. Higher flame temperatures increase radiation emissions. Kang and colleagues28 reported that peak temperatures within dimethyl ether jet diffusion flames were almost constant when γ was constant, but increased as γ decreased. As a result, the value of χR increased as the value of Re increased in their flames.29 The decrease in χR as the Re is increased beyond 20 000 is ascribed to increased entrainment and mixing, as well as greater strain on the flames. These effects counter the effects of decreases in γ on radiation emissions. A decrease in χR as Re increases is similar to trends reported for sooting jet diffusion flames11,12 and large-scale hydrogen-jet flames.30 Increasing the exit velocity of the fuel jet increases entrainment and turbulent mixing of air within the flame. This process has a tendency to reduce the soot volume fraction and, as a result, reduce radiation emissions.12,28 Increasing the exit velocity also increases the global strain (i.e., u/D) of the flame. Increasing the strain causes the peak and path-averaged soot volume fraction to decrease31 and can decrease the flame temperatures.32 Dong and colleagues33 noted that χR decreased by ∼30% for a factor-of-5 change in the global strain rate for C2H4/H2 flames. In this study, χR decreases by ∼10% as the global strain rate increases by a factor of 2 for 20 000 ≤ Re ≤ G

DOI: 10.1021/acs.energyfuels.7b02261 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels

surrounding air lowers the exhaust temperatures and product concentrations. The width of the flame in the infrared region at the height of peak emissions is ∼50 burner diameters, which is slightly more than the visible width of the flame. The greater width in the infrared results from emissions from carbon dioxide outside of the reaction zone. Time-averaged measurements of radiation intensity emitted by C6, A1, and A4 flames are presented in Figure 11, with filters

is that emissions from CO2 and H2O have a tendency to be relatively large. This is discussed further in the next section. An outlier in the trends for χR is C6 (n-dodecane), which had much lower radiation emissions (e.g., peak value of 0.24) than the other fuels, despite having a similar hydrogen content (percentage) as fuels C1 and C5. The significant reduction in χR when n-dodecane is burned agrees with trends reported previously;35 the χR value for dodecane burned in a laminar tubular burner was ∼0.24, whereas, for kerosene, it was ∼0.34. Note that fuel C6 is a n-paraffin whereas fuels C1 and C4 are primarily composed of iso-paraffins. Zhang et al.36 noted that iso-paraffins have a greater tendency to form benzene (a precursor to soot) than n-paraffins. Their study shows that the normal or iso-structure of the molecules can have a significant impact on the radiation emissions from jet-like fuels. 3.4. Radiation Intensity Distribution. Representative time-resolved (left and middle panels) and time-averaged (right panel) radiation intensity emissions are shown in Figure 10 for a C6 flame. Radiation intensity measurements allow the

Figure 10. Time-resolved (left and middle) and time-averaged (right) radiation intensities for C6 (dodecane) flame at Re = 20 000. A 4.37 ± 0.02 μm filter was used to isolate radiation primarily emitted by CO2.

spatial and spectral distribution of radiation emissions from the flame and plume to be identified. These are some of the first such measurements to be reported for diffusion flames that are burning large hydrocarbon fuels. In the time-resolved measurements, localized high-intensity regions are separated by lowintensity regions. This distribution of emissions is caused by large-scale mixing of the fuel, air, and combustion products. Larger radiating structures are evident starting near 150 burner diameters downstream and extending along the entire length of the flame. In the average measurements, relatively low intensity regions are apparent near the burner exit (0 ≤ X ≤ 50), where line-of-sight paths through the flame are the shortest, and CO2 and soot mole fractions are the smallest. This trend is consistent with results reported by Rankin et al.4,37 for gaseous sooting flames. Further downstream (50 ≤ X ≤ 200), higher intensities are observed, with peak intensities occurring near 200 burner diameters. This location corresponds to the peak heat fluxes and is near the visible flame height. Above 200 burner diameters, intensities decrease as entrainment of the

Figure 11. Time-averaged radiation emissions collected using bandpass filters that transmit radiation primarily from CO2 (top row) and soot and unburned hydrocarbons (bottom row). The intensity values are normalized by the peak intensity measured with each filter.

used to isolate radiation primarily emitted from CO2 (top row) and primarily from soot (bottom row). The measurements can be used to gain insight into the distribution of radiation emissions from CO2, as well as soot and unburned hydrocarbons in the flames. The two columns on the left are C6 flames at Re = 30 000 and 20 000, while panels on the right are flames burning fuels A1 and A4 at Re = 20 000. The values are normalized by the peak radiation intensity measured for each filter. This allows for easier comparison of relative emissions. H

DOI: 10.1021/acs.energyfuels.7b02261 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels

hydrocarbons between fuels A1 and A4, it is noteworthy that the χR values of the A1 and surrogate flames agree within 5%. This observation indicates that global radiation emissions are dominated by emissions from CO2 and H2O for these flames. This observation is further supported by considering that the peak radiation intensity emitted from soot and unburned hydrocarbons increased by 80% between the A1 and C6 flames, whereas the χR value only increased by 30%. The root mean square of fluctuations of the radiation intensity emitted from CO2 and unburned hydrocarbons and soot are reported in Figure 12 for the C6 flames. The values are

Direct comparison of the radiation intensity measured using the two filters could be misleading, because of the different spectral bandwidths of the two filters. Peak radiation intensities are observed between 150 and 250 burner diameters downstream of the burner exit for both filters used. The size of the region with the relatively higher radiation intensity (e.g., yellow in color scale) is the largest for the emissions from CO2 while being the smallest for emissions from soot and unburned hydrocarbons. These distributions occur because much of the soot and unburned hydrocarbons are consumed through the flame sheet, whereas CO2 is present within and outside the flame. Higher radiation emissions from CO2 are observed further upstream than that from soot (compare regions of yellow). This observation is attributed to relatively higher concentrations of CO2 being closer to the burner than soot. The magnitude and distribution of the radiation intensity emitted from CO2 is relatively similar for the four flames (see the top row of Figure 10). For similar Re flames, the peak intensity varies by