Influence of nozzle height to width ratio on ignition and NO_x emission characteristics of semicoke/bituminous coal blends in a 300 kW pulverized coal-fired furnace

Abstract

To improve the ignition behavior and to reduce the high NO_x emissions of blended pulverized fuels (PF) of semicoke (SC), large-scale experiments were conducted in a 300 kW fired furnace at various nozzle settings, i.e., ratios (denoted by h_f/b) of the height of the rectangular burner nozzle to its width of 1.65, 2.32, and 3.22. The combustion tests indicate that the flame stability, ignition performance, and fuel burnout ratio were significantly improved at a nozzle setting of h_f/b = 2.32. The smaller h_f/b delayed ignition and caused the flame to concentrate excessively on the axis of the furnace, while the larger h_f/b easily caused the deflection of the pulverized coal flame, and a high-temperature flame zone emerged close to the furnace wall. NO_x emissions at the outlet of the primary zone decreased from 447 to 354 mg/m³ (O₂ = 6%), and the ignition distance decreased from 420 to 246 mm when the h_f/b varied from 1.65 to 3.22. Furthermore, the ratio (denoted by S_R/S_C) of the strong reduction zone area to the combustion reaction zone area was defined experimentally by the CO concentration to evaluate the reduction zone. The S_R/S_C rose monotonously, but its restraining effects on NO_x formation decreased as h_f/b increased. The results suggested that in a test furnace, regulating the nozzle h_f/b conditions sharply reduces NO_x emissions and improves the combustion efficiency of SC blends possessing an appropriate jet rigidity.

Graphical abstract

Keywords

rectangular jet burner / nozzle height to width ratio / ignition characteristics / pyrolyzed semicoke (SC) and bituminous blend / NO_x formation

1 Introduction

Coal pyrolyzed semicoke (SC), which is consumed in pulverized-coal boilers due to its high calorific value, is abundant in reserves in China. The co-combustion of SC and raw coal, which is one promising disposal method, is preferred widely in practical applications. However, reducing NO_x emissions in the process of the co-combustion of SC and raw coal is one of the challenges to be faced by power plants. Nitrogen oxide (NO_x) compounds damage the ecological environment and threaten human health. At the same time, there exist some problems, such as difficulty in ignition, a low burnout rate, a low volatile content, and high NO_x emissions, in the SC combustion process. In recent years, studies on the co-combustion of SC and bituminous coal blends have been conducted. Yao et al. [1] performed thermogravimetric analysis to study the combustion characteristics and kinetic behavior of blends in an oxygen-enriched atmosphere. The co-firing experiment performed in the thermogravimetric analyzer showed observably different combustion characteristics depending on the proportion of bituminous coal in the blends. Zhang et al. [2] conducted thermogravimetric analysis and drop tube furnace experiments to study the combustion and NO formation characteristics during the co-combustion of bituminous coal with SC. The results showed that a blending proportion of 80% for bituminous coal had a good promoting effect on burnout and a significant inhibiting effect on NO formation. Nevertheless, as the blending proportion of SC increases, reducing NO_x emissions is confronted with challenges. Currently, the large-scale blending of SC in a tangentially fired utility boiler has generally become a promising disposal method. Thus, combustion condition optimization is one of the primary measures to reduce NO_x emissions from boilers burning SC and its blends [3]. In particular, the parametric optimization design of the low-NO_x tangential burner plays a critical part in improving the ignition and combustion characteristics of SC and inhibiting the formation of NO_x in the tangentially fired utility boiler.

Many numerical and experimental studies on large-scale tangentially fired boilers were widely and comprehensively performed. Researchers modeled tangentially fired pulverized fuel (PF) boilers by using CFD to study the prediction and decrease of temperature deviation and flue gas velocity deviation [4–6]. Guo et al. [7] proposed a compatible configuration strategy for burner streams of air combustion and oxy-fuel combustion in a 200 MW_e tangentially fired boiler to decrease the amount of oxygen consumption and reductive atmosphere near the membrane-wall, which may conceivably reduce corrosion and slagging. The NO_x, fluid flow and thermal characteristics were calculated by Habib et al. [8] for the flow in a tangentially fired furnace under different conditions of tilting burners. Tang et al. [9] developed a three-dimensional CFD model to simulate the flow characteristics and combustion process in a 200 MW tangentially fired boiler.

Obviously, most researchers have been emphasizing the overall combustion performance in large-scale tangentially fired boilers and have obtained systematic findings. However, separate detailed research measuring the flame field closely downstream of one of the four tangential burners in a hot-test furnace simulating actual burner operation conditions is a priority in laboratory studies. The PF ignition characteristics of the tangential burner play a vital role in increasing combustion efficiency and reducing NO_x emissions in the PF combustion of tangentially fired utility boilers. The selection of proper design parameters for co-firing coal with ignition characteristics significantly contributes to developing boiler PF tangential burners.

Many scientific researchers have studied the ignition and combustion characters of PF streams using many experimental systems, such as drop tube furnaces (DTFs), but the heating modes in these experiments are different from the mode of PF combustion in the furnace of an actual tangentially fired burner [10–15]. Therefore, a pilot-scale experimental system comprising a large-scale pulverized coal-fired furnace was built in the Clean Coal Combustion Laboratory in Harbin Institute of Technology (HIT) to simulate an actual industrial fired furnace. Compared with DTFs, the miniature pilot-scale experimental system has achieved better experimental results for industrial applications due to the different ignition characteristics achieved with various heating modes [14,15]. The researchers from HIT, Zeng et al. [16–18] and Zhao et al. [19], established a PF test burner in this pilot-scale test furnace, which simulated the ignition and combustion process of a bias PF tangential burner of actual tangentially fired boilers. The effects of primary air (PA) properties, pulverized coal (PC) concentration, raw coal equivalent moisture, and pulverized coal fineness on the ignition characteristics of parallel bias PF jets were systematically studied, which improved the design and operation parameters for rational tangential burner industrial applications. As mentioned above, the physical properties of the coal and the air supply settings were mainly focused on the research of bituminous coal combustion.

However, appropriate design parameters of the burner for the blending combustion of bituminous coal and SC are the key to achieving a successful PF combustion configuration. Based on Refs. [16–18], the SC co-firing characteristics of parallel PF jets with different burner nozzle height to width ratios (i.e., h_f/b) were studied in this work to achieve a strong and steady ignition, reduce NO_x emissions, and prevent slagging [20]. In this work, to study the ignition characteristics of the blending combustion of bituminous coal and SC, three tangential burners with different nozzle h_f/b values were produced, which were installed on a 300 kW pilot-scale PF combustion test furnace with six jet intersections [16], and a large SC blending ratio (50%) and highly volatile bituminous coal were used. The optimal parameters of the tangential burner nozzle were determined from the ignition characteristics, which corresponded to the standoff distance, combustion temperature distribution, PF burnout rate, and axial and radial NO_x distribution in co-firing ignition. The findings of this work may help to provide a better understanding of and guidance for employing SC for large-scale blending fuel applications in tangentially fired boilers.

2 Experimental methodology

The National Engineering Laboratory for Reducing Emissions from Coal Combustion (NELRECC) at the Harbin Institute of Technology in China fabricated the pilot-scale bias combustion simulator (PBCS) system, as shown in Fig. 1, which mainly consists of a furnace, a bias pulverized coal transportation system, an air supply system, a high-temperature flue gas simulating system, a tail flue gas system, a measurement and sampling system, and an operation monitoring system [16], whose main design parameters are listed in Table 1. The horizontal bias combustion technology tested in a 300 kW pilot furnace is generally adopted in the actual tangentially fired boiler. In this kind of tangential burner, the primary air jet is divided into rich and lean pulverized coal jets by the concentrator at the furnace corner. The rich side of the PF jet is sprayed into the furnace at the fire side, and the lean side of the PF jet is sprayed into the furnace at the back fire side. The inner boundaries of the two jets are first mixed on their central axis; the rich PC jet is then ignited by the fire-side high-temperature gas; the lean PC jet also begins to combust due to continuous entrainment from the surrounding high-temperature gas, and subsequently, the secondary air participates in combustion. To study the ignition characteristics of the rich-lean PC jet in the tangential boiler, the primary/secondary air and high-temperature flue gas are arranged in a reasonable way in the test system. As shown in Fig. 2(a), from the top to the bottom of the furnace, two parallel PF/PA jets are reasonably injected into the primary combustion zone of the furnace, mixed with two flames from a high-temperature combustion-supporting propane gas burner, and then ignited. The experiments are conducted with a basic bias concentration ratio (BCR) of 1:1. Other BCRs will be studied subsequently in future tests. Two jets of secondary air (SA) then gradually become involved in combustion at an axial distance of 0–540 mm. The temperatures of the input PA and SA secondary air are maintained separately at 80°C and 230°C, and their nozzle velocities are 18.37 m/s and 18.88 m/s. As shown in Fig. 2(a), x and r represent the distance between the measuring point and the burner outlet for the axial and radial directions, respectively. The main analysis zone is within an axial distance of 180 mm, 340 mm, 500 mm, 660 mm, and 820 mm and a radial distance of 0 mm, 20 mm, 40 mm, 60 mm, 80 mm, 100 mm, 120 mm, 150 mm, and 200 mm. The bias PF transportation system consists of two variable-frequency adjustable-screw feeders, two equal-dimension PA ducts, and a direct-flow jet burner. Figure 2 demonstrates a schematic and photos of the three pulverized coal direct-flow burners with different h_f/b values, which are installed subsequently at the top of the furnace. The h_f/b value is defined as the ratio of the net height to the net width of the burner nozzle, including its nozzle wall thickness and axial sampling hole. Multiple means of measurement and sampling are used to acquire accurate information on the PF ignition characteristics in the 300 kW PBCS, including the axial and radial temperatures, flame images, flue-gas concentration, and residual solid inside the furnace. The effect of h_f/b on the evolution of O₂, CO, NO_x, and NO precursors of HCN was studied using a Fourier transform infrared (FTIR) spectrometer. The axial and radial temperatures in each predetermined location of the primary combustion zone were measured using Inconel armored type-K thermocouples having a diameter of 2 mm, and the data were recorded on a computer [21–23]. The standoff distance is defined as the distance from the burner nozzle to the flame onset after the PF mixture achieved stable ignition [14]. The other parameters of the experiments are tabulated in Table 1. When every parameter reached the set point during the pilot-scale hot-condition experiments, combustion was adjusted according to the combustion temperature, flame stability in the primary zone, and the O₂ content at the furnace outlet. Data was collected only if the PBCS was operated in a continuous and stable state to ensure data accuracy and repeatability. Because of normal fluctuations that occur during coal combustion, measurement precision was improved by averaging the testing values of repeated measurements. The temperature measurements and average values deviated in the range of 0°C–12°C, and the measured flue-gas component concentrations and their average values deviated by less than 1%. Details of the experimental method can be found in Ref. [17].

Some problems, such as the difficulty in achieving stable combustion, low burnout rate, and high NO_x emissions, are encountered while increasing the blending proportion of SC. However, the study on the ignition characteristics of the large-scale blending combustion of bituminous coal and SC can expand the coal adaptability of boilers and consume SC in large quantities. Therefore, based on Ref. [24], a mass blending ratio of SC and bituminous coal slightly higher than 50%:50%, which were both sourced from Shenhua Group Corporation Limited, one of the largest and most modern coal enterprise in China, was used in the experiments, and the mixture of SC and bituminous coal is referred to as “blended coal” hereafter. The proximate and ultimate analytical data obtained for the blended coal is presented in Table 2. The pulverized coal used in this study had a full particle size distribution, and the PF fineness was R₉₀ = 9% (i.e., the surplus after sieving to 90 mm was 9%).

3 Results and discussion

3.1 Cold air flow field numerical simulation results

Cold aerodynamic field velocities were simulated along the longitudinal cross section through the common centerline of the primary air and secondary nozzles to employ the commercial CFD code Fluent 2020 R1. In the down fired furnace, the pulverized-coal ignition and combustion process mainly depends on the flow field in the upper furnace, because pulverized-coal combustion is almost complete in the lower furnace. Consequently, only the cold aerodynamic flow within the front 2000 mm was acquired in this paper, as depicted in Fig. 3 where the left part is the plane figure of the burner nozzle, and the right part is the three-dimensional geometry of the furnace drawn in Space (The coordinate axis x corresponds to r and z to x in Fig. 2(a)). The primary air jet and second air jet were defined as velocity inlets. The velocity and temperature of the primary air were 18.37 m/s and 353 K, respectively. The mass flow rate of the coal particles carried in the primary air was 0.004583 kg/s per nozzle. The velocity and temperature of the second air were 18.88 m/s and 503 K, respectively. The combustion products and part of the nonreacted air exhausted through the exit were treated as the pressure outlets.

This simulation is intended to obtain a detailed set of data of cold air flow fields. Figure 4 presents the cold air flow contour and streamlines with respect to h_f/b settings in the y direction. For all the h_f/b values, after the primary and secondary air met, the mixture was relatively uniform. A wide high-velocity distribution zone was formed in the central area under the burner nozzle, and a low velocity zone was formed in the region near the furnace wall. To some extent, the flame brush wall can be alleviated or avoided to relieve or avoid the slagging of the water-cooled wall during actual operation. From the streamline diagram, it was observed that the airflow on both sides of the mixed airflow in the center of the furnace formed two elliptic vortices with the same size and symmetric distribution. The airflow flowed upward along the furnace wall on both sides, changed direction to flow in an opposing direction near the top of the furnace, and finally flowed downward with the secondary air. The vortex, also known as the recirculation zone, is conducive to the mixing of the airflow and entrainment of the surrounding high-temperature gas. The residence time of the combustible gas and unburned particles in the furnace is also prolonged to make them fully burn out. After the secondary air was slanted out from the nozzle, within a considerable distance, the airflow direction did not change significantly, and the attenuation was also small, indicating that the jet was rigid, which subsequently promoted the rigidity of the primary air jet. Before the mixing jet velocity decayed to 3 m/s, the jet velocity attenuation rate increased slightly with the increase in h_f/b. It can be concluded that the penetration and rigidity of the mixture airflow differ from h_f/b [20]. Compared with the case where h_f/b is 1.65 and 2.32, the length of the recirculation zone decreased significantly at h_f/b = 3.22. Accordingly, the instability in the recirculation zone increased. An effective recirculation zone allows a high-temperature and low-O₂ zone to develop easily. Consequently, the coal particle residence time in the high-temperature zone lengthens, and the combustion proportion in the furnace increases, resulting in the consumption of more oxygen. Additionally, a low-speed zone is also found to develop near the wall. Under real-furnace conditions, this adverse behavior may result in a clearly increased heat load distribution and slag depositing locally in the furnace. For the case where h_f/b is 2.32, a more preferable penetration depth and jet rigidity were gained synthetically. As mentioned above, the cold-state numerical results provide a useful guidance in the event that designing an optimal burner nozzle structure is practical for improving the blended solid fuel combustion performance in a 300 kW pulverized coal-fired furnace.

3.2 Effects on ignition characteristics of PF

Figure 5 displays the profiles of the temperature along the furnace centerline and the ignition point position at different h_f/b values. The change of the pulverized coal and gas mixture temperature at different h_f/b values is similar with increasing the cross-section x (Zeng et al.) [17]. At an axial distance of 0<x<100 mm, the temperature increased significantly. This increase occurred because the coal mixture pulverized by radiant heating from the high-temperature furnace wall was the dominant heat transfer mechanism before the two parallel PC jets intersected at the centerline. At an axial distance of 100<x<200 mm, the two parallel PF jets began to mix on the centerline, and the volatile components emitted from bituminous coal (and emitted slightly from SC) were released in large quantities, keeping the temperature relatively constant. At an axial distance of 200<x<500 mm, the temperature increased dramatically. The reason for this is that volatiles combust mainly by violent combustion due to the continuous heating from high-temperature flue gas entrainment (convective heat transfer) and the radiant heat transfer from the inner wall of the furnace. The initial ignition position also appeared. At an axial distance of x>500 mm, the temperature reached the second peak value and remained at a very high level. The volatile matter had burned off, and the heat of combustion mainly came from the char (including bituminous char and SC). The combustion process can be divided into the heating stage, the volatile releasing stage, the ignition stage, and the char-combustion stage. At the heating stage, the gas temperature rose rapidly because blended fuels were heated. As exhibited in Fig. 7, the O₂ concentration decreased to approximately 15%, indicating slight volatile ignition (mainly from bituminous coal). However, the increase in temperature is attributed to the radiant heating at this stage. At the volatile releasing stage, the temperature and O₂ changed little, because the heat mainly accounts for volatiles releasing from fuels. At the ignition stage, the temperature increased quickly, and the O₂ concentration decreased to a low value, in which the volatiles ignited and combusted intensely. However, a small quantity of char from blended fuels is ignited by volatile combustion. Therefore, although char mainly burns at the char-combustion stage, some residual volatiles still participate in char oxidation. It is usually difficult to distinguish homogeneous ignition from heterogeneous ignition in such a large-scale test system. The two-stage ignition for blended fuels was observed and slight ignition was counted at the heating stage. Note that the volatiles and char were the main fuels at the ignition stage and at the char-combustion stage, respectively, but at the heating stage, there was slight ignition from volatiles. To determine the ignition point position, the inflection point from the temperature profile versus the path axial distance curve (d²T/dx² = 0) was used [25]. The ignition point positions for h_f/b = 1.65, 2.32 and 3.22 were 420 mm, 246 mm, and 298 mm, respectively, where the corresponding ignition temperatures were 1014°C, 1034°C, and 1074°C. In comparison with the case at a h_f/b of 1.65, when h_f/b = 2.32 and 3.22, the ignition point position was shorter because of the larger nozzle perimeter achieved with h_f/b, which led to the entrainment of sufficiently high-temperature flue gas to ignite the PC and promote ignition [26]. However, the ignition stand-off distance in the case when h_f/b = 3.22 was slightly further away than that achieve with the case of h_f/b = 2.32. The fluid rigidity of the jet was weak when h_f/b increased, causing the jet cross section maximum velocity line of the PF jet to turn far away from the centerline. Hence, the PC concentration (mainly that of large coal particles) and O₂ concentration were both weakened along the centerline. This resulted in a delayed ignition and an increase in ignition temperature [27–30].

Figure 6 shows the images of flames at x = 340 mm for different h_f/b values. The flame brightness and the flame stability were significantly improved when h_f/b = 2.32. This can be explained by the fact that with the h_f/b value increasing, the peripheral circumference (i.e., 206.06 mm, 213.92 mm, and 225.32 mm) of the two-row nozzles increased, and the heating surface area of the pulverized coal jet at the same distance increased, which enhanced convection heat transfer and radiative heat transfer. No matter the burner shape is chunky or slender, the burner with h_f/b = 2.32, which was approximately a square, had an appropriate structure, ensuring the suitable concentration distribution and rigidity of the PF jet. After the pulverized coal jet was ejected from the burner nozzle, it was quickly mixed and disturbed with the high-temperature propane gas and flame, which helped the pulverized coal to ignite and combust. The jet with a smaller expansion angle at h_f/b = 1.65 was so concentrated that the second air participated in the combustion process later, increasing the stand-off distance and resulting in a low combustion efficiency, flame instability or even flameout. Similarly, due to the high dispersion, the weak rigidity, and the large fullness of the pulverized coal jet for h_f/b = 3.22, ignition occurred earlier, the burner nozzle burned out easily, the flame was pressed against the furnace wall, and furnace slagging was caused. The PF concentration achieved with the burner at h_f/b = 2.32 was quite high, which enhanced the combustion reaction rate. This would dramatically increase the furnace temperature, in turn increasing the average kinetic energy of molecules and fuel molecule collision from the aspect of air dynamics. In addition, the effective airflow could better fill the furnace space with the burner at h_f/b = 2.32.

3.3 Gas species profiles along axial and radial direction

Figure 7 presents the profiles of the O₂, CO, and HCN concentration along the furnace centerline at different h_f/b values. According to the curve maximal slope of the gas species, the ignition point position was also determined, which was expected for the inflection point from the temperature profile. The standoff distance determined by the O₂ concentration change was nearly identical to that determined by temperature at h_f/b was 1.65. However, the deviation between the standoff distance determined by O₂ and that by temperature gradually increased as h_f/b increased, illustrating the inconsistency of the ignition point position determined by the two means. The reason for this is that the structure of the burner nozzle has an effect on the PF concentration field. The high concentration PF was gradually off-center when h_f/b was varied from 1.65 to 3.22, resulting in lower O₂ consumption rates along the axial direction than on the radial direction. As shown in Fig. 7, an inflection point existed in the CO and HCN curves, and the positions of the two point were consistent with the ignition position determined by O₂. This indicated that the inflection point of the CO and HCN curves was also determined to be the ignition point. Before the PF ejected from the nozzle to the inflection point, the volatile matter released, and the CO and HCN concentrations were low. The first so-called standoff distance determined by CO and HCN was too short to be considered as an ignition point. After the PF reached the inflection point, combustion became more intense, and char-combustion began. The second so-called standoff distance was not obviously evaluated to be an ignition point. In conclusion, the O₂ and CO (HCN) concentration variations were considered to determine the standoff distance during the PF combustion reaction under certain conditions.

Figure 8 shows the profiles of the gas temperature and O₂, CO, and NO_x concentrations at different h_f/b values while Fig. 9 shows their 2D profiles. For the cross-section x = 180 mm, the gas temperatures were all above 930°C. For 80≤r≤160 mm, the O₂ concentrations were quite low, while the CO concentrations were pretty high, which indicated that part of the outer pulverized coal started to ignite first. For 0≤r≤120 mm, at different h_f/b values, the peak values of the gas temperature reached 1011°C, 1139°C, and 1010°C, increasing initially and decreasing afterwards. The valley values of the O₂ concentration were 12.77%, 9.69%, and 10.9%, respectively, declining and then ascending, and the peak values of the CO concentration were respectively 2.351 × 10⁻³, 3.363 × 10⁻³, and 3.4668 × 10⁻³, first increasing and subsequently increasing gently. The NO_x concentration respectively peaked 643, 858, and 746 mg/m³ at 6% O₂, first increasing and then decreasing. The points where the peak values and valley values of these gas concentration deviated from the centerline as h_f/b increased, which resulted from the following two reasons. First, the two parallel PC jets intersected at the centerline to absorb heat so that the outer volatiles and fine coal particles first caught on fire due to high-temperature flue gas entrainment. Second, under the same cross-sectional areas of the three burner nozzles, the nozzle perimeter enlarged as h_f/b increased to enhance the ability to entrain high-temperature flue gas sufficiently; thus, the O₂ consumption was accelerated, the CO and NO_x production increased, and the heat released by the pulverized coal combustion within the same distance from the burner nozzle increased. However, the rigidity of the jet weakened, and the velocity of the jets decreased faster. Since h_f/b= 2.32, the ignition position was deflected from the centerline. Hence, the O₂ consumption slowed down, the CO production flattened, and the NO_x production decreased after the optimal h_f/b value was exceeded. For 120≤r≤200 mm, at different h_f/b values, the temperatures all reduced gently, the O₂ concentration gradually increased, the CO concentration decreased and tended to be flat, and the NO_x concentration after conversion computing increased with O₂ increasing due to the burner secondary air jet area.

At the cross-section of x = 340 mm, for 0≤r≤120 mm, the peak temperature at three h_f/b values was higher than that achieved at the cross-section of x = 180 mm, the central temperature was significantly higher than the surrounding temperature, and the valley values of the O₂ concentration decreased (as shown visually in Fig. 5), indicating that the large amount of heat released by the outer pulverized coal of the jets ignited the inner pulverized coal. Especially at h_f/b = 2.32, for 0≤r≤ 80 mm, the O₂ concentration was close to 0, the CO concentration peak increased vastly, and the NO_x concentration peak decreased. This occurred because, due to the progress of pulverized coal combustion, the O₂ near the central area was further consumed, and the CO generated rapidly at low O₂ concentrations. Compared with the formation rate of the NO, the consumption of O₂ was faster, and the reduction atmosphere lowered the NO_x concentration after conversion computing. Meanwhile, at h_f/b = 1.65 and 3.22, the O₂ concentration decreased slightly, the CO concentration peak increased slightly, and the NO_x concentration increased, indicating that the combustion was unstable and inadequate. For 120≤r ≤200 mm, compared with the cross section of x = 180 mm, the temperature, O₂ concentration, and CO concentration at different h_f/b values changed little, while the concentration of NO_x tended to flatten. This radial section range was probably located at the edge of the primary combustion zone where the combustion reaction was weak.

From the cross-section of x = 500 mm to 820 mm, the temperature peaks of the three cases increased slightly and finally increased gently. The temperature at h_f/b= 2.32 was always greater than that achieved with the other two, and the temperature was stable above 1200°C (as shown in Fig. 5), illustrating that the burner at h_f/b = 2.32 was conducive to strongly reinforcing combustion. Compared with the cross-section of x = 340 mm, at h_f/b = 1.65, the O₂ concentration of 0≤r≤40 mm, which was close to 2%, started to decrease significantly and then gradually rose to a peak at 80≤r≤200 mm, which decreased from 19% to 14% as the cross-section x increased. At h_f/b = 2.32, the O₂ concentration of 0≤r≤80 mm, which was close to 0%, showed few changes and then gradually rose to a peak at 80≤r≤200 mm, which decreased from 16% to 10% as the cross-section x increased. Meanwhile, at h_f/b = 3.22, the O₂ concentration of 0≤r≤20 mm began to decrease remarkably and stabilized at 4% as x increased, and the radial region with a zero O₂ concentration, which was off centerline, gradually expanded from 80≤r≤120 mm to 40≤r≤120 mm. It was observed that the reaction zone slightly widened as x increased, which occurred due to the expansion of the central jet as the combustion process progressed [31]. The whole pulverized coal in these three cross-sections had ignited completely, and the structures of the burners at different h_f/b values influenced the distribution of the volatiles. The CO concentration peak at h_f/b = 1.65 increased to the maximum in the cross-section of x = 82 0 mm, while the CO concentration peaks at h_f/b = 2.32 and 3.22 increased to the maximum in the cross-section of x = 500 mm and gently increased as x increased. This can be explained by the fact that the ignition at h_f/b = 1.65 was delayed, and the main combustion zone stayed further away than the latter two from the burner nozzle. In addition, the CO concentration peaks at h_f/b= 3.22 appeared at r = 100 mm in each cross-section because of the minimum of the O₂ concentration. As the cross-section x increased, the NO_x concentration at h_f/b = 1.65, 2.32 and 3.22 decreased as the CO concentration increased. In the cross-section of x = 820 mm, the NO_x concentration at h_f/b = 1.65 dropped to 374 mg/m³ at 6% O₂ in 0≤r≤20 mm, to 331 mg/m³ at 6% O₂ in 0≤r≤60 mm at h_f/b = 2.32, and to 384 mg/m³ at 6% O₂ in 60≤r≤120 mm at h_f/b = 3.22. It is known that the minimum NO_x concentrations were below 400 mg/m³ at 6% O₂, and the NO_x distribution area at h_f/b = 2.32 was wide and not scattered, unlike that of case 3.

Figure 9 presents the profiles of the NO_x concentration along the furnace centerline at different h_f/b values. The NO_x concentration along the axial direction in the cross-sections from x = 20 to x = 130 mm reduced sharply at different h_f/b values. The mean NO_x concentration in the cross-sections from x = 20 to x = 130 mm at h_f/b = 1.65, from x = 20 to x = 100 mm at h_f/b = 2.32, and from x = 20 to x = 100 mm at h_f/b = 3.22 reduced from 1300 to 659, 893 to 655, and 1138 to 786 mg/m³ at 6% O₂, respectively. The primary NO_x concentrations at different h_f/b values were so high due to conversion computing, which were then reduced as the O₂ concentration decreased, playing a more dominant role than that played for NO_x production. In the cross-sections from x = 100 to x = 910 mm, the PF combustion region was approximately divided into the preheating region, the growing flame region, and the continuous flame region in sequence [32,33]. First, the mixed pulverized coal particles that were heated produced volatile-N, which was mainly released in the form of the nitrogen oxide precursor HCN. With the gas temperature increasing, HCN was subsequently and continuously oxidized to become NO_x at a high oxygen concentration. Therefore, the oxygen and the HCN concentrations decreased, while the CO and the NO_x concentrations increased in the reheating region. Second, the peak values of the NO_x concentration in the growing flame region at h_f/b = 1.65, 2.32, and 3.22 were 911, 977, and 1031 mg/m³ at 6% O₂, respectively. The volatile-N and char-N were oxidized to produce more NO_x as the temperature increased and oxygen concentration declined rapidly, which was reversely reduced under reduction atmospheres so that the NO_x remained stable in the high concentration level. Finally, the NO_x concentration at h_f/b = 1.65, 2.32, and 3.22 first decreased sharply and finally decreased gently in the NO_x reduction zone, whereas the NO_x concentration at h_f/b = 2.32 was the lowest. These results can be attributed to three reasons. First, the CO and HCN concentrations began to rise substantially when the O₂ concentration at h_f/b = 1.65, 2.32, and 3.22 decreased to nearly 8%, which would gradually enhance the reductive atmosphere. Second, the O₂ concentration remained below 1% at h_f/b = 1.65 and 2.32 and below 4% at h_f/b = 3.22, and the CO concentration at h_f/b = 1.65, 2.32, and 3.22 peaked at 2.1243 × 10⁻², 2.8419 × 10⁻², and 1.6351 × 10⁻² owing to the abundant unburned coke combustion at low oxygen contents, which restrained NO_x formation tremendously. Finally, at h_f/b = 2.32, a shorter ignition distance was achieved than at h_f/b = 1.65, and a more stable pulverized-coal concentration distributed near the centerline than at h_f/b = 3.22 easily deviated from the axial direction, enlarging the continuous flame region, in which the NO_x concentration was low, and combustion was improved.

Figure 10 shows the 2D color-filled contours of the gas temperature and the O₂, CO, and NO_x concentrations at h_f/b = 2.32. The red region above the superhigh temperature 1197°C expands much more widely than at the other two h_f/b values, as its top is closer to the burner nozzle, and its bottom is obviously broad. This is attributed to the continuous flame region, as mentioned previously, with intense combustion. The blue region below the low-oxygen concentration of approximately 4.74%, whose shape is a reflection of the flame profile to a certain extent, lifts up to the furnace wall strikingly, which illustrates severe adherent burning, causing slagging, high-temperature corrosion or the burner nozzle to be easily overtemperature deformed [34,35]. Because of the large-scale consumption of O₂ and SC in a high-temperature atmosphere at h_f/b = 2.32, a large amount of CO, with a peak of up to 4.32 × 10⁻², is generated in the middle portion of the furnace to form a reducing atmosphere, effectively suppressing NO_x formation, and an ultralow NO_x concentration region develops below 502.1 mg/m³ at O₂ = 6%. Figure 11 shows the average concentrations of the gas components and average temperature in the high reduction zone at different h_f/b values. The region of CO>1.08 × 10⁻² is assumed experimentally as the high reduction zone. It was found that compared with h_f/b = 1.65 and 3.22, whose reduction rate was 13.1% and 20.9% respectively, the appropriate h_f/b of 2.32 had a better effect on NO_x emission. This was expected, since a proper nozzle structure formed a stronger reducing atmosphere effectively, which is instrumental for NO_x reduction because of the consumption of large-scale O₂ and production of a large amount of CO, as represented in Fig. 11. This is probably due to the augmentation of the PF residence time as a result of the low flow rate of the oxidizer, and a strongly reducing atmosphere could be formed by the reaction between oxygen and combustible gas species [36]. In addition, temperature has a great influence on CO denitrification, and the reduction efficiency of NO_x at 1550 K, i.e., 1280°C, is approximately 20% to 30% [37]. The rising of the reaction temperature had a positive effect on inhibiting char-N conversion into NO during the combustion process [38,39]. The flow structure is better, the jet rigidity is more preferable, and the jet is not attached to the wall when h_f/b equals 2.32 than when h_f/b equals 1.65 and 3.22. Meanwhile, the PF ignites moderately, and the flame temperature is high, restraining NO_x formation.

When investigating NO_x emissions, the char-N evolution routine should be considered because the amount of NO_x emissions will be determined eventually by the competition between NO formation and reduction. During the combustion of char, the main heterogeneous reactions of nitrogen, as proposed by De Soete [40], appear to be

(1){\mathrm{O}}_{\mathrm{2}}\mathrm{+}\left(\mathrm{-C}\right)\mathrm{+}\left(\mathrm{-CN}\right)\to \left(\mathrm{-CO}\right)\mathrm{+}\left(\mathrm{-CNO}\right)

(2)\left(\mathrm{-CNO}\right)\to \mathrm{NO+}\left(\mathrm{-C}\right)

(3)\mathrm{NO+C}\to \frac{\mathrm{1}}{\mathrm{2}}{\mathrm{N}}_{\mathrm{2}}\mathrm{+}\left(\mathrm{-CO}\right)

A small amount of HCN was detected in the combustion tests. Therefore, the homogeneous reaction of NO–CO played a major role in the reduction of NO, because the overall excessive air coefficient was 0.9 in this experiment, leading to a fuel-rich combustion. In the high reduction zone, the oxygen concentrations and consumption were both higher than those at h_f/b = 2.32, which resulted in a lower oxygen zone and a higher CO zone by enhancing the rate of Eq. (1), which will lower the NO_x formation in a high reduction atmosphere. Another possibility to be considered is that the resulting increasing particle temperature will also decrease the NO_x evolution by enhancing the rate of Eq. (3). Hence, the effect of aerodynamics achieved with different nozzles on the reducing capacity should be considered.

Reduction capacity is evaluated not only by the concentration of gas components but also by the percentage of the high reduction zone in the reaction zone. Figure 12 presents the ratio of the high reduction zone area (S_R) to the combustion zone area (S_C) for the carbon monoxide in different settings. As the height-width ratio increased, the S_R/S_C of CO increased monotonously. This demonstrates that the high reduction zone extended with the increase in h_f/b, which should decrease NO_x emissions. However, the ability to restrain the NO_x formation depended more on a high CO concentration value than on S_R/S_C. S_R/S_C may be used to predict or guide the range of the reduction zone in industrial boilers. In addition, the length of the recirculation zone was shortened at h_f/b = 3.22, as shown in Fig. 4, leading to the insufficient combustion of the blended PF. Therefore, it is observed from Fig. 11 that the O₂ average concentration in the reduction zone increased, and the CO average concentration decreased, which is unfavorable for inhibiting NO_x.

3.4 Effects on particle burnout at height-to-width ratios

The combustion efficiency of coal is often measured in terms of burnout. Figure 13 shows the gas temperature and the burnout of residuals sampled at an axial distance of 820 mm for blended SC and bituminous coal parallel PF jets at different height to width ratios. This demonstrated that burnout and temperature were consistent while the height-width ratio varied from 1.56 to 3.22, and there was a turning point for improving coal particle burnout. However, the blended coal burnout ratio tested was far below the value at the actual furnace exit. The reason for this is that, on the one hand, the main test combustion zone lacked oxygen due to its low excess air coefficient (0.9), which represented the primary combustion zone upstream of the over fire air (OFA) input into the actual furnace, on the other hand, the residence time of the blended coal particles tested was also short in the main combustion region (0.64–1.5 s). Additionally, as shown in Table 1, when h_f/b varied from 1.65 to 2.32 and 3.22, the unburnt carbon content in the fly ash was 9.73%, 0.74%, and 3.62%, respectively. These results show that reasonable h_f/b values improved the performance of combustion distribution, owning to a lower NO_x formation and a better burnout. An overall evaluation of the PC combustion characteristics and NO_x formation indicates that the burner at h_f/b = 2.32 is the optimum one.

Thus, in this work, it is shown that when the blending ratio of SC and bituminous coal is 50%:50%, the rectangular nozzle possesses a critical height-width ratio for positive NO_x reduction and a high burnout level. A smaller height to width ratio prolongs ignition and causes the flame to concentrate too much at the axis while a larger height to width ratio easily causes the deflection of the primary pulverized coal streams and the high temperature flame to attach to the wall whereas a suitable height to width ratio achieves a positive outcome to establish an acceptable flow field, corresponding to a better combustion and a lower NO_x formation performance at h_f/b = 2.32 than at the other two h_f/b values. These results provide a very useful reference for achieving pollutant control and a high burning efficiency in the existing PF furnace for burning bituminous coal and SC blended fuel.

The abovementioned numerical and experimental results clearly indicate the effect of the burner h_f/b on the flow-field rigidity, combustion and NO_x emissions. To obtain a better understanding of the complex formation and reduction paths of NO_x occurring during blending fuel combustion in different h_f/b settings, the data acquired will be used within thermal state CFD-simulations in the next step, and these results will be compared with the experimental results for further optimization of the burner geometry for pulverized coal combustion.

4 Conclusions

In this work, the influence of a rectangular jet burner nozzle with h_f/b on the ignition characteristics and NO_x formation of a 300 kW pulverized coal-fired furnace was investigated. The main results can be summarized as follows:

According to the gas components distribution, a small h_f/b value prolonged ignition and caused the flame to concentrate excessively on the axis, while a large h_f/b value easily caused the deflection of the primary pulverized coal streams and the high temperature flame to attach to the furnace wall.

The flame stability and performance of ignition and burnout were significantly improved at h_f/b = 2.32. The standoff distance and burnout rates of the residual solids sampled at an axial distance of 820 mm were 420 mm, 246 mm, 298 mm and 76.25%, 84.75%, 81.17% at h_f/b = 1.65, 2.32, and 3.22, respectively. The O₂ and CO (HCN) concentration variations can also be considered to determine the standoff distance when the jet rigidity was strong. In addition, the NO_x emissions at the outlet of the primary zone reduced by 20.8% to 354 mg/m³ (O₂ = 6%), and the carbon burnout ratio increased by 11.15% at h_f/b = 2.32.

The nozzle structure at different h_f/b values presented different aerodynamic characteristics (mainly different recirculation lengths and jet rigidity). Therefore, the distribution of oxygen concentrations and particle coal concentrations have an influence on NO_x formation. The S_R/S_C of CO increased monotonously as h_f/b increased. However, the ability to restrain NO_x formation depended more on the CO high concentration value and low oxygen atmosphere than on S_R/S_C.

Based on an evaluation of the flame stability, performance of the standoff distance, ignition, burnout, and low NO_x emissions, the recommended h_f/b value for the rectangular jet burner was 2.32 for burning bituminous coal and SC (1:1) blended fuel.

References

[1] Yao H, He B, Ding G, Thermogravimetric analyses of oxy-fuel co-combustion of semi-coke and bituminous coal. Applied Thermal Engineering, 2019, 156(6): 708–721

[2] Zhang J, Jia X, Wang C, Experimental investigation on combustion and NO formation characteristics of semi-coke and bituminous coal blends. Fuel, 2019, 247(7): 87–96

[3] Smrekar J, Potočnik P, Senegačnik A. Multi-step-ahead prediction of NO_x emissions for a coal-based boiler. Applied Energy, 2013, 106(6): 89–99

[4] Xu M, Yuan J, Ding S, Simulation of the gas temperature deviation in large-scale tangential coal fired utility boilers. Computer Methods in Applied Mechanics and Engineering, 1998, 155(3–4): 369–380

[5] Vuthaluru H B, Vuthaluru R. Control of ash related problems in a large scale tangentially fired boiler using CFD modeling. Applied Energy, 2010, 87(4): 1418–1426

[6] Liu Y, Fan W, Li Y. Numerical investigation of air-staged combustion emphasizing char gasification and gas temperature deviation in a large-scale tangentially fired pulverized-coal boiler. Applied Energy, 2016, 177(9): 323–334

[7] Guo J, Liu Z, Hu F, A compatible configuration strategy for burner streams in a 200 MW_e tangentially fired oxy-fuel combustion boiler. Applied Energy, 2018, 220(6): 59–69

[8] Habib M A, Ben-Mansour R, Abualhamayel H I. Thermal and emission characteristics in a tangentially fired boiler model furnace. International Journal of Energy Research, 2010, 34(13): 1164–1182

[9] Tang G, Wu B, Johnson K, Numerical study of a tangentially fired boiler for reducing steam tube overheating. Applied Thermal Engineering, 2016, 102(6): 261–271

[10] Khatami R, Stivers C, Joshi K, Combustion behavior of single particles from three different coal ranks and from sugar cane bagasse in O₂/N₂ and O₂/CO₂ atmospheres. Combustion and Flame, 2012, 159(3): 1253–1271

[11] Riaza J, Khatami R, Levendis Y A, Single particle ignition and combustion of anthracite, semi-anthracite and bituminous coals in air and simulated oxy-fuel conditions. Combustion and Flame, 2014, 161(4): 1096–1108

[12] Khatami R, Levendis Y A. An overview of coal rank influence on ignition and combustion phenomena at the particle level. Combustion and Flame, 2016, 164(2): 22–34

[13] Zhu M, Zhang H, Tang G, Ignition of single coal particle in a hot furnace under normal and micro-gravity condition. Proceedings of the Combustion Institute, 2009, 32(2): 2029–2035

[14] Su S, Pohl J H, Holcombe D, Techniques to determine ignition, flame stability and burnout of blended coals in p.f. power station boilers. Progress in Energy and Combustion Science, 2001, 27(1): 75–98

[15] Clements B R, Zhuang Q, Pomalis R, Ignition characteristics of co-fired mixtures of petroleum coke and bituminous coal in a pilot-scale furnace. Fuel, 2012, 97(2): 315–320

[16] Zeng G, Sun S, Dong H, Effects of combustion conditions on formation characteristics of particulate matter from pulverized coal bias ignition. Energy & Fuels, 2016, 30(10): 8691–8700

[17] Zeng G, Sun S, Yang X, Effect of the primary air velocity on ignition characteristics of bias pulverized coal jets. Energy & Fuels, 2017, 31(3): 3182–3195

[18] Zeng G, Sun S, Zhang W, Effects of bias concentration ratio on ignition characteristics of parallel bias pulverized coal jets. Energy & Fuels, 2017, 31(12): 14219–14227

[19] Zhao Y, Zeng G, Zhang L, Effects of fuel properties on ignition characteristics of parallel-bias pulverized-coal jets. Energy & Fuels, 2017, 31(11): 12804–12814

[20] Qin Y, Zhang Z, Wu S, Experimental study of jet characteristics of a rectangular nozzle with different height-to-width ratios in a tangentially-fired combustion flow field. Journal of Engineering for Thermal Energy and Power, 2000, 15(2): 125–127 (in Chinese)

[21] Zhang J, Kelly K E, Eddings E G, Ignition in 40 kW co-axial turbulent diffusion oxy-coal jet flames. Proceedings of the Combustion Institute, 2011, 33(2): 3375–3382

[22] Binner E, Zhang L, Li C, Bhattacharya S. In-situ observation of the combustion of air-dried and wet Victorian brown coal. Proceedings of the Combustion Institute, 2011, 33(2): 1739–1746

[23] Molina A, Murphy J J, Winter F, Pathways for conversion of char nitrogen to nitric oxide during pulverized coal combustion. Combustion and Flame, 2009, 156(3): 574–587

[24] Yan Y, Sun L, Peng Z, Effects of pyrolyzed semi-char blend ratio on coal combustion and pollution emission in a 0.35 MW pulverized coal-fired furnace. Frontiers in Energy, 2020, online, doi: 10.1007/s11708-020-0678-z

[25] Rastko J, Aleksandra M, Bartosz S. Sensitivity analysis of different devolatilization models on predicting ignition point position during pulverized coal combustion in O₂/N₂ and O₂/CO₂ atmospheres. Fuel, 2012, (11): 23–37

[26] Ti S, Chen Z, Li Z, Influence of primary air cone length on combustion characteristics and NO_x emissions of a swirl burner from a 0.5 MW pulverized coal-fired furnace with air staging. Applied Energy, 2018, 211(2): 1179–1189

[27] Qiao Y, Zhang L, Binner E, An investigation of the causes of the difference in coal particle ignition temperature between combustion in air and in O₂/CO₂. Fuel, 2010, 89(11): 3381–3387

[28] Sung Y, Moon C, Eom S, Coal-particle size effects on NO reduction and burnout characteristics with air-staged combustion in a pulverized coal-fired furnace. Fuel, 2016, 182(10): 558–567

[29] Chen Z, Li Z, Jing J, Gas/particle flow characteristics of two swirl burners. Energy Conversion and Management, 2009, 50(5): 1180–1191

[30] Jiang X, Zheng C, Yan C, Physical structure and combustion properties of super fine pulverized coal particle. Fuel, 2002, 81(4): 793–797

[31] Li Z, Liu Y, Chen Z, Effect of the air temperature on combustion characteristics and NO_x emissions from a 0.5 MW pulverized coal-fired furnace with deep air staging. Energy & Fuels, 2012, 26(4): 2068–2074

[32] Pedel J, Thornock J N, Smith P J. Large eddy simulation of pulverized coal jet flame ignition using the direct quadrature method of moments. Energy & Fuels, 2012, 26(11): 6686–6694

[33] Yamamoto K, Murota T, Okazaki T, Large eddy simulation of a pulverized coal jet flame ignited by a preheated gas flow. Proceedings of the Combustion Institute, 2011, 33(2): 1771–1778

[34] Zhu Z, Chi Z, Pan W, Aerodynamic experimental research in corner-fired furnace with burners of different height-width ratios. Journal of Shanghai Institute of Electric Power, 1994, 10(1): 20–27 (in Chinese)

[35] Wang C, Hao Y, Wang R, Numerical simulation of influence of depth-width ratio for boiler burner on combustion. Gas & Heat. 2010, 30(12): 11–15 (in Chinese)

[36] Moon C, Sung Y, Eom S, NO_x emissions and burnout characteristics of bituminous coal, lignite, and their blends in a pulverized coal-fired furnace. Experimental Thermal and Fluid Science, 2015, 62(1): 99–108

[37] Glarborg P, Kristensen P, Dam-Johansen K, Nitric oxide reburning by non-hydrocarbon fuels: implications for reburning with gasification gases. Energy & Fuels, 2000, 14(4): 828–838

[38] Gou X, Zhou J, Zhou Z, The influence of water vapor upon formation of NO in combustion of pulverized coal. Thermal Power Generation, 2007, 10(1): 4–8 (in Chinese)

[39] Wang Z, Zhao Y, Sun R, Effects of reaction condition on NO emission characteristic, surface behavior and microstructure of demineralized char during O₂/H₂O combustion process. Fuel, 2019, 253(10): 1424–1435

[40] De Soete G, Croiset E, Richard J R. Heterogeneous formation of nitrous oxide from char-bound nitrogen. Combustion and Flame, 1999, 117(1–2): 140–154

RIGHTS & PERMISSIONS

Higher Education Press