Effects of Burned Gas Recirculation on NOx Emissions from Natural

Article pubs.acs.org/EF

Effects of Burned Gas Recirculation on NOx Emissions from Natural Gas−Hydrogen−Oxygen Flames in a Burner with Separated Jets Sébastien Yon,*,† Jean-Charles Sautet,† and Toufik Boushaki‡ †

CORIA UMR 6614, Centre National de la Recherche Scientifique (CNRS), Université et INSA de Rouen, 76801 Saint Etienne du Rouvray, France ‡ ICARE, Centre National de la Recherche Scientifique (CNRS), Université d’Orleans, IUT GTE, 45070 Orleans, France ABSTRACT: Strict regulations of NOx emission standards have resulted in an optimized performance of combustion chambers. The present paper describes the effects of burned gas recirculation on NOx emissions from hythane−oxygen flames in a separated jet burner. Burners with separated fuel and oxidizer jets permit a high dilution of reactants and a large recirculation zone of burned gases in the combustion chamber, which favors the decrease of nitrogen oxide emissions. In this study, the oxyfuel burner was equipped with two nozzles: the first nozzle supplied a hythane flow (a mixture of natural gas and hydrogen), and the second nozzle supplied pure oxygen. The hydrogen content in the fuel varied from 0 to 20% in volume. The influence of the distance between the nozzles (12−100 mm) and the global equivalence ratio of the mixture was analyzed. Measurements of combustion product concentrations were carried out at the combustion chamber exit using a water-cooled probe and a NOx analyzer. The velocity fields and the size of the recirculation zone were determined by the particle image velocimetry (PIV) technique in the reacting flow. The results showed that an increase in the size of the recirculation zone leads to a decrease in NOx emissions (up to 96%) and that NOx emissions are inversely proportional to the measured surface area of the burned gas recirculation zone. Because of the reaction between recirculated CO2 and thermal NO and its chemical effects, NOx emissions decrease in lean combustion because of the destruction of NO.

1. INTRODUCTION The reduction of nitrogen oxide emissions has been extensively investigated in combustion research in response to increasingly stringent NOx emission standards. To reduce NOx emissions, various theoretical and technological directions have been explored, such as the effect of the flame temperature,1−5 the influence of the flame strain,6−8 or the flame radiation.9−13 Many studies dealing with NOx emissions have focused on diffusion flames;14−18 this kind of flame can be of great interest in the reduction of NOx emissions if the combustion is coupled with the recirculation of burned gases in the combustion chamber. Recent attempts using this technique have shown the efficiency of the recirculation of combustion products to reduce NOx emissions.19,20 With this technique, the burned gases are recirculated inside the combustion chamber and introduced into the flame, which permits a radical reduction of NOx emissions with compact equipment and at low cost.21 However, NOx emissions from a burner with separated jets and the effects of the size of the recirculation zone of burned gases in such a burner have not yet been studied. A new generation of burners with separated fuel (in the present study, natural gas) and oxidizer injectors shows great potential for the reduction of nitrogen oxide emissions. Indeed, the nozzle separation permits the development of a nonreacting zone between the two jets with regard to which the combustion products are recirculated and introduced into the flame. This dilution effect increases when the distance between the nozzles grows and favors the drive of burned gases in the recirculation zone.22,23 Indeed, the distance between the nozzles influences the flow dynamics acting on the location of the mixing point (the region where the jets start to interact © 2012 American Chemical Society

and mix together) because the further the nozzles are moved away, the more the jets interact downstream in the flow, favoring the dilution of the burned gases.24 Studies in free turbulent jets have shown that the space between the nozzles increases the size and force of the recirculation zone.25 However, the high dilution of reactants by burned gases is limited by the flame instability. The separation of the two nozzles increases the fluctuations of the bottom flame, leading in some case to extinction for high distance between the nozzles. To solve this problem of flame instability, hydrogen is added to the natural gas. A mixture of natural gas and hydrogen, called hythane, has the advantage of considerably modifying the properties of the fuel, thereby preserving the distribution installation and the burner geometry. Because of the properties of hydrogen found in the combustion process, specifically the high molecular diffusivity, wide flammability limits, high flame speed, and low ignition energy,26,27 hydrogen in the fuel allows for combustion systems to operate with leanfuel mixtures. An increase of flammability limits in the presence of hydrogen offsets the harmful effects of lean combustion or flame instabilities, such as local extinctions, energy losses by radiation, and flame stretching.28−30 Previous studies, which focused on flame lift-off heights and mixing zones for a natural gas−H2−O2 flame from burners with two jets, have shown that the hydrogen addition reduces the size of the recirculation zone and improves the ability of the combustion products to dilute Received: February 27, 2012 Revised: July 7, 2012 Published: July 12, 2012 4703

dx.doi.org/10.1021/ef300338m | Energy Fuels 2012, 26, 4703−4711

Energy & Fuels

Article

the flame without modifying thermal power, the global equivalence ratio, or the distance between the nozzles.31 To improve the flame stability and to limit NOx emissions in burners with separated jets, the air is substituted with pure oxygen, which is referred to as oxy-combustion. In air combustion, nitrogen results in a low combustion yield and high energy consumption because nitrogen contained in the air acts as energy ballast. The complete substitution of air with oxygen leads to an improved heat yield, a rise in the adiabatic flame temperature (2200 K for CH4 air combustion and 3090 K in CH4 oxy-combustion),32 a fuel consumption reduced by 50%, and from an environmental point of view, a decrease in the nitrogen oxide formation (of up to 95%) because of the reduced nitrogen quantities in the oxidant.33 However, a rising flame temperature favors nitrogen oxide formation via the NO thermal mechanism. In an oxy-fuel burner, NOx production is a result of the effects of air infiltration, the presence of N2 in the fuel, flame radiation, and aerodynamic straining.34 The findings on oxy-flames from burners with separated nozzles can be transposed to burner systems functioning with air because the recirculation rates are similar. In industrial processes, air used as an oxidant is often enriched in oxygen (from 30 to 40%), and tendencies found in a pure oxygen setup are therefore completely transposable. In this kind of burner with two separated jets, the properties in combustion of hythane and pure oxygen permit the improvement of the flame stability in lean combustion to fall the fuel consumption. Thus, the influence of the global equivalence ratio on NOx emissions is an important parameter to study in lean combustion and, more precisely, the impact of the size and composition of the recirculation zone of burned gases on NOx emissions. The present study was conducted with a diffusion flame from a separated jet burner to gain an understanding of the influence of the recirculation of burned gases on NOx emissions in terms of size and composition. Three parameters were used to modify the size of the burned gas recirculation zone: the distance between hythane and oxygen jets, the hydrogen volume fraction in natural gas, and the global equivalence ratio. NOx emissions were studied according to these parameters. The advantage of the burned gas recirculation zone is its size, which has the ability to radically modify NOx emissions without changing the characteristics of the combustion chamber. The experimental setup consisted of a burner functioning with hythane and pure oxygen, situated in the bottom wall of the combustion chamber. A water-cooled probe was used to carry outgas sampling, and NOx (NO + NO2) was detected with a gas analyzer using the chemiluminescence technique. A study of the jet aerodynamics through particle image velocimetry (PIV) allowed us to characterize the size of the recirculation zone and to understand the influence of combustion product recirculation on NOx emissions. In the first part of our work, we studied NOx emissions in relation to the hydrogen volume fraction of the fuel with a fixed equivalence ratio corresponding to the stoichiometric condition. In the second part of our study, we investigated NOx emissions as a function of the equivalence ratio without the addition of hydrogen.

the distance between the nozzles (from D = 12 to 100 mm), the hydrogen content in the fuel (from αH2 = 0 to 20%), and the global equivalence ratio (from Φ = 1 to 0.7). The burner depicted in Figure 1 consisted of two non-ventilated jets: a hythane jet (natural gas and hydrogen) and a pure oxygen jet.

Figure 1. Diagram of the burner with two separated jets. The separation distance between the nozzles (D) varied from 12 to 100 mm. The internal diameter of the nozzles was d = 6 mm. The natural gas had a density of 0.83 kg m−3 and a volume composition of 85% CH4, 9% C2H6, 3% C3H8, 2% N2, 1% CO2, and traces of higher hydrocarbon species. The hydrogen volume fraction in the fuel, αH2 = ρH2ṁ H2/(ρH2ṁ H2 + ρNGṁ NG) varied between 0 and 20% (with ṁ and ρ denoting the mass flow rate and the density and subscripts H2 and NG representing hydrogen and natural gas, respectively). The oxygen had a purity of 99.5% and a density of 1.354 kg m−3 (at 1 atm and 15 °C). The fuel flow rate (ṁ fuel) and the exit velocity of hythane (U0fuel) depended upon the hydrogen volume fraction in the fuel blend (ṁ fuel = ṁ NG + ṁ H2) and the equivalence ratio. For the pure natural gas configuration (αH2 = 0%) in stoichiometric proportion, thermal power P = 25 kW, ṁ fuel = 0.55 g s−1, and U0fuel = 23.7 m s−1. The flow rate and exit velocity of oxygen were fixed regardless of the configuration and corresponded to the value calculated for a thermal power of 25 kW in stoichiometric proportions (Φ = 1); thus, ṁ O2 = 1.954 g s−1 and U0O2 = 51.3 m s−1. Table 1 summarizes the parameters of this experimental study, including natural gas and hydrogen flow rates (ṁ NG and ṁ H2), the fuel exit velocity, Reynolds number, and the initial velocity ratio, r, between the fuel (natural gas + hydrogen) and oxygen. The regulation of the natural gas flow was provided by a mass flow controller (TYLAN RDM 280) with an accuracy of ±0.2%, while the oxygen and hydrogen flow rates were controlled by sonic throats connected to pressure gauges with an accuracy of ±0.7%. Flames from the oxy-burner occurred in a combustion chamber with a square cross-section of 60 × 60 cm2 and a height of 100 cm, as shown in Figure 2. To limit wall contact and to keep a constant wall temperature, the lateral walls were refractory-lined inside (12 mm thick lining with a thermal conductivity of 0.1 W m−1 K−1) and watercooled outside of the combustion chamber. The exhaust section was closed by a convergent with a small square cross-section opening of 12 × 12 cm2, allowing for the exhaust of flue gases. Six windows on every lateral wall created the optical accesses necessary to the PIV setup. The combustion chamber reached the equilibrium temperature after 25 min, and the combustion chamber reached the chemical equilibrium after 40 min. The chemical equilibrium, corresponding to the moment at which air contained in the chamber is totally consumed, ensured the sole presence of burned gases inside the chamber. 2.2. Measurements of NOx Emissions. The combustion gases were sampled at the exit of the combustion chamber, i.e., at the height Z = 1240 mm. Analyses were performed using a NOx analyzer

2. EXPERIMENTAL SECTION 2.1. Burner, Flow Control System, and Combustion Chamber. The aim of this study was to investigate the effects of the combustion product diluted flame on NOx emissions considering 4704


Energy & Fuels

Article

Table 1. Parameters of the Studya hythane jet

a

−1

−1

Φ

P (kW)

αH2 (%)

ṁ NG (g s )

ṁ H2 (g s )

1 1 1 1 1 0.9 0.8 0.7

25 25 25 25 25 22.5 20 17.5

0 5 10 15 20 0 0 0

0.5556 0.5473 0.5384 0.5288 0.5184 0.5000 0.4444 0.3889

0 0.0031 0.0065 0.0101 0.0140 0 0 0

r

oxygen jet −1

U0fuel

(m s ) 23.7 24.5 25.5 26.5 27.6 21.3 18.9 16.6

−1

Re

ṁ O2 (g s )

10761 10632 10498 10358 10212 9684 8608 7532

1.954 1.954 1.954 1.954 1.954 1.954 1.954 1.954

U0O2

−1

(m s ) 51.3 51.3 51.3 51.3 51.3 51.3 51.3 51.3

Re

U0fuel/U0O2

21823 21823 21823 21823 21823 21823 21823 21823

0.46 0.48 0.50 0.52 0.54 0.42 0.37 0.32

Inside diameter of the nozzles d = 6 mm. lens with a focal length of +85 mm (1:14.1 Nikkor, Nikon) was placed perpendicular to the light source and collected the signal of Mie scattering emitted by the seeded particles [zirconium oxide (ZrO2)] in the reacting flow. An interference filter (532 nm, 3 ± 0.6 nm bandwidth, and 35% peak transmittance minimum) was used to reject the bright luminosity from the oxy-flame in front of the lens of the PIV setup. To determine the dynamic fields of the flow, calculations of the cross-correlation images were carried out by the Davis software to find the average particle displacement in each sub-region of the image. The interrogation window dimensions were 64 × 64 pixels2, with a coverage of 50%, and, hence, a grid step of 32 pixels. Post-processing was carried out to detect and correct the aberrant vectors that appeared in the cross-correlation calculations. The sub-pixel displacement was estimated by means of Gaussian peak fitting. A maximum displacement of eight pixels would correspond to less than 2% uncertainty in the final velocity measurement. To characterize the flow and to improve spatial resolution, it was necessary to take measurements of four different heights: 0−93.5, 58−172, 136−250, and 212−328 mm, with a spatial resolution of the interrogation windows of 2.93 × 2.93, 3.57 × 3.57, 3.57 × 3.57, and 3.64 × 3.64 mm2, respectively. This allowed us to study the different pertinent zones of the flow to describe the recirculation zone, the fusion zone (where the two jets begin to mix and interact), and the combination zone (where the flow tends to have similar behavior to a single jet). The processing of 500 image pairs made it possible to obtain 500 instantaneous velocity vector fields for each height and configuration. The mean velocity field was obtained by averaging the 500 instantaneous fields. Figure 3 shows the mean 2D velocity field and the characteristic zones of a flow with the two separated jets for the stoichiometric configuration, with αH2 = 20% and D = 60 mm. This study focused on the recirculation zone of combustion products located between the two jets (see the red square in Figure 3).

Figure 2. Schematic of the combustion chamber. (HORIBA PG-250) connected to a water-cooled probe (1 mm in diameter) with a heated sample line transfer. The process used to measure nitrogen oxide was based on the chemiluminescence method. In this situation, a portion of nitrogen monoxide (NO) contained in the sample gas reacted with ozone (O3), leading to the production of NO2. Part of this generated NO2 was in its excited state (NO2*) and returned to the normal state by emitting a photon. The NO concentration was derived by measuring the emitted photon flux. NO2 was converted to NO with the molybdenum−carbon catalyst. The measurement of converted NO yielded the NO2 concentration. The typical temporal duration of these NOx measurements was 300 s with a sampling rate of 1 measurement per second. The NOx emissions average was determined on the basis of 300 measurements per configuration. The measurements of the NOx analyzer ranged from 0 to 1000 ppm, with a repeatability of ±0.5% of full scale. For a nozzle distance of D = 12 mm, the accuracy of NOx emissions was ±0.5%, and for a nozzle distance of D = 40 mm, the accuracy of NOx rates was ±5%. Despite the accuracy, the NOx measurement results indicated significant tendencies concerning NOx emissions. 2.3. Velocity Measurements. PIV was used to study the flow aerodynamics. This non-intrusive method allowed us to obtain twodimensional (2D) images of the flow and to deduce the instantaneous 2D velocity fields. A double-pulsed Nd:YAG laser (Big Sky CFR200, Quantel) with a wavelength of 532 nm and a frequency of 10 Hz was used as a light source (120 mJ/pulse and a pulse duration of 8 ns). An optical system consisting of three consecutive lenses created the laser sheet with a thickness of 500 μm and a height of 80 mm. A charge-coupled device (CCD) camera with a dynamic range of 16 bits (2040 × 2040 pixel2, Image Pro X Lavision) was oriented perpendicular to the laser sheet. A

3. RESULTS AND DISCUSSION 3.1. Determination of the Size of the Recirculation Zone. The size of the recirculation zone was studied in the XZ plane based on the PIV in the reacting flow. This zone was defined by the distance between the burner exit (Z = 0 mm) and a characteristic height, called the mixing point, Zmp (Figure 4). The Zmp point was located at the height where the velocity between the two jets ceased to be negative, which was equivalent to the completed recirculation of the burned gases. After the mixing point Zmp, the two jets started to mix and interact together; this corresponded to the beginning of the fusion zone of the jets. The shape of the recirculation zone resembled a trapezium, where the base of the trapezium was formed by the distance between the nozzles, D, minus the internal diameter of the nozzles. Zmp was the height of the trapezium, and the second base, B, of the trapezium was the distance between the jets at 4705


Energy & Fuels

Article

the Zmp height, as shown in Figure 4. The motions observed in the YZ plane had no influence on the flow behavior. The PIV measurements in the YZ plane indicated that the Uz mean velocity was near 0, except in the jet plane (Y = 0). In the jet plane, the Uz mean velocity followed the same tendency as in the XZ plane. Furthermore, the jet interaction zone in the YZ plane was very thin and independent of the configuration. The PIV measurements in the YZ plane showed that the recirculation zone was only influenced by 2D motions in the XZ plane. To study the velocity fields in the recirculation zone, where the velocities were low (less than 1 m s−1), it was necessary to obtain important temporal and spatial resolutions. Table 2 summarizes the different magnifications, the spatial resolutions of the interrogation windows, and the time intervals, Δt, between two PIV images, corresponding to each burner configuration. Figure 4 displays the 2D velocity fields, refined for the recirculation zone. When the temporal and spatial resolutions were increased, the information obtained regarding the natural gas and oxygen jet velocities was not relevant; however, the velocity field between the two jets was more accurate. The correlation between the nozzle distance, D, and the size of the recirculation zone, Szr, for cases with and without the addition of hydrogen is depicted in Figure 5. This figure clearly shows an increase of the recirculation zone with a greater nozzle distance, D. When D increased, the hythane jet and the oxygen jet interacted more intensely downstream, the height of the mixing point Zmp increased, and the surface of the recirculation zone grew in size. In the case of a pure natural gas configuration (αH2 = 0%), when D increased from 12 to 100 mm, the size of the recirculation zone grew by a factor of 400 (from 22 mm2 ± 10% to 8900 mm2 ± 3%). The increase of the recirculation zone was less significant when the fuel jet moved away from the oxygen jet. The hydrogen volume fraction also influenced the size of the recirculation zone, in particular for large distances between

Figure 3. Mean 2D velocity field and the characteristic zones of a flow with separated jets (case: Φ = 1, αH2 = 20%, and D = 60 mm). The red square displays the location of the refined PIV in the recirculation zone.

Figure 4. Instantaneous and average 2D velocity fields, refined for the recirculation zones of the flow (Φ = 0.7, αH2 = 20%, and D = 60 mm). 4706


Energy & Fuels

Article

Table 2. Magnification, Spatial Resolution of the Interrogation Windows, and Time Interval Δt of PIV with the Distance between the Nozzles magnification (mm pixel−1) interrogation windows (mm2) Δt (μs)

D = 12 mm

D = 24 mm

D = 40 mm

D = 60 mm

D = 80 mm

D = 100 mm

0.0316 2.02 × 2.02 150

0.0328 2.09 × 2.09 150

0.0364 2.33 × 2.33 200

0.0392 2.51 × 2.51 200

0.0419 2.68 × 2.68 250

0.0456 2.92 × 2.92 250

Figure 5. Surface of the recirculation zone as a function of the distance between the nozzles for hydrogen volume fractions αH2 = 0 and 20%.

Figure 6. Temperature of the fumes at the chamber exit (Z = 1240 mm) versus the size of the recirculation zone of burned gases Szr, for the hydrogen content in the fuel at 0, 10, and 20%.

the nozzles. In the case of D = 100 mm, when the hydrogen volume fraction increased from αH2 = 0% (pure natural gas) to αH2 = 20%, the size of the recirculation zone was reduced by 12%. The high molecular diffusivity of hydrogen favored the mixing between the jets, which allowed for an early fusion of the jets. From a thermal point of view, the combustibility and heat release rates of hydrogen encouraged the reaction and the quick expansion of the burned gases, which reduced the size of the recirculation zone.35 3.2. Temperature of the Burned Gases. It was important to measure the temperature in the combustion chamber to study its influence on NOx emissions. Because the flame structure differed significantly depending upon the configuration (D and αH2), the temperature of the fumes was the common criterion chosen to follow the evolution of the temperature as a function of the study parameters. This measurement of the combustion product temperature was performed inside the convergent section of the combustion chamber (Z = 1240 mm) using K thermocouples. Results are illustrated in Figure 6 as a function of the surface of the recirculation zone of combustion products, Szr. The increase of the size of the recirculation zone of burned gases induced a decrease of the burned gas temperature. In fact,

in the case of pure natural gas (αH2 = 0%), the temperature of the fumes decreased from 649 ± 3 °C for Szr = 23 ± 2 mm2 (D = 12 mm) to 593 ± 2 °C for Szr = 8933 ± 160 mm2 (D = 100 mm). The relationship between the increase of the recirculation zone and a greater distance between the nozzles promoted the dilution of the flame with combustion products, therefore reducing the flame temperature. Figure 6 shows that the hydrogen addition resulted in a slight increase of the fume temperatures. Note that the variation in fume temperatures was very similar regardless of the distance between the jets. For the configuration of D = 12 mm and a hydrogen addition of 20% (Szr = 20 ± 2 mm2), the temperature of the fumes increased from 649 ± 3 °C to 674 ± 3 °C, i.e., 25 °C higher than for the pure natural gas flame. The higher fume temperature was due to the increased temperature of the hydrogen-enriched flame. Moreover, with the hydrogen addition, the reduced size of the recirculation zone limited the dilution of the flame with the combustion products and increased the temperature of fumes.36 3.3. NOx Emissions. Because of the high temperatures occurring in oxy-fuel combustion, the predominant NOx formation is a result of the thermal NO mechanism.37 The NOx formation in the present system was caused by the presence of nitrogen in the fuel (2% in volume) and air leaks. 4707


Energy & Fuels

Article

between D = 40 and 100 mm. The NOx rate variation in the first stage was 600 ppm (from 690 to 90 ppm), while this variation was about 65 ppm (from 90 to 25 ppm) for the second stage (in the case of αH2 = 0%). A greater spacing between the jets resulted in higher recirculation rates of the reactants caused by the combustion products and the temperature decreased, thus limiting the NOx formation through the thermal NO mechanism. With regard to the hydrogen effect, the results indicated that NOx emissions increased the greater the hydrogen volume fraction in the fuel. This increase can be explained by the rise in flame temperature, which mainly promoted the formation of thermal NO. The high diffusivity of hydrogen was able to reduce the size of the recirculation zone, leading to a decreased combustion product dilution of the flame. This limited the temperature reduction and, consequently, the NOx rate. 3.4. NOx Concentrations in Relation to the Size of the Recirculation Zone. This section is focused on the relationship between NOx emissions and the size of the recirculation zone. The results of the size of the surface area of the recirculation zone (Figure 5) and NOx emissions (Figure 7) as a function of the distance between the nozzles, D, indicated high NOx emissions for a small recirculation zone size and weak NOx emissions for a large burned gas recirculation zone, regardless of the hydrogen volume fraction. The relationship between NOx emissions and the size of the recirculation zone conformed to a hyperbolic decrease. This tendency prompted us to study the inverse ratio of NOx emissions as a function of the size of the recirculation zone (Figure 8). For the case of a burner with two separated jets, NOx emissions at the exit of the fumes were inversely proportional to the size of the recirculation zone of the burned gases (1/ NOx ∝ Szr). This tendency allowed us to derive the amount of NOx emissions from the size of the recirculation zone and proved the efficiency of nozzle separation to reduce NOx concentrations. The size of the recirculation zone of the burned gases and its impact on flame temperatures was essential to the reduction of NOx emissions with the Zeldovich mechanism.

Figure 7 shows the variation of NOx emissions as a function of the distance between the nozzles, D, for the hydrogen volume

Figure 7. NOx emissions as a function of the distance between the nozzles D for the hydrogen volume fraction αH2 = 0, 10, and 20%.

fraction in the fuel αH2 = 0, 10, and 20%. The effect of the separation distance between the jets on NOx emissions was significant for all cases of hydrogen content in the fuel, as shown in Figure 7. NOx emissions noticeably decreased when the distance D increased. For the pure natural gas case (αH2 = 0%), the NOx rate decreased from 690 to 25 ppm. The NOx rate distribution with D was observed in two stages: a very high decay between D = 12 and 40 mm, followed by a slow decay

Figure 8. Inverse of NOx emissions according the surface area of the recirculation zone for different hydrogen volume fractions αH2. 4708


Energy & Fuels

4709

25 32 37 41 160 147 135 120 ± ± ± ± 8933 8669 8367 7872 593 587 548 507 36 48 55 58 156 145 129 112 ± ± ± ± 5854 5759 5581 5312 596 591 552 509 49 52 58 78 95 77 60 51 ± ± ± ± 3359 3323 3255 3187 613 606 562 515 91 99 120 138 37 32 27 23 ± ± ± ± 1287 1260 1224 1175 626 613 579 537 230 265 288 310 21 18 15 13 ± ± ± ± 364 358 346 332

Szr (mm2) Szr (mm2) Szr (mm2) Szr (mm2)

690 726 752 787 23 22 20 19 649 624 588 560 1 0.9 0.8 0.7

Φ

2 2 2 2 ± ± ± ±

645 618 584 562

Szr (mm2)

D = 100 mm

T (°C) NOx (ppm) D = 80 mm



T (°C) NOx (ppm)

Table 3 summarizes the results for fume temperatures, T, the size of the recirculation zone, Szr, and NOx emissions in relation to the global equivalence ratio (from Φ = 0.7 to 1), without the addition of hydrogen (αH2 = 0%). A decreased fuel flow rate enabled the fuel jet to further interact with the oxygen jet upstream of the flow. Increasing the velocity difference between the two jets led to a greater diversion of the fuel jet toward the oxidizer jet; the jets mixed more upstream and reduced the size of the recirculation zone, Szr. Decreasing the global equivalence ratio resulted in lower flame temperatures. This decrease appeared to be the result of a reduction in the combustion product temperature, as shown in Table 3. For example, for the configuration with D = 12 mm, the temperature of the fumes decreased from 649 ± 3 °C for Φ = 1 to 560 ± 2 °C for Φ = 0.7. Figure 9 illustrates NOx emissions and equivalence ratios for different distances between the jets. As shown below, increasing the spacing between the jets decreased the NOx rate. A decrease of the global equivalence ratio was followed by an increase in NOx emissions. Note that this effect is opposite the size of the recirculation zone. For a nozzle distance of D = 40 mm, NOx emissions increased from 91 ppm for Φ = 1 to 140 ppm for Φ = 0.7. This tendency of NOx emissions was due to the destruction and formation of NOx through the thermal and Fenimore mechanisms. The mole fraction of CO2 in combustion products decreased with a lower global equivalence ratio Φ, because of the presence of excess oxygen in lean combustion. This reduction of CO2 emissions in a lean flame induced a decrease in the amount of recirculated CO2, which chemically reduced the nitrogen oxide

D = 24 mm

(1)

T (°C)

+ 2.165(1 − Φ)O2

NOx (ppm)

→ Φ[(1.12CO2 + 0.01CO2 + 0.02N2) + 2.09H 2O]

Szr (mm2)

+ 0.02N2] + 2.165O2

D = 12 mm

Φ[0.85CH4 + 0.09C2H6 + 0.03C3H8 + 0.01CO2

T (°C)

Table 3. Temperature of Fumes T, the Size of the Recirculation Zone Szr, and NOx Emissions According to the Global Equivalence Ratio (from Φ = 0.7 to 1) without Hydrogen Addition (αH2 = 0%)

3.5. NOx Emissions in Relation to the Equivalence Ratio. The study of NOx emissions as a function of the size of the recirculation zone for different hydrogen volume fractions (from αH2 = 0 to 20%) showed that NOx emissions heavily depend upon recirculated combustion products. For oxy-fuel combustion of diffusion flames, Park et al.38 showed that NOx emissions were linked to recirculated CO2 in the flame. They focused on the role of the recirculated CO2 and its chemical effects on the formation and destruction of NOx through Fenimore (prompt NO) and Zeldovich (thermal NO) mechanisms. In oxy-fuel combustion, the recirculated CO2 modifies oxidation reaction pathways and affects NOx emission behavior in both thermal NO and prompt NO mechanisms. The interruption of the formation and destruction of NOx depends upon the behavior of the recirculated CO2. From an experimental point of view, the unpredictability of air infiltration, the presence of nitrogen in the fuel, the size of the recirculation zone, and the aerodynamic straining of the flow make the study of NOx emissions difficult. The parameter retained to vary the amount of recirculated CO2 in oxy-fuel combustion is the global equivalence ratio, Φ. On the basis of the combustion equation (eq 1) of natural gas and pure oxygen, a variation in the global equivalence ratio, Φ, modifies the CO2 rate in combustion products.

NOx (ppm)

Article


Energy & Fuels

Article

tional to the size of the recirculation zone of the combustion products. With CO2 was recirculated, the global equivalence ratio had a significant effect on NOx emissions because recirculated CO2 reacted with prompt NO and thermal NO and reduced NOx emissions. The decrease of recirculated CO2, in combination with lean combustion, limited the reaction between NO and CO2. NOx emissions in lean combustion were higher than in a stoichiometric configuration. We conclude that, in oxy-fuel combustion occurring in burners with two separated jets, NOx emissions depend upon the size and composition of the recirculation zone of the burned gases, which are central to the goal of limiting NOx emissions at the exit of the fumes.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

REFERENCES

(1) Drake, M. C.; Blint, R. J. Combust. Flame 1989, 76, 151−167. (2) Chen, R. H.; Driscoll, J. F. Proc. Combust. Inst. 1990, 23, 281− 288. (3) Barlow, R. S.; Carter, C. D. Combust. Flame 1996, 104, 288−299. (4) Drake, M. C.; Blint, R. J. Combust. Flame 1991, 83, 185−203. (5) Naha, S.; Aggarwal, S. K. Combust. Flame 2004, 139, 90−105. (6) Chen, R. Y.; Driscoll, J. F. Proceedings of the 23rd Symposium (International) on Combustion; The Combustion Institute, Pittsburgh, PA, 1991; p 281. (7) Røkke, N. A.; Hustad, J. E.; Sønju, O. K. Combust. Flame 1994, 97, 88−106. (8) Kim, M.; Yoon, M. Proc. Combust. Inst. 2007, 31, 1609−1616. (9) Røkke, N. A.; Hustad, J. E.; Sønju, O. K.; Williams, F. A. Proc. Combust. Inst. 1992, 24, 385−393. (10) Turns, S. R.; Myhr, F. H. Combust. Flame 1991, 87, 319−335. (11) Kim, S. H.; Kim, M.; Yoon, Y.; Jeung, I. S. Proc. Combust. Inst. 2002, 29, 1951−1956. (12) Peters, N.; Donnerhack, S. Proceedings of the 18th Symposium (International) on Combustion; The Combustion Institute, Pittsburgh, PA, 1981; p 33. (13) Turns, S. R.; Lovett, J. A.; Sommer, H. J. Combust. Flame 1989, 77, 405−409. (14) Hwang, C. H.; Lee, S.; Lee, C. E. Int. J. Hydrogen Energy 2008, 33, 832−841. (15) Peters, N.; Donnerhack, S. Proc. Combust. Inst. 1981, 18, 33−42. (16) Driscoll, J. F.; Chen, R. H.; Yoon, Y. Combust. Flame 1992, 88, 37−49. (17) Santoro, V. S.; Kyritsis, D. C.; Smooke, M. D.; Gomez, A. Proc. Combust. Inst. 2002, 29, 2227−2233. (18) Frank, J. H.; Barlow, R. S.; Lundquist, C. Proc. Combust. Inst. 2000, 28, 447−454. (19) Gopalakrishnan, P.; Bobba, M. K.; Seizman, J. M. Proc. Combust. Inst. 2007, 31, 3401−3408. (20) Nada, Y.; Parwatha, I. G.; Fukushige, S.; Noda, S. Nippon Kikai Gakkai Ronbunshu, B-hen 2008, 74, 707−714. (21) Shinomori, K.; Katou, K.; Shimokuri, D.; Ishizuka, S. Proc. Combust. Inst. 2011, 33, 2735−2742. (22) Sautet, J. C.; Boushaki, T.; Salentey, L.; Labegorre, B. Combust. Sci. Technol. 2006, 178, 2075−2096. (23) Boushaki, T.; Sautet, J. C.; Labegorre, B. Combust. Flame 2009, 156, 2043−2055. (24) Salentey, L. Ph.D. Thesis, University of Rouen, Rouen, France, 2002.

Figure 9. NOx emissions as a function of the global equivalence ratio Φ for different distances between the nozzle D.

destruction.39 In diffusion oxy-flames with recirculating burned gases, prompt NO and thermal NO reacted with CO2 and decreased NOx emissions.38 The decrease of the CO2 amount in the recirculation zone led to higher NOx emissions. In the burner with separated jets, we found that NOx emissions were highly dependent upon the size and composition of the recirculation zone of the burned gases.

4. CONCLUSION We experimentally investigated the effects on NOx emissions of recirculating burned gases in natural gas−hydrogen−oxygen flames from a burner with separated jets. A burner composed of two nozzles and developed inside a combustion chamber supplied the oxy-flame. Measurements of NOx concentrations were carried out at the exit of the combustion chamber using a water-cooled probe and a NOx analyzer. A greater distance between the nozzles increased the size of the recirculation zone and favored the dilution of the reactants. Diluting the reactants with combustion products led up to a 96% decrease in NOx emissions. Without the addition of hydrogen, the NOx rate decreased from 690 ppm for D = 12 mm to 25 ppm for D = 100 mm. The dilution of the reactants with combustion products decreased both the flame temperature and the NOx formation, particularly through the thermal mechanism. The high molecular diffusivity and combustibility of hydrogen promoted mixing between the jets and reduced the size of the recirculation zone. The dilution of the oxy-flame with combustion products was less significant when hydrogen was added and the flame temperature increased. In the case of D = 12 mm, NOx emissions increased up to 39% when 20% hydrogen was added to the fuel. In oxy-fuel combustion with non-premixed flames, NOx emissions were inversely propor4710


Energy & Fuels

Article

(25) Zhakatakev, T. A. Circulatory Flow in the Inner Jet Zone of a System of Free Turbulent Jet; Plenum Publishing Corporation: New York, 1994. (26) Choudhuri, A. R.; Gollahalli, S. R. Int. J. Hydrogen Energy 2000, 25, 451−462. (27) Cozzi, F.; Coghe, A. Int. J. Hydrogen Energy 2006, 31, 669−677. (28) Anderson, D. N. Effect of Hydrogen Injection on Stability and Emissions of an Experimental Premixed Prevaporized Propane Burner; National Aeronautics and Space Administration (NASA): Washington, D.C., 1975; Report TM X-3301. (29) Tseng, C. J. Proceeding of the 36th Intersociety Energy Conversion Engineering Conference; Savannah, GA, 2001; pp 589−594. (30) Briones, A. M.; Aggarwal, S. K.; Katta, V. R. Combust. Flame 2008, 153, 367−386. (31) Yon, S.; Sautet, J. C. Appl. Therm. Eng. 2012, 32, 83−92. (32) Perthuis, E. La Combustion Industrielle; Technip: Paris, France, 1983. (33) Genies, B. A.D.E.M.E 1996, 153−196. (34) Sung, C. J.; Law, C. K. Proc. Combust. Inst. 1998, 27, 1411− 1418. (35) Kim, S. H.; Arghode, V. K.; Linck, M. B.; Gupta, A. K. Int. J. Hydrogen Energy 2009, 34, 1054−1062. (36) Zhang, Y.; Wu, J.; Ishizuka, S. Int. J. Hydrogen Energy 2009, 34, 519−527. (37) Fenimore, C. P. Proceedings of the 17th Symposium (International) on Combustion; The Combustion Institute, Pittsburgh, PA, 1979; pp 661−669. (38) Park, J.; Park, J. S.; Kim, H. P.; Kim, J. S.; Kim, S. C.; Choi, J. G.; Cho, H. C.; Cho, K. W.; Park, H. S. Energy Fuels 2007, 21, 121−129. (39) Watanabe, H.; Marumo, T.; Okazaki, K. Energy Fuels 2012, 26, 938−951.

4711


Effects of Burned Gas Recirculation on NOx Emissions from Natural

Recommend Documents