- Email: [email protected]
Thermal field investigation under distributed combustion conditions
Applied Energy 160 (2015) 477–488
Contents lists available at ScienceDirect
Applied Energy journal homepage: www.elsevier.com/locate/apenergy
Thermal field investigation under distributed combustion conditions Ahmed E.E. Khalil, Ashwani K. Gupta ⇑ Department of Mechanical Engineering, University of Maryland, College Park, MD 20742, USA
h i g h l i g h t s Examined the thermal field behavior under conventional and distributed combustion. N2 and CO2 are used to simulate entrained combustion gases from within the combustor. Reduced oxygen concentration (<15%) is key to achieve thermal field uniformity. Distributed combustion reduced temperature variation by 50% axially & 60% radially. Significant noise and emissions reduction was achieved via distributed combustion.
a r t i c l e
i n f o
Article history: Received 18 July 2015 Received in revised form 7 September 2015 Accepted 11 September 2015
Keywords: Colorless distributed combustion Ultra low NOx High intensity distributed combustion Reactive gas entrainment Thermal field uniformity
a b s t r a c t Distributed combustion has demonstrated significant performance gains, especially on combustion efficiency and near zero pollutants emission. Controlled mixture preparation between air, fuel and internal hot reactive gases prior to mixture ignition is a critical requirement to achieve distributed combustion condition. Though distributed combustion have been extensively studied using a variety of geometries, heat loads and intensities, and fuels, limited information is available on the role of hot reactive gas entrainment and the resultant thermal field uniformity. In this paper, the impact of internal entrainment of hot reactive gases on thermal field uniformity and pollutants emission is investigated. A mixture of nitrogen and carbon dioxide was introduced to the fresh air stream prior to mixing with the fuel and its subsequent combustion to simulate the product gases from within the combustor. Increase in the amounts of nitrogen and carbon dioxide (simulating increased entrainment) significantly reduced pollutants emission, enhanced thermal field uniformity, and increased the reaction volume to occupy larger portion of the combustor. This was evident through spatial temperature measurements in the combustor along with the enhanced distribution of the flame visible signature and OH⁄ chemiluminescence signal. The temperature data demonstrated that lowering oxygen concentration from 21% to 15%, through increased entrainment, promoted distributed combustion conditions with lower overall temperature rise throughout the combustor. In addition, the peak temperature regions associated with swirl burners disappeared, eliminating most of the hot spots in the combustor. The enhanced thermal field uniformity and reduced temperature variation provided ultra-low emissions, demonstrating the impact of enhanced thermal flowfield uniformity on emissions. Experiments performed at different equivalence ratios and entrained gas temperatures demonstrated similar behavior of thermal field uniformity and ultra-low emissions. Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction The increase in fossil fuels as a local energy source to satisfy the states’ energy needs, along with concerns about global warming and climate change, have motivated combustion researchers to develop new combustion methods that has minimal impact on the environment and high efficiency. The potential for natural gas and shale gas deployment in electricity and power generation ⇑ Corresponding author. Tel.: +1 301 405 5276; fax: +1 301 314 9477. E-mail address: [email protected] (A.K. Gupta). http://dx.doi.org/10.1016/j.apenergy.2015.09.058 0306-2619/Ó 2015 Elsevier Ltd. All rights reserved.
have been fostered by their increased availability as a local energy source and their lower carbon emission (compared to other fossil fuels such as coal). To ensure this mandated environmentally friendly performance, future combustion systems shall achieve even lower pollutants (including NOx, CO, unburned hydrocarbons and soot) while minimizing CO2 emissions. To this end, distributed combustion (Colorless Distributed Combustion, CDC [1–4]), among other technologies (i.e., FLOX, MILD) [5–7], has been shown to provide significant benefits of reducing the emissions of NO and CO along with stable combustion, reduced noise, no flame fluctuations and combustion instability. CDC combustors have also
478
A.E.E. Khalil, A.K. Gupta / Applied Energy 160 (2015) 477–488
demonstrated fuel flexibility, handling a variety of liquid and gaseous fuels with no modification to the combustor [8]. The flames in distributed combustion do not show any visible flame signatures so that the flame so formed is termed colorless due to negligible visible emission as compared to conventional flames. The performance benefits of CDC have been demonstrated using a variety of geometrical arrangement for the combustor flowfield, air, and fuel injection schemes [9–12]. The focus in these investigations was on the internal entrainment of hot reactive species and gases and their subsequent mixing with the freshly introduced air and fuel. This entrainment and the subsequent adequate mixing prior to ignition forms pivotal step in achieving distributed reactions. Distributed reactions are characterized by a lower reaction rate over the entire volume of the combustor, as opposed to the concentrated flame front characterized by high reaction rates with local hot spots, to result in the same fuel consumption with lower temperature rise in the combustor. This low reaction rate is achieved through lowering the oxygen concentration in the reactants, and increasing the temperature of the reactants (both achieved simultaneously through the entrainment of hot reactive gases from within the combustor). The distributed combustion regime not only avoids the formation of thin reaction zone but also the hot-spot zones in the flame, which help mitigate thermal NOx formation and emission from the Zeldovich thermal mechanism [13,14]. The impact of the amount of gas entrainment and recirculation on reaction distribution and pollutants emission have been investigated with focus on determining the minimum entrainment requirements for distributed reactions to occur (hot gas recirculation/oxygen concentration) [15]. In these investigations, a swirl burner was used with focus on determining emissions (NO and CO) and flame behavior (radicals’ chemiluminescence signal) with different amounts of recirculation. A mixture of nitrogen and carbon dioxide (90–10% by volume) is used to simulate the product gases. These investigations showed that reaction distribution is significantly enhanced with increased entrainment, lowering oxygen concentration in the ready-to-ignite mixture. This reaction distribution was fostered at oxygen concentration below 15%, to result in ultra-low emissions. However, limited information is available on the thermal field uniformity within the reaction zone. Swirl burners thermal field have been studied in details with focus on mean and fluctuating temperatures. Hedman and Warren [16] have used coherent anti-Stokes Raman spectroscopy (CARS) to measure temperature in a dual swirl combustor, where they showed that the temperature increases across the centerline then decreases near the exit, with significant radial change across the swirl boundary. Similar behavior was also shown by Keck et al. [17]. Cheng et al. [18] used R-type thermocouples showing the same radial temperature change across the swirl stabilized combustor. The temperature profile has also been numerically examined [19,20] with similar behavior as those measured experimentally. In this paper, the thermal field uniformity is evaluated from the temperature measurement throughout the reaction field under distributed combustion conditions. Comparison between the temperature values under distributed reaction conditions and normal combustion will aim to outline the significant thermal field uniformity expected under distributed combustion condition. Various conditions will be examined to identify the cause of this field uniformity and its direct impact. These conditions include variations of stoichiometry, heat load, and different diluents temperature. Nitrogen and carbon dioxide were selected as they represent majority of the product gases. They were mixed in a 90% N2–10% CO2 by volume simulating product gases near stoichiometry conditions. It is recognized that this ratio changes as the equivalence
ratio becomes leaner, the diluting gases mixture (90–10%) was kept constant. Any minor deviation from the actual composition will have minimal impact on the results as nitrogen and carbon dioxide behave similarly in flames. Laminar flame speed and flame temperature for methane-air flames diluted with nitrogen and/or carbon dioxide have shown to exhibit similar behavior [21,22]. Diluting the reactants with a nitrogen–carbon dioxide–water vapor mixture also resulted in similar behavior to that of nitrogen [23]. 2. Experimental facility The experiments were performed using a swirl burner fueled with methane. Details of this swirl burner can be found elsewhere [24]. The performance of this swirl burner has been studied in terms of emissions and velocity profiles using methane and hydrogen enriched methane [24,25]. To simulate product gas entrainment and recirculation, and reduce oxygen concentration in the mixture prior to ignition, different amounts of N2–CO2 mixture were added to the air upstream of the burner. Fuel was injected at the center of the swirler in a non-premixed configuration. Air and nitrogen flowrates were controlled by laminar flow controllers with an accuracy of ±0.8% of reading and ±0.2% of full scale leading to an overall accuracy of 1.5% of the reading. Methane and carbon dioxide flowrates were controlled through gravimetric flow controllers with an accuracy of 1.5% of full scale. Detailed temperature measurements were performed using a K-type thermocouple with NI-DAQ (data acquisition) system calibrated using blackbody calibrator (Omega BB-4A) resulting in an accuracy of ±3%. This accuracy is established based on the maximum difference between the thermocouple reading and the reference thermocouple in the calibrator. The thermocouple was mounted on a traverse mechanism to allow measurement at different axial and radial positions from within the flames. The thermocouple was continuously sampled with the mean of 50 readings reported herein as the average temperature. An ICCD (Intensified Charge-Coupled Device) camera coupled to a narrow band filter for OH⁄ chemiluminescence detection (UV interference filter centered at 307 nm with a FWHM of ±10 nm) was used to evaluate the flame behavior upon the insertion of the thermocouple into the combustor. It is imperative to ensure that the flame characteristic do not change upon the movement of the thermocouple. This is especially critical under distributed combustion conditions where the flame might anchor on the thermocouple. For pollutants emission, the products of combustion were continuously sampled at the exit of the burner. The concentration of NO was measured using a NO–NOx chemiluminescent gas analyzer, CO concentration was measured using the non-dispersive infrared method and O2 concentration (used to correct the NO and CO emissions at standard 15% oxygen concentration) was measured using galvanic cell method. During a single experiment, measurements were repeated at least three times for each configuration and the uncertainty was estimated to be about ±0.5 ppm for NO and ±10% for CO emission. The experiments were repeated at least three times to ensure good repeatability of the experimental data obtained. The experimental rig is shown in Fig. 1 with the flame at two different conditions of normal air combustion wherein the swirl structure is dominant (left photo), and reduced oxygen concentration combustion showing near distributed combustion with less visible emission (right photo). 3. Experimental investigations Table 1 summarizes the conditions for the specific experimental conditions reported here along with the variables manipulated for each case. For each heat load, the fuel flow rate was kept constant
479
A.E.E. Khalil, A.K. Gupta / Applied Energy 160 (2015) 477–488
Fig. 1. Experimental test rig with flame using normal air (left) and reduced oxygen concentration air (right).
Table 1 Experimental parameters.
1 2 3 4 5 6 7 8 9
Heat load (kW)
Heat release intensity (MW/m3 atm)
Mixture temperature (K)
Equivalence ratio
Oxygen concentration (%)
Inlet Reynold’s number
3.25 3.25 3.25 3.25 3.25 4.875 3.25 3.25 3.25
2.4 2.4 2.4 2.4 2.4 3.6 2.4 2.4 2.4
300 300 300 300 300 300 600 600 600
0.9 0.7 0.6 0.9 0.7 0.9 0.9 0.7 0.9
21 21 21 13.8 16.51 21 12.38 14.91 13.8
5000 6000 7000 7000 7500 7000 11,000 10,500 9500
Table 2 Temperature field parameters.
1 2 3 4 5 6 7 8 9
Oxygen concentration (%)
Diluents temp. (K)
Maximum axial temp. (K)
Combustor exit temp. (K)
Temp. difference (max-exit)
21 21 21 13.8 16.51 21 12.38 14.91 13.8
– – – 300 300 – 600 600 600
1248 1191 1144 1104 1093 1350 1144 1106 1164
938 918 897 953 918 1064 1018 995 985
310 273 247 151 175 286 126 111 179
while the air flow rate was changed to change the equivalence ratio. For every equivalence ratio, air and fuel flow rates were kept constant while the amount of N2–CO2 mixture was increased to lower the oxygen concentration in the mixture prior to ignition. The nitrogen-carbon dioxide mixture was heated to 600 K to study the impact of mixture temperature on the combustion behavior and results obtained are given in Table 2.
lence ratios shown herein and agrees favorably well with the trends reported in the literature [16–18]. The high temperature at about 0.6X/R is attributed to the flame stabilization at the tip of the swirler. Decreasing the equivalence ratio resulted in lower temperature values while maintaining the profile shape (see Figs. 2 and 3). The radial temperature profile is shown in Fig. 3 at three different heights, 1.9R, 3.8R and 5R, where R is the radius of the combustor inlet. The profiles were consistent at all the equivalence ratios examined. The profiles, at each height, looked similar but the temperature value decreased with decrease in equivalence ratio similar to the axial profiles behavior. The increase at r = 1.125R for X = 1.9R can be related to the thermocouple being in the thin flame zone, where the maximum temperature was recorded. Radial temperature profile was also obtained at a height of 7.6R, where it was found to be similar to that of 5R. The increase in temperature near the exit (at 28R) is attributed to the smaller exit diameter, increasing the gases velocity and radial mixing with gases away from the central axis location that are expected to be at higher temperatures. 4.2. Dilution impact on thermal field
4. Results and discussion 4.1. Effect of equivalence ratio The initial experiments focused on measuring the temperatures to obtain axial and radial profiles with no entrainment (cases 1, 2, and 3), establishing a baseline for the combustor performance. These measurements are shown in Figs. 2 and 3. In both figures, the abscissa (height, X, and radial position, r, respectively) is normalized by the inlet radius, R. In Fig. 2, one can see that the temperature increases through the combustor up to a distance of 7X/R where the maximum temperature occurs on the centerline. Afterwards, the temperature start to decrease until the exit at 30.8X/R. This behavior is consistent for the three equiva-
Upon examination of the temperature profiles for normal air combustion (no dilution), the temperature measurements with dilution under distributed combustion conditions were determined to reveal the flames signatures. Fig. 4 shows the visible flame signature under different dilution amounts that resulted in lower oxygen concentration in the fresh reactants. Fig. 5 shows the axial temperature profile at an equivalence ratio of 0.9 with no dilution (21% O2) and maximum stable dilution (13.8% O2), representing cases 1 and 4. For no dilution, the difference between the maximum and exit temperature was about 300%, which was significantly reduced to 100% under distributed combustion conditions. In addition, the temperature increase/ decrease ramp is a much smoother under low oxygen concentra-
480
A.E.E. Khalil, A.K. Gupta / Applied Energy 160 (2015) 477–488
1300 Phi=0.9, O2=21%
1250
Phi=0.7, O2=21%
1200
Phi=0.6, O2=21%
Temperature [K]
1150 1100 1050 1000 950 900 850 800
0
5
10
15
20
25
30
35
X/R Fig. 2. Axial temperature profiles along the longitudinal central axis for different equivalence ratio with no dilution.
Phi=0.9, O2=21%
Phi=0.7, O2=21%
Phi=0.6, O2=21%
Temperature [K]
1500
h=1.9R
1400 1300 1200 1100 1000
Temperature [K]
1500
h=3.8R
1400 1300 1200 1100 1000
Temperature [K]
1500
h=5R
1400 1300 1200 1100 1000
0
0.5
1
1.5
2
r/R Fig. 3. Radial temperature profiles for different equivalence ratio with no dilution at different heights.
tion conditions. Under such conditions, the flame is slightly lifted (as seen from a comparison of the extremes in Fig. 4), which led to lower temperatures directly above the swirler in the dilution case up to about a distance of X = 2R. The radial profiles for the two cases discussed herein (cases 1 and 4) are shown in Fig. 6. One can see that there is no increase
in temperature at 1.9R for the lower oxygen concentration case; instead, the temperature is decreased as one progressively moves away from the longitudinal centerline of the combustor, see blue symbols for both cases. The temperature profile at higher heights (3.8R, 5R, and 7.6R) show minimal variation (50 K) for the dilution case as compared to 100 K for case 1 (using 21% oxygen).
481
A.E.E. Khalil, A.K. Gupta / Applied Energy 160 (2015) 477–488
Fig. 4. Flame images for different oxygen concentration (with change in diluents amount) at an equivalence ratio of 0.9.
1300
Phi=0.9, O2=21% Phi=0.9, O2=13.8%
Temperature [K]
1200 1100 1000 900 800 700 600
0
5
10
15
20
25
30
35
X/R Fig. 5. Axial temperature profile at equivalence ratio of 0.9 with and without dilution.
O2=21%, h=1.9R O2=13.8%, h=1.9R
O2=21%, h=3.8R O2=13.8%, h=3.8R
O2=21%, h=5R O2=13.8%, h=5R
O2=21%, h=7.6R O2=13.8%, h=7.6R
1500 1400
Temperature [K]
1300 1200 1100 1000 900 800 700
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
r/R Fig. 6. Radial temperature profile at equivalence ratio of 0.9 with and without dilution at different heights.
In addition, for the profiles at 5R and 7.6R, the temperature is more uniform radially as compared to that at 3.8R. Dilution also resulted in lower temperature values at different heights as compared to normal air combustion. To further understand the behavior at h = 1.9R, OH⁄ chemiluminescence signal was compared for both cases with focus on flame position. Fig. 7 shows this comparison, where it is evident that lowering oxygen concentration via entrainment resulted in a decrease in the chemiluminescence signal as compared to the no dilution
case. The low oxygen concentration case (13.8%) is also shown on the right image but at a different scale (being 0–2500 a.u. instead of 0–10,000 a.u.). It can be seen that at a height of 1.9R, the OH⁄ signal decreases as one moves radially away from the center then increases again. This explains the temperature profile shown in Fig. 6 where the temperature first drops and then increases again. Also OH⁄ shows that this is the height where the reaction starts to occur, with minimal reaction occurring upstream of this location, which agrees well with the low temperature measured at the center
482
A.E.E. Khalil, A.K. Gupta / Applied Energy 160 (2015) 477–488
h=3.8R
h=1.9R
Phi=0.9, O2 =13.8%
Phi=0.9, O2 =13.8%
Phi=0.9, O2 =21% 0
10000
0
2500
Fig. 7. OH⁄ chemiluminescence intensity distribution for standard air combustion (left) and reduced oxygen concentration (center) and with thermocouple position at 1.9R (right).
at position lower than 1.9R for this case, see Fig. 5. This in contrast to the case with no dilution, where at this height, high reaction rate occurs at the swirl lobes resulting in high temperatures. Further downstream at a height of 3.8R, for the low oxygen concentration case, the reaction occurs across the radius of the reactor resulting in a more uniform radial temperature profile. Moreover, Fig. 7 shows that the thermocouple had limited impact on the OH⁄ chemiluminescence as no field disturbance was noticed. Similar axial temperature behavior was also observed at lower equivalence ratio (cases 2 and 5); where the temperature rise across the center was lower for the lower oxygen concentration case, see Fig. 8. The temperature difference between the maximum temperature and the exit temperature was found to be less in the low oxygen concentration case. In addition, the maximum temperature was reduced in the low oxygen concentration case. Radial temperature profiles showed a similar trend to that of Fig. 6 where the temperature profile was more uniform for the low oxygen concentration case. 4.3. Oxygen concentration impact on thermal field The impact of lowering the oxygen concentration over using a lower equivalence ratio was examined and the results on temperature distribution for cases 1, 3, 4, and 6 are shown in Fig. 9. Case 1 represents the base case at an equivalence ratio of 0.9. Case 3 represents dilution through adding more air (lowering equivalence
ratio to 0.6 while maintaining oxygen concentration at 21%). This air addition resulted in an increase in inlet velocity and Reynold’s number (7000). Case 4 represents dilution through lowering oxygen concentration through N2–CO2 addition. The added gas mixtures also increased the velocity and Reynold’s number to match the conditions of case 3. Case 6 represents an increase in air and fuel flow rates to maintain an equivalence ratio of 0.9% and 21% oxygen concentration with an increased velocity and Reynold’s number to match cases 3 and 4. The centerline temperature distribution are shown in Fig. 9. Comparing the different cases, one can see that increase in the flowrates (air and fuel) resulted in the similar temperature profile but of increased values (cases 1 and 6), limiting the impact of increased inlet velocity on its own. Comparing cases 1 and 3, one can see that the temperature profile shape is preserved, but the values are lowered. This is attributed to the leaner conditions (lower equivalence ratio) and the resulting lower temperature for this lower equivalence ratio. The temperature profile only changes upon lowering of the oxygen concentration to result in distributed combustion (case 4). Case 4 has a lower temperature rise (peak) across the combustor as compared to the other cases. The maximum temperature is even lower than that resulting from an equivalence ratio of 0.6. This behavior is supported through OH⁄ chemiluminescence and visible flame signature. Only at lower oxygen concentration did the flame transfer to a distributed regime across the combus-
1200
Phi=0.7, O2=21%
Temperature [K]
1100
Phi=0.7, O2=16.51%
1000 900 800 700 600 500
0
5
10
15
20
25
30
X/R Fig. 8. Axial temperature profile for equivalence ratio of 0.7 with and without dilution.
35
483
A.E.E. Khalil, A.K. Gupta / Applied Energy 160 (2015) 477–488
1400
Case1, Phi=0.9, O2=21% Case 3, Phi=0.6, O2=21%
1300
Case 4, Phi=0.9, O2=13.8% Case 6, Phi=0.9, O2=21%
Temperature [K]
1200 1100 1000 900 800 700 600 500
0
5
10
15
20
25
30
35
X/R Fig. 9. Axial temperature profile for cases 1, 3, 4, and 6.
Case 1, Phi=0.9, O2=21%
Case 3, Phi=0.6, O2=21%
Case 4, Phi=0.9, O2=13.8%
0
Case 6, Phi=0.9, O2=21% 16000
Fig. 10. OH⁄ Chemiluminescence intensity distribution for cases 1, 3, 4, and 6.
tor. For cases 1, 3, and 6, the swirl structure was preserved with hot spots existing at the swirl lobes at the boundaries of the central toroidal recirculation zone. This swirl structure only disappeared through lowering oxygen concentration (case 4) to result in distributed combustion, see Fig. 10. 4.4. Effect of dilution temperature In all the previous experiments, the N2–CO2 mixture was introduced to the fresh reactants at room temperature. These experiments were aimed at identifying the impact of oxygen concentration reduction (which can be achieved in actual combustion systems through product gas recirculation). However, such recirculation will increase the fresh mixture temperature. In the case of internal recirculation, the product gases recirculated will have temperature of about 1500 K or higher. In these experiments, the N2–CO2 mixture was preheated to higher temperatures with view to understand the impact of diluents temperature on thermal field uniformity and emissions. These experiments were performed at the same heat load of 3.25 kW and heat release intensity (HRI) of 2.4 MW/m3 atm. Fig. 11 shows the axial temperature profile for the base case of no dilution (O2 = 21%), dilution mixture at 300 K (O2 = 13.8%), and two dilution cases (O2 = 13.8% and 12.38%) at diluents temperature of 600 K. As the temperature of
the diluents was increased, the combustor was stable at lower oxygen concentrations as opposed to the diluents at room temperature [15]. As the dilution temperature was increased, the overall temperature profile shifted to higher values while maintaining the same profile. This can be seen comparing the two cases with oxygen concentration of 13.8%. The maximum temperature slightly increased from 1100 K to 1160 K. In addition, the reaction zone shifted slightly upstream (as the temperature increase was higher in the case of diluents temperature = 600 K). For the same temperature, increase in the dilution amounts resulted in a lower temperatures throughout the combustor. For all the dilution cases shown here, the temperatures were lower as compared to the case with no dilution, with a smooth temperature increase/decrease throughout the flow regime. For the case of no dilution and diluents at 300 K, the combustor exit temperature was almost the same (945 K), when the diluents temperature increased, the combustor exit temperature increased, respectively. As more diluents were added (further reduction in oxygen), the combustor exit temperature was found to be higher (985 K for 13.8%, 1018 K for 12.38%). This increase in temperature can be related to the increased energy input for that case (higher amounts of N2–CO2 mixture at 600 K to further reduce O2 concentration).
484
A.E.E. Khalil, A.K. Gupta / Applied Energy 160 (2015) 477–488
1300
O2=21% O2=13.8%, T=600K O2=12.38%, T=600K O2=13.8%
1200
Temperature [K]
1100 1000 900 800 700 600
0
5
10
15
20
25
30
35
X/R Fig. 11. Axial temperature profile for different dilution cases at equivalence ratio of 0.9 (cases 1, 4, 7, and 9).
O2=21%, h=1.9R O2=13.8%, T=600K, h=1.9R O2=12.38%, T=600K, h=1.9R
O2=21%, h=3.8R O2=13.8, T=600K, h=3.8R O2=12.38%, T=600K, h=3.8R
O2=21%, h=5R O2=13.8%, T=600K, h=5R O2=12.38%, T=600K, h=5R
1500 1400
Temperature [K]
1300 1200 1100 1000 900 800 700
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
r/R Fig. 12. Radial temperature profile for different dilution cases at equivalence ratio of 0.9 and diluents temperature of 600 K (cases 1, 7, and 9).
Fig. 12 shows the radial temperature profile for the reduced oxygen concentration with diluents at 600 K. The temperature profiles are similar to those obtained with the diluents at room temperature (see Fig. 6). The radial temperature values slightly decreased with increased dilution (lowered oxygen concentration) for the different heights shown here. The profile at h = 1.9R is consistent with what was shown in Fig. 6 and explained by the reaction zone location (see Fig. 7). The diluents temperature influence on the thermal field uniformity was also examined. The radial temperature profiles were compared for oxygen concentration of 13.8% using diluents temperature of 300 K and 600 K. The results obtained are shown in Fig. 13. Increasing the temperature of the diluents resulted in an increase in the temperature values at the examined heights while maintaining a similarity in the temperature profile shape. More importantly, the radial temperature profile uniformity was maintained at h = 3.8R, 5R and 7.6R as a result of the distributed combustion conditions achieved at this oxygen concentration (13.8%) and dilution temperatures. Experiments were also performed at an equivalence ratio of 0.7. Increase in the diluents temperature allowed stable operation at lower oxygen concentration. Fig. 14 shows the axial temperature
profile for equivalence ratio of 0.7 with no dilution, O2 = 16.51% (T = 300 K) and O2 = 14.91% (T = 600 K). Similar to the profiles for equivalence ratio of 0.9, increasing the diluents temperature slightly increased the temperature values while maintaining the uniform temperature profile. Also increased diluents temperature resulted in an increase in the combustor exit temperature. 4.5. Thermal field uniformity The temperature field for each case was analyzed to obtain the maximum temperature along the longitudinal axis and the difference between that maximum and exit temperature to provide information on the temperature field uniformity. As the oxygen concentration was reduced, to steer towards distributed reaction conditions, the maximum temperature was significantly reduced with a smaller temperature difference. At equivalence ratio = 0.9, the maximum temperature was reduced from 1248 K (O2 = 21%) to 1104 K (13.8%) with the temperature difference reduced from 310 to 151 K. Further oxygen reduction (assisted by increased diluents temperature) led to a slight increase in the maximum temperature (1144 K) but a reduced temperature difference (126 K). The same behavior was also
485
A.E.E. Khalil, A.K. Gupta / Applied Energy 160 (2015) 477–488
T=300K, h=1.9R
T=300K, h=3.8R
T=300K, h=5R
T=300K, h=7.6R
T=600K, h=1.9R
T=600K, h=3.8R
T=600K, h=5R
T=600K, H=7.6R
1300
Temperature [K]
1200 1100 1000 900 800 700
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
r/R Fig. 13. Radial temperature profile for different dilution temperatures at equivalence ratio of 0.9 and oxygen concentration of 13.8% (cases 4 and 7).
1300
O2=21% O2=16.51%, T=300K O2=14.91%, T=600K
Temperature [K]
1200 1100 1000 900 800 700 600
0
5
10
15
20
25
30
35
X/R Fig. 14. Axial temperature profile for different dilution temperatures at equivalence ratio of 0.7 (cases 2, 5 and 8).
demonstrated at an equivalence ratio of 0.7. For all the cases, lowering oxygen concentration in the fresh reactants led to 40–60% reduction in temperature variation. The same analysis was performed on different radial profiles. Table 3 gives the maximum temperature and minimum temperature along with their difference for the profiles with maximum variations. Similar to longitudinal profiles, lowering oxygen concentration resulted in lower temperature variation (60 K instead of 262 K in case of phi = 0.9). Increase in diluents temperature slightly affected the radial profile variation. For all cases, low oxygen concentration resulted in 60% or more reduction in temperature variation. The thermal field uniformity can be expressed in percentage temperature variation as:
DT ¼ ½ðT max T exit Þ=ðT exit T in Þ 100
ð1Þ
where DT is the percentage temperature change, Tmax is the maximum temperature along measured line, Texit is the combustor exit temperature and Tin is the air inlet temperature, 300 K. Following the above definition in Eq. (1), the percent temperature change for standard air combustion, cases 1, 2, 3 and 6, was 48.6%, 44.2%, 41.4% and 37.5%, respectively. On the other hand,
for the cases demonstrating distribution combustion, this percentage varies between 23.1% and 28.3% for the diluents at 300 K (cased 4 and 5). For the diluents at 600 K, this percentage varied around 17% (cases 7 and 8) and 26% (case 9). Thus lowering the oxygen concentration enhanced thermal field uniformity and reduced the percentage change from 48.6% (case 1) down to 23.1% (case 4). Upon increasing the diluents temperature, this value decreased to 17% (case 7). The same trend can be seen for the lower equivalence ratio of 0.7 (case 2: 44.2%, case 5: 28.3%, and case 8: 26%) Similar behavior was observed for radial temperature distribution. For normal air combustion, the percentage temperature variation varied between 15% and 23%, cases 1–3, while this value was about 7% for distributed combustion conditions, cases 4, 5, 7 and 9. These percentages were obtained using the following equations:
DT ¼ ½ðT max T min Þ=ðT max T in Þ 100
ð2Þ
where Tmin is the minimum temperature along the measured line. The temperature values used for calculations are those of the height presenting maximum temperature variation, see Table 3. These values emphasize the thermal field uniformity achieved under distributed combustion conditions. This uniformity can be
486
A.E.E. Khalil, A.K. Gupta / Applied Energy 160 (2015) 477–488
Table 3 Radial temperature field parameters. Oxygen concentration (%) 1 2 3 4 5 7 9 a b
Diluents temp. (K)
21 21 21 13.8 16.51 12.38 13.8
Height a
– – – 300 300 600 600
1.9R 1.9Ra 1.9Ra 7.6Ra,b 7.6Ra,b 7.6Ra,b 7.6Ra,b
Maximum radial temp. (K)
Minimum radial temp. (K)
Temp. difference (max–min)
1449 1395 1247 1156 1126 1204 1204
1187 1139 1100 1096 1080 1130 1146
262 256 147 60 46 74 58
This is the height with maximum fluctuations, temperature profiles at other heights demonstrated lower temperature variation. The profile at 1.9R was excluded as this was before the reaction initiation as indicated through chemiluminescence data.
further enhanced by increasing the temperature of the diluents and decreasing the oxygen concentration as evident from the trend shown in comparing cases 1 (no dilution), 4 (13.8% O2, T = 300 K), and 7 (12.38%O2, T = 600 K), as these values were: axially: 48.6%, 23.1%, 17.5%, and radially: 22.8%, 7%, 7%, respectively.
4.6. Pollutants emission The benefits of distributed combustion are not only limited to the thermal field uniformity (which is a direct result of the lower reaction rate distributed over a larger volume of the combustion), but also on the combustor performance in terms of reduced emissions. These benefits have been previously examined for different combustors with focus on stability and pollutants emission under
different geometries [1–4]. In this work, the emissions for the configurations examined herein are discussed in order to link thermal field uniformity within the combustor with emissions.
4.6.1. Noise emission One of the characteristic benefits of distributed combustion is the reduced noise emission. During experiments, the combustor noise was significantly reduced when the combustor transitioned to distributed combustion mode (through added dilution, lowering oxygen concentration). For an equivalence ratio of 0.9, the combustor noise recorded was 80.3 dB (A-weighted) operating at no dilution (21% O2, case 1). This noise was reduced to 63.5 dB upon transition to distributed combustion mode (13.8% O2, no preheat to the diluents, case 4) leading to a significant noise reduction.
0.006
21% O2
0.005
Signal
0.004 0.003 0.002 0.001 0
0
1000
2000
3000
4000
5000
6000
7000
8000
Frequency [Hz] 0.006
13.8% O2 0.005
Signal
0.004 0.003 0.002 0.001 0
0
1000
2000
3000
4000
5000
6000
7000
8000
Frequency [Hz] Fig. 15. Noise spectrum from the combustor operating under normal conditions and with decreased oxygen concentration (distributed combustion condition).
A.E.E. Khalil, A.K. Gupta / Applied Energy 160 (2015) 477–488 Table 4 NO and CO emissions (corrected to 15% O2 in exhaust gas) for the different cases.
1 2 3 4 5 6 7 8 9
Oxygen concentration (%)
Diluents temp. (K)
NO (ppm)
CO (ppm)
21 21 21 13.8 16.51 21 12.38 14.91 13.8
– – – 300 300 – 600 600 600
13.3 6.6 2.26 1.8 0.8 13.3 1.9 0.75 2.66
24.8 3.11 3.7 5.3 2.87 40 4.5 2.87 3.93
The flow noise (reactants flowing without reaction) is about 54 dB. This indicates that distributed combustion resulted in approximately 17 dB decreased noise emission than normal combustion case and only 9.5 dB increase over the non-reacting condition. The spectral noise signature was also captured for the combustor operating under normal combustion (case 1) and distributed combustion conditions (case 4) and is shown in Fig. 15. The background noise was subtracted from both signals. Comparing both signals, one can see the disappearance of the combustor dominant frequency from the swirl (at around 500 Hz) when the combustor operated at distributed combustion mode to result in the low overall noise as demonstrated by some 17 dB reduced noise emission. 4.6.2. NO and CO emission The thermal field uniformity reflected on the pollutants emission also. NO and CO emissions for the cases discussed herein are shown in Table 4 where the distributed combustion cases (low oxygen concentration cases) offered the lowest NO emissions without any increase in CO emission or any associated instabilities. These emissions were measured at the exit (at X = 30.8R). Under distributed combustion conditions, NO emissions were 2.7 ppm or lower for the cases at equivalence ratio of 0.9 (for diluents temperature of 300 K and 600 K, cases 4, 7 and 9). This is in contrast to 13.3 ppm NO recorded from normal flame case (no dilution, case 1). CO emission was also lowered with distributed combustion (5.3 ppm or less as opposed to 24.8 ppm of CO for normal combustion case). The same behavior was also demonstrated at equivalence ratio of 0.7, where distributed combustion (cases 5 and 8) emitted less than 1 ppm of NO while normal flame (case 2) emitted 6.6 ppm of NO. CO emissions were about the same. These pollutants emission outlines the significant role of distributed combustion on pollutants emissions reduction.
487
to the concentration thin reaction zone at the shear layer of the swirl flow (at the boundaries of the central toroidal recirculation zone). Visible flame signature demonstrated the same behavior as lowering the oxygen concentration through increased dilution resulted in faint blue distributed flame as opposed to that obtained from normal swirl flame. The results revealed that the reaction distribution is a result of lower oxygen concentration rather than the dilution itself. Lowering the equivalence ratio diluted the flame (due to the added nitrogen) but did not result in distributed reactions (as the oxygen concentration remained at 21%). On the other hand, lowering oxygen concentration, while maintaining the same flow conditions (as that of the low equivalence ratio) resulted in a distributed reaction along with the disappearance of the swirl structure to outline the importance of oxygen concentration as a means to achieve distributed combustion. Increase in the diluents temperature slightly increased the temperature values across the combustor while maintaining the smooth temperature change/ramp within. Increase in the diluents temperature led to stabilizing the flame at lower oxygen concentrations to result in improved thermal field uniformity at these lower concentrations as compared to that with diluents at room temperature. Achieving distributed combustion, evidenced by distributed visible emissions and thermal flow field, resulted in significant reduction of harmful pollutants emission. This includes noise as well as chemical pollutants. The combustor noise was reduced by 17 dB when the combustor transition from swirl flame to distributed combustion mode. In addition, the characteristic frequency of the noise associated with the swirl burner disappeared upon transition to distributed combustion. Distributed combustion also lowered NO and CO emissions as compared to normal swirl flames. NO emissions were reduced by some 80% as the combustor transitioned to distributed combustion with similar reduction in CO at an equivalence ratio of 0.9. Similar reductions were also achieved at an equivalence ratio of 0.7, thus revealing the impact of distributed combustion and its thermal field uniformity on pollutants emission and combustor performance. Acknowledgment This research was supported by ONR and is gratefully acknowledged. References
5. Conclusions Temperature profiles measured along the centerline of a swirl burner and across the combustor radius at defined heights has demonstrated the thermal field uniformity achieved under distributed combustion conditions. Temperature variation along the centerline and across the combustor radius was significantly reduced upon lowering the oxygen concentration of the reactants. This is achieved through using N2–CO2 mixture simulating internal hot reactive combustion gases recirculation and entrainment from within the combustor. Temperature measurements across the combustor axis revealed that distributed combustion lowered the maximum temperature and resulted in approximately 50% reduction in temperature variation across the combustor. Similar temperature variation reduction was demonstrated at different radial positions examined. The thermal field uniformity data supported the OH⁄ chemiluminescence data which also showed distributed reaction to occur across the combustor at low oxygen concentrations as opposed
[1] Arghode VK, Gupta AK. Effect of flow field for colorless distributed combustion (CDC) for gas turbine combustion. J Appl Energy 2010;87:1631–40. [2] Arghode VK, Gupta AK. Role of thermal intensity on operational characteristics of ultra-low emission colorless distributed combustion. J Appl Energy 2013;111:930–56. [3] Khalil AEE, Gupta AK. Swirling distributed combustion for clean energy conversion in gas turbine applications. J Appl Energy 2011;88:3685–93. [4] Khalil AEE, Gupta AK. Distributed swirl combustion for gas turbine application. J Appl Energy 2011;88:4898–907. [5] Wunning JA, Wunning JG. Flameless oxidation to reduce thermal NO formation. Prog Energy Combust Sci 1997;23:81–94. [6] Lammerl O, Schutz H, Schmitz G, Luckerath R, Stohr M, Noll B, et al. FLOX combustion at high power density and high flame temperature. J Eng Gas Turb Power 2010;132(12):121503. [7] Weber R, Smart JP, VdKamp W. On the (MILD) combustion of gaseous, liquid, and solid fuels in high temperature preheated air. Proc Combust Inst 2005;30:2623–9. [8] Khalil AEE, Gupta AK. Fuel flexible distributed combustion with swirl for gas turbine applications. J Appl Energy 2013;109:327–34. [9] Khalil AEE, Gupta AK, Bryden MK, Lee SC. Mixture preparation effects on distributed combustion for gas turbine applications. J Energy Resour Technol 2012;134(3):032201. [10] Khalil AEE, Gupta AK. Dual injection distributed combustion for gas turbine application. J Energy Resour Technol 2014;136(1):011601.
488
A.E.E. Khalil, A.K. Gupta / Applied Energy 160 (2015) 477–488
[11] Khalil AEE, Gupta AK. Velocity and turbulence effects on high intensity distributed combustion. J Appl Energy 2014;125:1–9. [12] Khalil AEE, Arghode VK, Gupta AK. Novel mixing for ultra-high thermal intensity distributed combustor. J Appl Energy 2013;105:327–34. [13] Correa SM. A review of NOx formation under gas-turbine combustion conditions. Combust Sci Technol 1992;87:329–62. [14] Tsuji H, Gupta AK, Hasegawa T, Katsuki M, Kishimoto K, Morita M. High temperature air combustion: from energy conservation to pollution reduction. Boca Raton (Florida): CRC Press; 2003. [15] Khalil AEE, Gupta AK. Impact of internal entrainment on high intensity distributed combustion. J Appl Energy; 2015. doi: http://dx.doi.org/10.1016/j. apenergy.2015.07.044 [in press]. [16] Hedman PO, Warren DL. Turbulent velocity and temperature measurements from gas-fueled technology combustor with a practical fuel injector. Combust Flame 1995;100:185–92. [17] Keck O, Meier W, Stricker W, Aigner M. Establishment of a confined swirling natural gas/air flame as a standard flame: temperature and species distributions from laser Raman measurements. Combust Sci Technol 2002;174(8):117–51. [18] Cheng TS, Chao YC, Wu DC, Yuan T, Lu CC, Cheng CK, et al. Effects of fuel-air mixing on flame structure and NOx emissions in swirling methane jet flames. 27th Symposium (int’l) on combustion, vol. 27. The Combustion Institute; 1998. p. 1229–37.
[19] Schmitt P, Poinsot T, Schuermans B, Geigle KP. Large-Eddy simulation and experimental study of heat transfer, nitric oxide emissions and combustion instability in a swirled turbulent high pressure burner. J Fluid Mech 2007;570:17–46. [20] Datta A, Som SK. Combustion and emission characteristics in a gas turbine combustor at different pressures and swirl conditions. Appl Therm Eng 1999;19(9):949–67. [21] Al-Hamamre Z, Diezinger S, Talukdar P, Von Issendorff F, Trimis D. Combustion of low calorific gases from landfills and waste pyrolysis using porous medium burner technology. Proc Saf Environ Protect 2006;84(B4):297–308. [22] Stone R, Clarke A, Beckwith P. Correlations for the laminar-burning velocity of methane/diluent/air mixtures obtained in free-fall experiments. Combust Flame 1998;114:546–55. [23] Galmiche B, Halter F, Foucher F, Dagaut P. Effects of dilution on laminar burning velocity of premixed methane/air flames. Energy Fuels 2011;25:948–54. [24] Kim HS, Arghode VK, Linck MB, Gupta AK. Hydrogen addition effects in a confined swirl-stabilized methane–air flame. Int J Hydrogen Energy 2009;34 (2):1054–62. [25] Kim HS, Arghode VK, Gupta AK. Flame characteristics of hydrogen-enriched methane–air premixed swirling flames. Int J Hydrogen Energy 2009;34 (2):1063–73.
Contents lists available at ScienceDirect
Applied Energy journal homepage: www.elsevier.com/locate/apenergy
Thermal field investigation under distributed combustion conditions Ahmed E.E. Khalil, Ashwani K. Gupta ⇑ Department of Mechanical Engineering, University of Maryland, College Park, MD 20742, USA
h i g h l i g h t s Examined the thermal field behavior under conventional and distributed combustion. N2 and CO2 are used to simulate entrained combustion gases from within the combustor. Reduced oxygen concentration (<15%) is key to achieve thermal field uniformity. Distributed combustion reduced temperature variation by 50% axially & 60% radially. Significant noise and emissions reduction was achieved via distributed combustion.
a r t i c l e
i n f o
Article history: Received 18 July 2015 Received in revised form 7 September 2015 Accepted 11 September 2015
Keywords: Colorless distributed combustion Ultra low NOx High intensity distributed combustion Reactive gas entrainment Thermal field uniformity
a b s t r a c t Distributed combustion has demonstrated significant performance gains, especially on combustion efficiency and near zero pollutants emission. Controlled mixture preparation between air, fuel and internal hot reactive gases prior to mixture ignition is a critical requirement to achieve distributed combustion condition. Though distributed combustion have been extensively studied using a variety of geometries, heat loads and intensities, and fuels, limited information is available on the role of hot reactive gas entrainment and the resultant thermal field uniformity. In this paper, the impact of internal entrainment of hot reactive gases on thermal field uniformity and pollutants emission is investigated. A mixture of nitrogen and carbon dioxide was introduced to the fresh air stream prior to mixing with the fuel and its subsequent combustion to simulate the product gases from within the combustor. Increase in the amounts of nitrogen and carbon dioxide (simulating increased entrainment) significantly reduced pollutants emission, enhanced thermal field uniformity, and increased the reaction volume to occupy larger portion of the combustor. This was evident through spatial temperature measurements in the combustor along with the enhanced distribution of the flame visible signature and OH⁄ chemiluminescence signal. The temperature data demonstrated that lowering oxygen concentration from 21% to 15%, through increased entrainment, promoted distributed combustion conditions with lower overall temperature rise throughout the combustor. In addition, the peak temperature regions associated with swirl burners disappeared, eliminating most of the hot spots in the combustor. The enhanced thermal field uniformity and reduced temperature variation provided ultra-low emissions, demonstrating the impact of enhanced thermal flowfield uniformity on emissions. Experiments performed at different equivalence ratios and entrained gas temperatures demonstrated similar behavior of thermal field uniformity and ultra-low emissions. Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction The increase in fossil fuels as a local energy source to satisfy the states’ energy needs, along with concerns about global warming and climate change, have motivated combustion researchers to develop new combustion methods that has minimal impact on the environment and high efficiency. The potential for natural gas and shale gas deployment in electricity and power generation ⇑ Corresponding author. Tel.: +1 301 405 5276; fax: +1 301 314 9477. E-mail address: [email protected] (A.K. Gupta). http://dx.doi.org/10.1016/j.apenergy.2015.09.058 0306-2619/Ó 2015 Elsevier Ltd. All rights reserved.
have been fostered by their increased availability as a local energy source and their lower carbon emission (compared to other fossil fuels such as coal). To ensure this mandated environmentally friendly performance, future combustion systems shall achieve even lower pollutants (including NOx, CO, unburned hydrocarbons and soot) while minimizing CO2 emissions. To this end, distributed combustion (Colorless Distributed Combustion, CDC [1–4]), among other technologies (i.e., FLOX, MILD) [5–7], has been shown to provide significant benefits of reducing the emissions of NO and CO along with stable combustion, reduced noise, no flame fluctuations and combustion instability. CDC combustors have also
478
A.E.E. Khalil, A.K. Gupta / Applied Energy 160 (2015) 477–488
demonstrated fuel flexibility, handling a variety of liquid and gaseous fuels with no modification to the combustor [8]. The flames in distributed combustion do not show any visible flame signatures so that the flame so formed is termed colorless due to negligible visible emission as compared to conventional flames. The performance benefits of CDC have been demonstrated using a variety of geometrical arrangement for the combustor flowfield, air, and fuel injection schemes [9–12]. The focus in these investigations was on the internal entrainment of hot reactive species and gases and their subsequent mixing with the freshly introduced air and fuel. This entrainment and the subsequent adequate mixing prior to ignition forms pivotal step in achieving distributed reactions. Distributed reactions are characterized by a lower reaction rate over the entire volume of the combustor, as opposed to the concentrated flame front characterized by high reaction rates with local hot spots, to result in the same fuel consumption with lower temperature rise in the combustor. This low reaction rate is achieved through lowering the oxygen concentration in the reactants, and increasing the temperature of the reactants (both achieved simultaneously through the entrainment of hot reactive gases from within the combustor). The distributed combustion regime not only avoids the formation of thin reaction zone but also the hot-spot zones in the flame, which help mitigate thermal NOx formation and emission from the Zeldovich thermal mechanism [13,14]. The impact of the amount of gas entrainment and recirculation on reaction distribution and pollutants emission have been investigated with focus on determining the minimum entrainment requirements for distributed reactions to occur (hot gas recirculation/oxygen concentration) [15]. In these investigations, a swirl burner was used with focus on determining emissions (NO and CO) and flame behavior (radicals’ chemiluminescence signal) with different amounts of recirculation. A mixture of nitrogen and carbon dioxide (90–10% by volume) is used to simulate the product gases. These investigations showed that reaction distribution is significantly enhanced with increased entrainment, lowering oxygen concentration in the ready-to-ignite mixture. This reaction distribution was fostered at oxygen concentration below 15%, to result in ultra-low emissions. However, limited information is available on the thermal field uniformity within the reaction zone. Swirl burners thermal field have been studied in details with focus on mean and fluctuating temperatures. Hedman and Warren [16] have used coherent anti-Stokes Raman spectroscopy (CARS) to measure temperature in a dual swirl combustor, where they showed that the temperature increases across the centerline then decreases near the exit, with significant radial change across the swirl boundary. Similar behavior was also shown by Keck et al. [17]. Cheng et al. [18] used R-type thermocouples showing the same radial temperature change across the swirl stabilized combustor. The temperature profile has also been numerically examined [19,20] with similar behavior as those measured experimentally. In this paper, the thermal field uniformity is evaluated from the temperature measurement throughout the reaction field under distributed combustion conditions. Comparison between the temperature values under distributed reaction conditions and normal combustion will aim to outline the significant thermal field uniformity expected under distributed combustion condition. Various conditions will be examined to identify the cause of this field uniformity and its direct impact. These conditions include variations of stoichiometry, heat load, and different diluents temperature. Nitrogen and carbon dioxide were selected as they represent majority of the product gases. They were mixed in a 90% N2–10% CO2 by volume simulating product gases near stoichiometry conditions. It is recognized that this ratio changes as the equivalence
ratio becomes leaner, the diluting gases mixture (90–10%) was kept constant. Any minor deviation from the actual composition will have minimal impact on the results as nitrogen and carbon dioxide behave similarly in flames. Laminar flame speed and flame temperature for methane-air flames diluted with nitrogen and/or carbon dioxide have shown to exhibit similar behavior [21,22]. Diluting the reactants with a nitrogen–carbon dioxide–water vapor mixture also resulted in similar behavior to that of nitrogen [23]. 2. Experimental facility The experiments were performed using a swirl burner fueled with methane. Details of this swirl burner can be found elsewhere [24]. The performance of this swirl burner has been studied in terms of emissions and velocity profiles using methane and hydrogen enriched methane [24,25]. To simulate product gas entrainment and recirculation, and reduce oxygen concentration in the mixture prior to ignition, different amounts of N2–CO2 mixture were added to the air upstream of the burner. Fuel was injected at the center of the swirler in a non-premixed configuration. Air and nitrogen flowrates were controlled by laminar flow controllers with an accuracy of ±0.8% of reading and ±0.2% of full scale leading to an overall accuracy of 1.5% of the reading. Methane and carbon dioxide flowrates were controlled through gravimetric flow controllers with an accuracy of 1.5% of full scale. Detailed temperature measurements were performed using a K-type thermocouple with NI-DAQ (data acquisition) system calibrated using blackbody calibrator (Omega BB-4A) resulting in an accuracy of ±3%. This accuracy is established based on the maximum difference between the thermocouple reading and the reference thermocouple in the calibrator. The thermocouple was mounted on a traverse mechanism to allow measurement at different axial and radial positions from within the flames. The thermocouple was continuously sampled with the mean of 50 readings reported herein as the average temperature. An ICCD (Intensified Charge-Coupled Device) camera coupled to a narrow band filter for OH⁄ chemiluminescence detection (UV interference filter centered at 307 nm with a FWHM of ±10 nm) was used to evaluate the flame behavior upon the insertion of the thermocouple into the combustor. It is imperative to ensure that the flame characteristic do not change upon the movement of the thermocouple. This is especially critical under distributed combustion conditions where the flame might anchor on the thermocouple. For pollutants emission, the products of combustion were continuously sampled at the exit of the burner. The concentration of NO was measured using a NO–NOx chemiluminescent gas analyzer, CO concentration was measured using the non-dispersive infrared method and O2 concentration (used to correct the NO and CO emissions at standard 15% oxygen concentration) was measured using galvanic cell method. During a single experiment, measurements were repeated at least three times for each configuration and the uncertainty was estimated to be about ±0.5 ppm for NO and ±10% for CO emission. The experiments were repeated at least three times to ensure good repeatability of the experimental data obtained. The experimental rig is shown in Fig. 1 with the flame at two different conditions of normal air combustion wherein the swirl structure is dominant (left photo), and reduced oxygen concentration combustion showing near distributed combustion with less visible emission (right photo). 3. Experimental investigations Table 1 summarizes the conditions for the specific experimental conditions reported here along with the variables manipulated for each case. For each heat load, the fuel flow rate was kept constant
479
A.E.E. Khalil, A.K. Gupta / Applied Energy 160 (2015) 477–488
Fig. 1. Experimental test rig with flame using normal air (left) and reduced oxygen concentration air (right).
Table 1 Experimental parameters.
1 2 3 4 5 6 7 8 9
Heat load (kW)
Heat release intensity (MW/m3 atm)
Mixture temperature (K)
Equivalence ratio
Oxygen concentration (%)
Inlet Reynold’s number
3.25 3.25 3.25 3.25 3.25 4.875 3.25 3.25 3.25
2.4 2.4 2.4 2.4 2.4 3.6 2.4 2.4 2.4
300 300 300 300 300 300 600 600 600
0.9 0.7 0.6 0.9 0.7 0.9 0.9 0.7 0.9
21 21 21 13.8 16.51 21 12.38 14.91 13.8
5000 6000 7000 7000 7500 7000 11,000 10,500 9500
Table 2 Temperature field parameters.
1 2 3 4 5 6 7 8 9
Oxygen concentration (%)
Diluents temp. (K)
Maximum axial temp. (K)
Combustor exit temp. (K)
Temp. difference (max-exit)
21 21 21 13.8 16.51 21 12.38 14.91 13.8
– – – 300 300 – 600 600 600
1248 1191 1144 1104 1093 1350 1144 1106 1164
938 918 897 953 918 1064 1018 995 985
310 273 247 151 175 286 126 111 179
while the air flow rate was changed to change the equivalence ratio. For every equivalence ratio, air and fuel flow rates were kept constant while the amount of N2–CO2 mixture was increased to lower the oxygen concentration in the mixture prior to ignition. The nitrogen-carbon dioxide mixture was heated to 600 K to study the impact of mixture temperature on the combustion behavior and results obtained are given in Table 2.
lence ratios shown herein and agrees favorably well with the trends reported in the literature [16–18]. The high temperature at about 0.6X/R is attributed to the flame stabilization at the tip of the swirler. Decreasing the equivalence ratio resulted in lower temperature values while maintaining the profile shape (see Figs. 2 and 3). The radial temperature profile is shown in Fig. 3 at three different heights, 1.9R, 3.8R and 5R, where R is the radius of the combustor inlet. The profiles were consistent at all the equivalence ratios examined. The profiles, at each height, looked similar but the temperature value decreased with decrease in equivalence ratio similar to the axial profiles behavior. The increase at r = 1.125R for X = 1.9R can be related to the thermocouple being in the thin flame zone, where the maximum temperature was recorded. Radial temperature profile was also obtained at a height of 7.6R, where it was found to be similar to that of 5R. The increase in temperature near the exit (at 28R) is attributed to the smaller exit diameter, increasing the gases velocity and radial mixing with gases away from the central axis location that are expected to be at higher temperatures. 4.2. Dilution impact on thermal field
4. Results and discussion 4.1. Effect of equivalence ratio The initial experiments focused on measuring the temperatures to obtain axial and radial profiles with no entrainment (cases 1, 2, and 3), establishing a baseline for the combustor performance. These measurements are shown in Figs. 2 and 3. In both figures, the abscissa (height, X, and radial position, r, respectively) is normalized by the inlet radius, R. In Fig. 2, one can see that the temperature increases through the combustor up to a distance of 7X/R where the maximum temperature occurs on the centerline. Afterwards, the temperature start to decrease until the exit at 30.8X/R. This behavior is consistent for the three equiva-
Upon examination of the temperature profiles for normal air combustion (no dilution), the temperature measurements with dilution under distributed combustion conditions were determined to reveal the flames signatures. Fig. 4 shows the visible flame signature under different dilution amounts that resulted in lower oxygen concentration in the fresh reactants. Fig. 5 shows the axial temperature profile at an equivalence ratio of 0.9 with no dilution (21% O2) and maximum stable dilution (13.8% O2), representing cases 1 and 4. For no dilution, the difference between the maximum and exit temperature was about 300%, which was significantly reduced to 100% under distributed combustion conditions. In addition, the temperature increase/ decrease ramp is a much smoother under low oxygen concentra-
480
A.E.E. Khalil, A.K. Gupta / Applied Energy 160 (2015) 477–488
1300 Phi=0.9, O2=21%
1250
Phi=0.7, O2=21%
1200
Phi=0.6, O2=21%
Temperature [K]
1150 1100 1050 1000 950 900 850 800
0
5
10
15
20
25
30
35
X/R Fig. 2. Axial temperature profiles along the longitudinal central axis for different equivalence ratio with no dilution.
Phi=0.9, O2=21%
Phi=0.7, O2=21%
Phi=0.6, O2=21%
Temperature [K]
1500
h=1.9R
1400 1300 1200 1100 1000
Temperature [K]
1500
h=3.8R
1400 1300 1200 1100 1000
Temperature [K]
1500
h=5R
1400 1300 1200 1100 1000
0
0.5
1
1.5
2
r/R Fig. 3. Radial temperature profiles for different equivalence ratio with no dilution at different heights.
tion conditions. Under such conditions, the flame is slightly lifted (as seen from a comparison of the extremes in Fig. 4), which led to lower temperatures directly above the swirler in the dilution case up to about a distance of X = 2R. The radial profiles for the two cases discussed herein (cases 1 and 4) are shown in Fig. 6. One can see that there is no increase
in temperature at 1.9R for the lower oxygen concentration case; instead, the temperature is decreased as one progressively moves away from the longitudinal centerline of the combustor, see blue symbols for both cases. The temperature profile at higher heights (3.8R, 5R, and 7.6R) show minimal variation (50 K) for the dilution case as compared to 100 K for case 1 (using 21% oxygen).
481
A.E.E. Khalil, A.K. Gupta / Applied Energy 160 (2015) 477–488
Fig. 4. Flame images for different oxygen concentration (with change in diluents amount) at an equivalence ratio of 0.9.
1300
Phi=0.9, O2=21% Phi=0.9, O2=13.8%
Temperature [K]
1200 1100 1000 900 800 700 600
0
5
10
15
20
25
30
35
X/R Fig. 5. Axial temperature profile at equivalence ratio of 0.9 with and without dilution.
O2=21%, h=1.9R O2=13.8%, h=1.9R
O2=21%, h=3.8R O2=13.8%, h=3.8R
O2=21%, h=5R O2=13.8%, h=5R
O2=21%, h=7.6R O2=13.8%, h=7.6R
1500 1400
Temperature [K]
1300 1200 1100 1000 900 800 700
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
r/R Fig. 6. Radial temperature profile at equivalence ratio of 0.9 with and without dilution at different heights.
In addition, for the profiles at 5R and 7.6R, the temperature is more uniform radially as compared to that at 3.8R. Dilution also resulted in lower temperature values at different heights as compared to normal air combustion. To further understand the behavior at h = 1.9R, OH⁄ chemiluminescence signal was compared for both cases with focus on flame position. Fig. 7 shows this comparison, where it is evident that lowering oxygen concentration via entrainment resulted in a decrease in the chemiluminescence signal as compared to the no dilution
case. The low oxygen concentration case (13.8%) is also shown on the right image but at a different scale (being 0–2500 a.u. instead of 0–10,000 a.u.). It can be seen that at a height of 1.9R, the OH⁄ signal decreases as one moves radially away from the center then increases again. This explains the temperature profile shown in Fig. 6 where the temperature first drops and then increases again. Also OH⁄ shows that this is the height where the reaction starts to occur, with minimal reaction occurring upstream of this location, which agrees well with the low temperature measured at the center
482
A.E.E. Khalil, A.K. Gupta / Applied Energy 160 (2015) 477–488
h=3.8R
h=1.9R
Phi=0.9, O2 =13.8%
Phi=0.9, O2 =13.8%
Phi=0.9, O2 =21% 0
10000
0
2500
Fig. 7. OH⁄ chemiluminescence intensity distribution for standard air combustion (left) and reduced oxygen concentration (center) and with thermocouple position at 1.9R (right).
at position lower than 1.9R for this case, see Fig. 5. This in contrast to the case with no dilution, where at this height, high reaction rate occurs at the swirl lobes resulting in high temperatures. Further downstream at a height of 3.8R, for the low oxygen concentration case, the reaction occurs across the radius of the reactor resulting in a more uniform radial temperature profile. Moreover, Fig. 7 shows that the thermocouple had limited impact on the OH⁄ chemiluminescence as no field disturbance was noticed. Similar axial temperature behavior was also observed at lower equivalence ratio (cases 2 and 5); where the temperature rise across the center was lower for the lower oxygen concentration case, see Fig. 8. The temperature difference between the maximum temperature and the exit temperature was found to be less in the low oxygen concentration case. In addition, the maximum temperature was reduced in the low oxygen concentration case. Radial temperature profiles showed a similar trend to that of Fig. 6 where the temperature profile was more uniform for the low oxygen concentration case. 4.3. Oxygen concentration impact on thermal field The impact of lowering the oxygen concentration over using a lower equivalence ratio was examined and the results on temperature distribution for cases 1, 3, 4, and 6 are shown in Fig. 9. Case 1 represents the base case at an equivalence ratio of 0.9. Case 3 represents dilution through adding more air (lowering equivalence
ratio to 0.6 while maintaining oxygen concentration at 21%). This air addition resulted in an increase in inlet velocity and Reynold’s number (7000). Case 4 represents dilution through lowering oxygen concentration through N2–CO2 addition. The added gas mixtures also increased the velocity and Reynold’s number to match the conditions of case 3. Case 6 represents an increase in air and fuel flow rates to maintain an equivalence ratio of 0.9% and 21% oxygen concentration with an increased velocity and Reynold’s number to match cases 3 and 4. The centerline temperature distribution are shown in Fig. 9. Comparing the different cases, one can see that increase in the flowrates (air and fuel) resulted in the similar temperature profile but of increased values (cases 1 and 6), limiting the impact of increased inlet velocity on its own. Comparing cases 1 and 3, one can see that the temperature profile shape is preserved, but the values are lowered. This is attributed to the leaner conditions (lower equivalence ratio) and the resulting lower temperature for this lower equivalence ratio. The temperature profile only changes upon lowering of the oxygen concentration to result in distributed combustion (case 4). Case 4 has a lower temperature rise (peak) across the combustor as compared to the other cases. The maximum temperature is even lower than that resulting from an equivalence ratio of 0.6. This behavior is supported through OH⁄ chemiluminescence and visible flame signature. Only at lower oxygen concentration did the flame transfer to a distributed regime across the combus-
1200
Phi=0.7, O2=21%
Temperature [K]
1100
Phi=0.7, O2=16.51%
1000 900 800 700 600 500
0
5
10
15
20
25
30
X/R Fig. 8. Axial temperature profile for equivalence ratio of 0.7 with and without dilution.
35
483
A.E.E. Khalil, A.K. Gupta / Applied Energy 160 (2015) 477–488
1400
Case1, Phi=0.9, O2=21% Case 3, Phi=0.6, O2=21%
1300
Case 4, Phi=0.9, O2=13.8% Case 6, Phi=0.9, O2=21%
Temperature [K]
1200 1100 1000 900 800 700 600 500
0
5
10
15
20
25
30
35
X/R Fig. 9. Axial temperature profile for cases 1, 3, 4, and 6.
Case 1, Phi=0.9, O2=21%
Case 3, Phi=0.6, O2=21%
Case 4, Phi=0.9, O2=13.8%
0
Case 6, Phi=0.9, O2=21% 16000
Fig. 10. OH⁄ Chemiluminescence intensity distribution for cases 1, 3, 4, and 6.
tor. For cases 1, 3, and 6, the swirl structure was preserved with hot spots existing at the swirl lobes at the boundaries of the central toroidal recirculation zone. This swirl structure only disappeared through lowering oxygen concentration (case 4) to result in distributed combustion, see Fig. 10. 4.4. Effect of dilution temperature In all the previous experiments, the N2–CO2 mixture was introduced to the fresh reactants at room temperature. These experiments were aimed at identifying the impact of oxygen concentration reduction (which can be achieved in actual combustion systems through product gas recirculation). However, such recirculation will increase the fresh mixture temperature. In the case of internal recirculation, the product gases recirculated will have temperature of about 1500 K or higher. In these experiments, the N2–CO2 mixture was preheated to higher temperatures with view to understand the impact of diluents temperature on thermal field uniformity and emissions. These experiments were performed at the same heat load of 3.25 kW and heat release intensity (HRI) of 2.4 MW/m3 atm. Fig. 11 shows the axial temperature profile for the base case of no dilution (O2 = 21%), dilution mixture at 300 K (O2 = 13.8%), and two dilution cases (O2 = 13.8% and 12.38%) at diluents temperature of 600 K. As the temperature of
the diluents was increased, the combustor was stable at lower oxygen concentrations as opposed to the diluents at room temperature [15]. As the dilution temperature was increased, the overall temperature profile shifted to higher values while maintaining the same profile. This can be seen comparing the two cases with oxygen concentration of 13.8%. The maximum temperature slightly increased from 1100 K to 1160 K. In addition, the reaction zone shifted slightly upstream (as the temperature increase was higher in the case of diluents temperature = 600 K). For the same temperature, increase in the dilution amounts resulted in a lower temperatures throughout the combustor. For all the dilution cases shown here, the temperatures were lower as compared to the case with no dilution, with a smooth temperature increase/decrease throughout the flow regime. For the case of no dilution and diluents at 300 K, the combustor exit temperature was almost the same (945 K), when the diluents temperature increased, the combustor exit temperature increased, respectively. As more diluents were added (further reduction in oxygen), the combustor exit temperature was found to be higher (985 K for 13.8%, 1018 K for 12.38%). This increase in temperature can be related to the increased energy input for that case (higher amounts of N2–CO2 mixture at 600 K to further reduce O2 concentration).
484
A.E.E. Khalil, A.K. Gupta / Applied Energy 160 (2015) 477–488
1300
O2=21% O2=13.8%, T=600K O2=12.38%, T=600K O2=13.8%
1200
Temperature [K]
1100 1000 900 800 700 600
0
5
10
15
20
25
30
35
X/R Fig. 11. Axial temperature profile for different dilution cases at equivalence ratio of 0.9 (cases 1, 4, 7, and 9).
O2=21%, h=1.9R O2=13.8%, T=600K, h=1.9R O2=12.38%, T=600K, h=1.9R
O2=21%, h=3.8R O2=13.8, T=600K, h=3.8R O2=12.38%, T=600K, h=3.8R
O2=21%, h=5R O2=13.8%, T=600K, h=5R O2=12.38%, T=600K, h=5R
1500 1400
Temperature [K]
1300 1200 1100 1000 900 800 700
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
r/R Fig. 12. Radial temperature profile for different dilution cases at equivalence ratio of 0.9 and diluents temperature of 600 K (cases 1, 7, and 9).
Fig. 12 shows the radial temperature profile for the reduced oxygen concentration with diluents at 600 K. The temperature profiles are similar to those obtained with the diluents at room temperature (see Fig. 6). The radial temperature values slightly decreased with increased dilution (lowered oxygen concentration) for the different heights shown here. The profile at h = 1.9R is consistent with what was shown in Fig. 6 and explained by the reaction zone location (see Fig. 7). The diluents temperature influence on the thermal field uniformity was also examined. The radial temperature profiles were compared for oxygen concentration of 13.8% using diluents temperature of 300 K and 600 K. The results obtained are shown in Fig. 13. Increasing the temperature of the diluents resulted in an increase in the temperature values at the examined heights while maintaining a similarity in the temperature profile shape. More importantly, the radial temperature profile uniformity was maintained at h = 3.8R, 5R and 7.6R as a result of the distributed combustion conditions achieved at this oxygen concentration (13.8%) and dilution temperatures. Experiments were also performed at an equivalence ratio of 0.7. Increase in the diluents temperature allowed stable operation at lower oxygen concentration. Fig. 14 shows the axial temperature
profile for equivalence ratio of 0.7 with no dilution, O2 = 16.51% (T = 300 K) and O2 = 14.91% (T = 600 K). Similar to the profiles for equivalence ratio of 0.9, increasing the diluents temperature slightly increased the temperature values while maintaining the uniform temperature profile. Also increased diluents temperature resulted in an increase in the combustor exit temperature. 4.5. Thermal field uniformity The temperature field for each case was analyzed to obtain the maximum temperature along the longitudinal axis and the difference between that maximum and exit temperature to provide information on the temperature field uniformity. As the oxygen concentration was reduced, to steer towards distributed reaction conditions, the maximum temperature was significantly reduced with a smaller temperature difference. At equivalence ratio = 0.9, the maximum temperature was reduced from 1248 K (O2 = 21%) to 1104 K (13.8%) with the temperature difference reduced from 310 to 151 K. Further oxygen reduction (assisted by increased diluents temperature) led to a slight increase in the maximum temperature (1144 K) but a reduced temperature difference (126 K). The same behavior was also
485
A.E.E. Khalil, A.K. Gupta / Applied Energy 160 (2015) 477–488
T=300K, h=1.9R
T=300K, h=3.8R
T=300K, h=5R
T=300K, h=7.6R
T=600K, h=1.9R
T=600K, h=3.8R
T=600K, h=5R
T=600K, H=7.6R
1300
Temperature [K]
1200 1100 1000 900 800 700
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
r/R Fig. 13. Radial temperature profile for different dilution temperatures at equivalence ratio of 0.9 and oxygen concentration of 13.8% (cases 4 and 7).
1300
O2=21% O2=16.51%, T=300K O2=14.91%, T=600K
Temperature [K]
1200 1100 1000 900 800 700 600
0
5
10
15
20
25
30
35
X/R Fig. 14. Axial temperature profile for different dilution temperatures at equivalence ratio of 0.7 (cases 2, 5 and 8).
demonstrated at an equivalence ratio of 0.7. For all the cases, lowering oxygen concentration in the fresh reactants led to 40–60% reduction in temperature variation. The same analysis was performed on different radial profiles. Table 3 gives the maximum temperature and minimum temperature along with their difference for the profiles with maximum variations. Similar to longitudinal profiles, lowering oxygen concentration resulted in lower temperature variation (60 K instead of 262 K in case of phi = 0.9). Increase in diluents temperature slightly affected the radial profile variation. For all cases, low oxygen concentration resulted in 60% or more reduction in temperature variation. The thermal field uniformity can be expressed in percentage temperature variation as:
DT ¼ ½ðT max T exit Þ=ðT exit T in Þ 100
ð1Þ
where DT is the percentage temperature change, Tmax is the maximum temperature along measured line, Texit is the combustor exit temperature and Tin is the air inlet temperature, 300 K. Following the above definition in Eq. (1), the percent temperature change for standard air combustion, cases 1, 2, 3 and 6, was 48.6%, 44.2%, 41.4% and 37.5%, respectively. On the other hand,
for the cases demonstrating distribution combustion, this percentage varies between 23.1% and 28.3% for the diluents at 300 K (cased 4 and 5). For the diluents at 600 K, this percentage varied around 17% (cases 7 and 8) and 26% (case 9). Thus lowering the oxygen concentration enhanced thermal field uniformity and reduced the percentage change from 48.6% (case 1) down to 23.1% (case 4). Upon increasing the diluents temperature, this value decreased to 17% (case 7). The same trend can be seen for the lower equivalence ratio of 0.7 (case 2: 44.2%, case 5: 28.3%, and case 8: 26%) Similar behavior was observed for radial temperature distribution. For normal air combustion, the percentage temperature variation varied between 15% and 23%, cases 1–3, while this value was about 7% for distributed combustion conditions, cases 4, 5, 7 and 9. These percentages were obtained using the following equations:
DT ¼ ½ðT max T min Þ=ðT max T in Þ 100
ð2Þ
where Tmin is the minimum temperature along the measured line. The temperature values used for calculations are those of the height presenting maximum temperature variation, see Table 3. These values emphasize the thermal field uniformity achieved under distributed combustion conditions. This uniformity can be
486
A.E.E. Khalil, A.K. Gupta / Applied Energy 160 (2015) 477–488
Table 3 Radial temperature field parameters. Oxygen concentration (%) 1 2 3 4 5 7 9 a b
Diluents temp. (K)
21 21 21 13.8 16.51 12.38 13.8
Height a
– – – 300 300 600 600
1.9R 1.9Ra 1.9Ra 7.6Ra,b 7.6Ra,b 7.6Ra,b 7.6Ra,b
Maximum radial temp. (K)
Minimum radial temp. (K)
Temp. difference (max–min)
1449 1395 1247 1156 1126 1204 1204
1187 1139 1100 1096 1080 1130 1146
262 256 147 60 46 74 58
This is the height with maximum fluctuations, temperature profiles at other heights demonstrated lower temperature variation. The profile at 1.9R was excluded as this was before the reaction initiation as indicated through chemiluminescence data.
further enhanced by increasing the temperature of the diluents and decreasing the oxygen concentration as evident from the trend shown in comparing cases 1 (no dilution), 4 (13.8% O2, T = 300 K), and 7 (12.38%O2, T = 600 K), as these values were: axially: 48.6%, 23.1%, 17.5%, and radially: 22.8%, 7%, 7%, respectively.
4.6. Pollutants emission The benefits of distributed combustion are not only limited to the thermal field uniformity (which is a direct result of the lower reaction rate distributed over a larger volume of the combustion), but also on the combustor performance in terms of reduced emissions. These benefits have been previously examined for different combustors with focus on stability and pollutants emission under
different geometries [1–4]. In this work, the emissions for the configurations examined herein are discussed in order to link thermal field uniformity within the combustor with emissions.
4.6.1. Noise emission One of the characteristic benefits of distributed combustion is the reduced noise emission. During experiments, the combustor noise was significantly reduced when the combustor transitioned to distributed combustion mode (through added dilution, lowering oxygen concentration). For an equivalence ratio of 0.9, the combustor noise recorded was 80.3 dB (A-weighted) operating at no dilution (21% O2, case 1). This noise was reduced to 63.5 dB upon transition to distributed combustion mode (13.8% O2, no preheat to the diluents, case 4) leading to a significant noise reduction.
0.006
21% O2
0.005
Signal
0.004 0.003 0.002 0.001 0
0
1000
2000
3000
4000
5000
6000
7000
8000
Frequency [Hz] 0.006
13.8% O2 0.005
Signal
0.004 0.003 0.002 0.001 0
0
1000
2000
3000
4000
5000
6000
7000
8000
Frequency [Hz] Fig. 15. Noise spectrum from the combustor operating under normal conditions and with decreased oxygen concentration (distributed combustion condition).
A.E.E. Khalil, A.K. Gupta / Applied Energy 160 (2015) 477–488 Table 4 NO and CO emissions (corrected to 15% O2 in exhaust gas) for the different cases.
1 2 3 4 5 6 7 8 9
Oxygen concentration (%)
Diluents temp. (K)
NO (ppm)
CO (ppm)
21 21 21 13.8 16.51 21 12.38 14.91 13.8
– – – 300 300 – 600 600 600
13.3 6.6 2.26 1.8 0.8 13.3 1.9 0.75 2.66
24.8 3.11 3.7 5.3 2.87 40 4.5 2.87 3.93
The flow noise (reactants flowing without reaction) is about 54 dB. This indicates that distributed combustion resulted in approximately 17 dB decreased noise emission than normal combustion case and only 9.5 dB increase over the non-reacting condition. The spectral noise signature was also captured for the combustor operating under normal combustion (case 1) and distributed combustion conditions (case 4) and is shown in Fig. 15. The background noise was subtracted from both signals. Comparing both signals, one can see the disappearance of the combustor dominant frequency from the swirl (at around 500 Hz) when the combustor operated at distributed combustion mode to result in the low overall noise as demonstrated by some 17 dB reduced noise emission. 4.6.2. NO and CO emission The thermal field uniformity reflected on the pollutants emission also. NO and CO emissions for the cases discussed herein are shown in Table 4 where the distributed combustion cases (low oxygen concentration cases) offered the lowest NO emissions without any increase in CO emission or any associated instabilities. These emissions were measured at the exit (at X = 30.8R). Under distributed combustion conditions, NO emissions were 2.7 ppm or lower for the cases at equivalence ratio of 0.9 (for diluents temperature of 300 K and 600 K, cases 4, 7 and 9). This is in contrast to 13.3 ppm NO recorded from normal flame case (no dilution, case 1). CO emission was also lowered with distributed combustion (5.3 ppm or less as opposed to 24.8 ppm of CO for normal combustion case). The same behavior was also demonstrated at equivalence ratio of 0.7, where distributed combustion (cases 5 and 8) emitted less than 1 ppm of NO while normal flame (case 2) emitted 6.6 ppm of NO. CO emissions were about the same. These pollutants emission outlines the significant role of distributed combustion on pollutants emissions reduction.
487
to the concentration thin reaction zone at the shear layer of the swirl flow (at the boundaries of the central toroidal recirculation zone). Visible flame signature demonstrated the same behavior as lowering the oxygen concentration through increased dilution resulted in faint blue distributed flame as opposed to that obtained from normal swirl flame. The results revealed that the reaction distribution is a result of lower oxygen concentration rather than the dilution itself. Lowering the equivalence ratio diluted the flame (due to the added nitrogen) but did not result in distributed reactions (as the oxygen concentration remained at 21%). On the other hand, lowering oxygen concentration, while maintaining the same flow conditions (as that of the low equivalence ratio) resulted in a distributed reaction along with the disappearance of the swirl structure to outline the importance of oxygen concentration as a means to achieve distributed combustion. Increase in the diluents temperature slightly increased the temperature values across the combustor while maintaining the smooth temperature change/ramp within. Increase in the diluents temperature led to stabilizing the flame at lower oxygen concentrations to result in improved thermal field uniformity at these lower concentrations as compared to that with diluents at room temperature. Achieving distributed combustion, evidenced by distributed visible emissions and thermal flow field, resulted in significant reduction of harmful pollutants emission. This includes noise as well as chemical pollutants. The combustor noise was reduced by 17 dB when the combustor transition from swirl flame to distributed combustion mode. In addition, the characteristic frequency of the noise associated with the swirl burner disappeared upon transition to distributed combustion. Distributed combustion also lowered NO and CO emissions as compared to normal swirl flames. NO emissions were reduced by some 80% as the combustor transitioned to distributed combustion with similar reduction in CO at an equivalence ratio of 0.9. Similar reductions were also achieved at an equivalence ratio of 0.7, thus revealing the impact of distributed combustion and its thermal field uniformity on pollutants emission and combustor performance. Acknowledgment This research was supported by ONR and is gratefully acknowledged. References
5. Conclusions Temperature profiles measured along the centerline of a swirl burner and across the combustor radius at defined heights has demonstrated the thermal field uniformity achieved under distributed combustion conditions. Temperature variation along the centerline and across the combustor radius was significantly reduced upon lowering the oxygen concentration of the reactants. This is achieved through using N2–CO2 mixture simulating internal hot reactive combustion gases recirculation and entrainment from within the combustor. Temperature measurements across the combustor axis revealed that distributed combustion lowered the maximum temperature and resulted in approximately 50% reduction in temperature variation across the combustor. Similar temperature variation reduction was demonstrated at different radial positions examined. The thermal field uniformity data supported the OH⁄ chemiluminescence data which also showed distributed reaction to occur across the combustor at low oxygen concentrations as opposed
[1] Arghode VK, Gupta AK. Effect of flow field for colorless distributed combustion (CDC) for gas turbine combustion. J Appl Energy 2010;87:1631–40. [2] Arghode VK, Gupta AK. Role of thermal intensity on operational characteristics of ultra-low emission colorless distributed combustion. J Appl Energy 2013;111:930–56. [3] Khalil AEE, Gupta AK. Swirling distributed combustion for clean energy conversion in gas turbine applications. J Appl Energy 2011;88:3685–93. [4] Khalil AEE, Gupta AK. Distributed swirl combustion for gas turbine application. J Appl Energy 2011;88:4898–907. [5] Wunning JA, Wunning JG. Flameless oxidation to reduce thermal NO formation. Prog Energy Combust Sci 1997;23:81–94. [6] Lammerl O, Schutz H, Schmitz G, Luckerath R, Stohr M, Noll B, et al. FLOX combustion at high power density and high flame temperature. J Eng Gas Turb Power 2010;132(12):121503. [7] Weber R, Smart JP, VdKamp W. On the (MILD) combustion of gaseous, liquid, and solid fuels in high temperature preheated air. Proc Combust Inst 2005;30:2623–9. [8] Khalil AEE, Gupta AK. Fuel flexible distributed combustion with swirl for gas turbine applications. J Appl Energy 2013;109:327–34. [9] Khalil AEE, Gupta AK, Bryden MK, Lee SC. Mixture preparation effects on distributed combustion for gas turbine applications. J Energy Resour Technol 2012;134(3):032201. [10] Khalil AEE, Gupta AK. Dual injection distributed combustion for gas turbine application. J Energy Resour Technol 2014;136(1):011601.
488
A.E.E. Khalil, A.K. Gupta / Applied Energy 160 (2015) 477–488
[11] Khalil AEE, Gupta AK. Velocity and turbulence effects on high intensity distributed combustion. J Appl Energy 2014;125:1–9. [12] Khalil AEE, Arghode VK, Gupta AK. Novel mixing for ultra-high thermal intensity distributed combustor. J Appl Energy 2013;105:327–34. [13] Correa SM. A review of NOx formation under gas-turbine combustion conditions. Combust Sci Technol 1992;87:329–62. [14] Tsuji H, Gupta AK, Hasegawa T, Katsuki M, Kishimoto K, Morita M. High temperature air combustion: from energy conservation to pollution reduction. Boca Raton (Florida): CRC Press; 2003. [15] Khalil AEE, Gupta AK. Impact of internal entrainment on high intensity distributed combustion. J Appl Energy; 2015. doi: http://dx.doi.org/10.1016/j. apenergy.2015.07.044 [in press]. [16] Hedman PO, Warren DL. Turbulent velocity and temperature measurements from gas-fueled technology combustor with a practical fuel injector. Combust Flame 1995;100:185–92. [17] Keck O, Meier W, Stricker W, Aigner M. Establishment of a confined swirling natural gas/air flame as a standard flame: temperature and species distributions from laser Raman measurements. Combust Sci Technol 2002;174(8):117–51. [18] Cheng TS, Chao YC, Wu DC, Yuan T, Lu CC, Cheng CK, et al. Effects of fuel-air mixing on flame structure and NOx emissions in swirling methane jet flames. 27th Symposium (int’l) on combustion, vol. 27. The Combustion Institute; 1998. p. 1229–37.
[19] Schmitt P, Poinsot T, Schuermans B, Geigle KP. Large-Eddy simulation and experimental study of heat transfer, nitric oxide emissions and combustion instability in a swirled turbulent high pressure burner. J Fluid Mech 2007;570:17–46. [20] Datta A, Som SK. Combustion and emission characteristics in a gas turbine combustor at different pressures and swirl conditions. Appl Therm Eng 1999;19(9):949–67. [21] Al-Hamamre Z, Diezinger S, Talukdar P, Von Issendorff F, Trimis D. Combustion of low calorific gases from landfills and waste pyrolysis using porous medium burner technology. Proc Saf Environ Protect 2006;84(B4):297–308. [22] Stone R, Clarke A, Beckwith P. Correlations for the laminar-burning velocity of methane/diluent/air mixtures obtained in free-fall experiments. Combust Flame 1998;114:546–55. [23] Galmiche B, Halter F, Foucher F, Dagaut P. Effects of dilution on laminar burning velocity of premixed methane/air flames. Energy Fuels 2011;25:948–54. [24] Kim HS, Arghode VK, Linck MB, Gupta AK. Hydrogen addition effects in a confined swirl-stabilized methane–air flame. Int J Hydrogen Energy 2009;34 (2):1054–62. [25] Kim HS, Arghode VK, Gupta AK. Flame characteristics of hydrogen-enriched methane–air premixed swirling flames. Int J Hydrogen Energy 2009;34 (2):1063–73.