Influence of Separated Overfire Air Ratio and Location on Combustion

Oct 27, 2015 - radiation heat transfer, P-1 model(14). heat absorption coefficient of combustion gases, weighted sum of gray gases (WSGG) method(15). ...
1 downloads 9 Views 7MB Size
Article pubs.acs.org/EF

Influence of Separated Overfire Air Ratio and Location on Combustion and NOx Emission Characteristics for a 600 MWe DownFired Utility Boiler with a Novel Combustion System Lun Ma,† Qingyan Fang,*,† Dangzhen Lv,‡ Cheng Zhang,† Gang Chen,*,† Yiping Chen,‡ and Xuenong Duan‡ †

State Key Laboratory of Coal Combustion, Huazhong University of Science and Technology, Wuhan, Hubei 430074, People’s Republic of China ‡ Electric Power Research Institute, Hunan Power Company, Changsha, Hunan 410000, People’s Republic of China ABSTRACT: To reduce NOx emissions without introducing an obvious increase in the carbon content of fly ash, a novel combustion system was applied to a 600 MWe Foster Wheeler (FW) down-fired boiler. This approach mainly consisted of moving fuel-lean nozzles from the arches to the front/rear walls, rearranging staged air, and introducing separated overfire air (SOFA). The aim of this work was to evaluate the overall performance of the novel combustion system relative to different SOFA ratios (i.e., 15, 20, 25, and 30%) and different SOFA locations in the upper furnace (1.0, 2.0, and 3.0 m above the arches) using numerical simulations and experimental measurements. Both numerical and experimental results showed that, with increasing the SOFA ratio from 15 to 20%, NOx emissions were greatly reduced but the carbon content in the fly ash increased slightly. With a further increase from 20 to 30%, NOx emissions slightly decreased but the carbon content in the fly ash increased substantially. Considering both the environmental and economic effects, 20% was chosen as the optimal SOFA ratio. With increasing the SOFA location height in the upper furnace from 1.0 to 3.0 m above the arches, the average gas temperature after the superheater and the carbon content in the fly ash at the furnace outlet somewhat increased but NOx emissions decreased. Considering various factors, the location (2.0 m above the arches) was chosen as the optimal SOFA location. The performance of this boiler in actual operation was good after performing modifications with the optimal SOFA ratio and location.

1. INTRODUCTION Nitrogen oxides are major environmental problems, and coal combustion is the main resource of anthropogenic NOx, which is an extremely toxic pollutant involved in the generation of acid rain and photochemical smog.1,2 A large proportion of nitrogen oxides come from coal-fired power plants; thus, controlling and reducing NOx emissions from utility boilers are very significant. Coal with low volatility and poor reactivity (such as anthracite and lean coal) is abundant, and down-fired boilers are commonly used to fire these coals because, in comparison to other types of boilers (such as tangentially or opposed fired utility boilers), this type of boiler has a hightemperature level in the lower furnace, long travel paths for pulverized-coal particles, and other advantages. However, this type of boiler has high NOx emissions (approximately 1100− 1800 mg/m3, O2 at 6%, before the flue gas denitrification system).3−8 The latest “thermal power plant air pollutant emission standards” (issued by the Chinese Environmental Protection Department) in China limits NOx emissions below 200 mg/m3 (after the flue gas denitrification system) for downfired boilers. For many existing down-fired boilers, adopting low-NOx modification is necessary to minimize NOx emissions. To reduce and control NOx emissions as much as possible, many technical schemes have been used for the actual modification of down-fired boilers, such as redesigning the burners, staging the air/fuel, and applying overfire air (OFA). Overall, the main purpose of these technologies is to decay the mixing between the fuel and oxygen in the combustion air to © 2015 American Chemical Society

create a fuel-rich lean-oxygen atmosphere, in which NOx can be reduced. In general, of these technical schemes, the simplest, low-cost, and highest efficiency technique is the introduction of OFA. However, the introduction of OFA also changes the dominant combustion condition and may elicit the problem of high burnout loss, which is economically disadvantageous for power plants. Four main types of down-fired utility boilers have been commonly applied to power plants: the Foster Wheeler (FW), the Mitsui Babcock Energy Limited (MBEL), the Babcock & Wilcox (B&W), and the Stein down-fired boilers. Of these boilers, FW down-fired boilers represent a large proportion of boilers in China. For this type of boiler, some low NOx modifications have been reported, i.e., “fuel preheat nozzle”, “vent-to-OFA technology”, and “combined high-efficiency and low-NOx technology” (CHELNO).9−11 The latter two techniques reduced NOx emissions considerably by introducing OFA but brought the obvious increase of carbon content in the fly ash. In this work, a novel combustion system was applied to a 600 MWe FW down-fired boiler to solve high NOx emissions without introducing an obvious increase in the carbon content of the fly ash. This system included moving fuel-lean nozzles from the arches to the front/rear walls, rearranging staged air, and introducing separated overfire air (SOFA). The aim of this Received: July 10, 2015 Revised: October 27, 2015 Published: October 27, 2015 7630

DOI: 10.1021/acs.energyfuels.5b01569 Energy Fuels 2015, 29, 7630−7640

Article

Energy & Fuels

Figure 1. Schematics of the down-fired boiler and the original combustion system (mm).

Figure 2. Schematics of the novel combustion system of the down-fired boiler. divided into fuel-rich and fuel-lean coal/air streaming. Originally, the fuel-rich and fuel-lean coal/air mixtures are symmetrically channeled through nozzles on the arches. Secondary air is divided into seven layers of air, labeled A, B, C, D, E, F1, and F2. The A-, B-, and C-layer secondary air are ejected through the arches, and the remaining secondary air are supplied from the front and rear walls below the arches. Apart from C-layer secondary air nozzles, other secondary air nozzles are symmetrically arranged. 2.2. Novel Combustion System. As illustrated in panels a and b of Figure 2, the following steps are included for the novel combustion system: (1) The original 72 fuel-lean coal/air nozzles are merged into 36 nozzles, and their locations are moved from the arches to the front/ rear walls. The fuel-lean/air (or named as vent air) stream is fed into the furnace at a declination angle of approximately 30°. (2) At the original fuel-lean location (centered over the furnace on the arches), A-layer secondary air is injected through circular nozzles instead of through the original annular nozzles. (3) D- and E-layer secondary air are supplied through annular ports around the fuel-lean nozzles to cool the burner nozzles (named as D-layer secondary air). (4) The F1- and

paper was to evaluate the overall performance of the novel combustion system under different SOFA ratios and SOFA locations in the upper furnace for the down-fired boiler. Numerical simulations of the combustion and NOx emission characteristics in the furnace were conducted under four SOFA ratios (i.e., 15, 20, 25, and 30%) and three SOFA location settings in the upper furnace (1.0, 2.0, and 3.0 m above the arches). Meanwhile, industrial measurements were also performed to investigate the influence of different SOFA ratios on combustion and NOx emissions.

2. METHODOLOGY 2.1. Utility Boiler and Original Combustion System. Figure 1 shows the furnace schematics and the original combustion system. The width of the furnace is 34.480 m, and other dimensions are shown in Figure 1. There are 72 burners with double cyclones positioned symmetrically on the arches, and the primary air/coal mixture is 7631

DOI: 10.1021/acs.energyfuels.5b01569 Energy Fuels 2015, 29, 7630−7640

Article

Energy & Fuels F2-layer secondary air are merged into a new secondary air scheme (called the F-layer secondary air), and it is supplied at a declination angle of 20° through the novel nozzles instead of the original slit ports. (5) A part of the secondary air is separated from the secondary air box to be directed as SOFA to form a low-oxygen and reducing atmosphere in the lower furnace. The direct-flow SOFA nozzles (72 nozzles) are arranged symmetrically on the front and rear walls in the upper furnace zone at a declination angle of approximately 30°. Figure 2c shows the detailed principle schematic of the novel combustion system. The introduction of the fuel-lean/air stream below the arches and the novel staged air at a declination angle is in favor of (i) further increasing the residence time of pulverized-coal particles to obtain more reaction times and promoting fuel-rich/air stream ignition because of the fuel-lean/air stream carrying much of the moisture from the coal and (ii) extending the flame penetration depth and postponing the mixing of the coal/air stream with the secondary air to inhibit NOx formation. Positing SOFA in the upper furnace is to strengthen air-staging conditions and enlarge the reducing zone in the lower furnace region as far as possible. With this technology, the most important merit is to reduce NOx emissions without causing an obvious increase in the carbon content of the fly ash after introducing SOFA. 2.3. Numerical Simulation. Pulverized coal combustion is a series of complex physical and chemical reactions consisting of turbulent flow, combustion, and heat and mass transfer. In this work, a computational fluid dynamics (CFD) program, Fluent (version 6.3.26), was used to conduct the numerical simulations. The numerical methods and models related to this study are shown in Table 1. The

Figure 3. Geometry of the CFD and grid division. and gas temperature along the grid independence test line. The meshes with 4 270 000 and 3 820 000 cells presented similar results. Therefore, the mesh consisting of 3 820 000 cells was ultimately adopted for the simulations. The coal characteristics and operating parameters are given in Table 2. The total pulverized-coal mass rate is 93.4 kg/s (as received). A total of 90% of the total coal mass is supplied through the fuel-rich nozzles, while the remaining 10% is ejected through the fuel-lean nozzles. The dimensional distribution of the coal particles follows the Rosin− Rammler formula, with a size range from 5 to 250 μm, an average diameter of 54 μm, and a spread parameter of 1.05. The reaction kinetic data of devolatilization and char combustion (including the pre-exponential factor and active energy) are listed in Table 3. The total air mass flow is 648 kg/s, and the pressure at the furnace outlet is maintained in the level of about −70 Pa. The simulations are performed under four SOFA ratios, i.e., 15, 20, 25, and 30%, and three SOFA locations in the upper furnace, i.e., 1.0, 2.0, and 3.0 m above the arches (shown in Figure 2b). The details about the case settings are shown in Table 4. 2.4. In Situ Measurements. Full-scale experimental measurements of the boiler performance before and after the combustion modifications were performed under full load. During the experiments, coal samples (sampled from the different coal hopper exits) were mixed before the property tests. Table 2 lists the coal properties. The local gas temperatures near the burner region were measured using a fine-wire thermocouple (the measurement range of 773−2273 K) along the fuel-rich burner axis, and the gas temperatures near the rightside wall were also measured through the observing doors. The finewire thermocouple was kept in the furnace for less than 1 min to protect the thermocouple from serious ash deposit. At each point, the temperature is measured 5 times, and then the temperature data are averaged. A MSI EURO-type flue gas analyzer was employed with a grid-based method to measure the species concentrations in the flue gas. After dust removal and dehumidification, the tested gas enters into the sensor chamber and then enters into the cell via a permeable membrane. The tested gas is reacted with electrolyte, generating an electrical signal that is proportional to the gas concentration. The signal can be converted to the species concentrations. At each point, the species concentrations are measured 3 times, and then the data are also averaged. To verify simulation results, O2 and NOx concentrations in the flue gas were measured at the entrance of the air preheater to avoid the error from the in-leakage of the air preheater. Under different SOFA ratios (5, 15, 20, and 25%), O 2 and NO x concentrations were measured at the exit of the air preheater and the steam parameters and the de-superheating water rate were recorded. The carbon content in the fly and bottom ashes was tested using a proximate analyzer.

Table 1. Mathematical Methods and Models for This Numerical Simulation item

methods and models

gas-phase turbulent fluid flow coal particle motion radiation heat transfer heat absorption coefficient of combustion gases coal devolatilization char combustion gas-phase turbulent combustion fuel NOx thermal NOx

realizable k−ε method12 stochastic particle method13 P-1 model14 weighted sum of gray gases (WSGG) method15 two competing rates16 diffusion/kinetic models17 mixture fraction/PDF18 De Soete’s model19 extended Zeldovich mechanism20

two-competing-rate model includes the following: devolatilization 1 at a low temperature and devolatilization 2 at a high temperature. K v1

⎯ α1volatitle(g) + (1 − α1)char(s), and Devolatilization 1 is coal(s) ⎯→ K v2

⎯ α2volatitle(g) + (1 − α2)char(s). In devolatilization 2 is coal(s) ⎯→ this work, prompt NOx was not taken into account because of its low proportion in the pulverized furnace; thus, only fuel NOx and thermal NOx were taken into account. De Soete’s model was used to predict the formation of fuel NOx, and it was assumed that fuel N was distributed between the volatile and char. The mass fractions of volatile N and char N were 23.09 and 76.91% in this paper, respectively. Volatile N converted to intermediates, such as HCN and NH3 (HCN/NH3 = 9:1). Char N directly converted to NO (the conversion fraction was set at 0.5). More details regarding the application of these models can be found in refs 21−24. On the basis of the real physical structure, the geometric model of the furnace was divided into several sections for accurate representation. Different sections were meshed separately with hexahedral cells and coupled by interfaces to improve the quality of the mesh (seen in Figure 3). There were more cells in the near burner zone because of the intense combustion, and the mesh line was consistent with the airflow direction in the furnace. The grid independence test was conducted under the cell numbers of 2 800 000, 3 260 000, 3 820 000, and 4 270 000, respectively. Figure 4 shows the grid independence test based on the gas vertical velocity 7632

DOI: 10.1021/acs.energyfuels.5b01569 Energy Fuels 2015, 29, 7630−7640

Article

Energy & Fuels

Figure 4. Grid independence test: (a) gas vertical velocity and (b) gas temperature along the grid independence test line.

Table 2. Proximate and Ultimate Analyses of Coal and the Operation Parameters coal characteristics ultimate analysis (wt %, air dried)a

proximate analysis (wt %, air dried) volatile matter

moisture

ash

fixed carbon

C

8.27

2.00

39.80

49.93

51.56

item

O

S

N

lower heating value (kJ/kg, as received)

1.24

0.68

17370

original combustion system

novel combustion system

air ratio (%)

velocity (m/s)

temperature (K)

air ratio (%)

velocity (m/s)

temperature (K)

11.11 11.11 10.00 10.00 15.00 4.00 5.00 5.00 23.78

17.98 15.75 31.05 30.07 29.89 5.41 5.26 5.92 8.34

393 393 594 594 594 594 594 594 594

11.11 11.11 10.00 10.00 15.00 7.78

18.44 22.81 33.35 32.50 32.30 32.10

403 403 642 642 642 642

10.00 20.00 5.00

42.10 42.84

642 642 300

fuel-rich flow fuel-lean flow A-layer secondary air B-layer secondary air C-layer secondary air D-layer secondary air E-layer secondary air F1-layer secondary air F2-layer secondary air SOFA leaking air a

H

2.26 2.46 operation parameters

5.00

300

C, carbon; H, hydrogen; O, oxygen; N, nitrogen; and S, sulfur.

3. RESULTS AND DISCUSSION 3.1. Validation of the Numerical Results. To validate the simulations, the comparison between the measured and calculated gas temperatures is carried out (after the modification, the SOFA ratio is 20% and the SOFA location is 2.0 m above the arches). Figure 5a shows the trend of temperature changes along the axis of one burner, and Figure 5b compares the measured and calculated gas temperatures for the different observation doors (ports 1, 2, 3, and 4) after using the novel combustion system. The overall trend of gas temperatures achieves well consistency between the measured and calculated results, except for some deviation. This may be because any variation of in-furnace operating parameters will strongly influence the combustion or some ash may deposit on the thermocouple, although the device was tried to be kept clean. The calculated and measured carbon contents in the fly ash and O2 and NOx concentrations in the flue gas are summarized in Table 5. These predicted values are also consistent with the measured values. Therefore, the mesh and models adopted in this work are reasonable for investigating the characteristics of flow, combustion, and NOx emissions for the studied boiler.

Table 3. Reaction Kinetic Data of Devolatilization and Char Combustion24,25 A (pre-exponential factor)

reaction devolatilization 1 (α1 = 0.3) devolatilization 2 (α2 = 1) char combustion

E (activation energy)

3.75 × 105 s−1

7.366 × 104 J mol−1

1.46 × 1013 s−1

2.511 × 105 J mol−1

0.0016 kg m−2 s−1 Pa−1

83700 J mol−1

Table 4. Case Settings for the Simulations case settings case case case case case case

1 2 3 4 5 6

SOFA ratio (%)

stoichiometric ratio in the lower furnace

15 20 25 30 20 20

0.95 0.88 0.83 0.77 0.88 0.88

location of SOFA ports location location location location location location

2 2 2 2 1 3

2.0 2.0 2.0 2.0 1.0 3.0

m m m m m m

above above above above above above

the the the the the the

arches arches arches arches arches arches

7633

DOI: 10.1021/acs.energyfuels.5b01569 Energy Fuels 2015, 29, 7630−7640

Article

Energy & Fuels

Figure 5. Comparison of the experimental and calculated gas temperatures.

trajectories of the coal particles from the arches, and most of the particles follow the airflow travel path before flowing downward into the lower part of the lower furnace and redirect upward. The typical temperature distribution at the crosssection of one fuel-rich nozzle is presented in Figure 6c. After the fuel-rich and fuel-lean air mixtures enter into the furnace, the volatile in the pulverized coal releases and ignites quickly to release a large amount of heat. Thus, two high-temperature zones are shown in the lower furnace: one near the fuel-rich nozzle and the other near the fuel-lean nozzle. Because coal combusts continually and releases significant heat, a relatively high-temperature zone in the central furnace zone is formed. With an increasing furnace height, the temperature begins to decrease gradually for the coal burnout and heat transfer by radiation and convection between the flue gas and water walls. Panels d and e of Figure 6 display the O2 and NO concentration distributions, respectively. Because of the intense coal combustion, O2 is consumed rapidly and NO is produced in the region not far from the burner outlet. With a further increase in the furnace height, the NO concentration decreases because a greater amount of the present NO in the gas is gradually reduced into N2 by the unburned char and by HCN, NH3, and other intermediate products in the flue gas.25 3.3. Combustion and NOx Emission Characteristics with Different SOFA Ratios. Figures 7−9 present the flow field, temperature distribution, and NO concentration distribution of one fuel-rich nozzle cross-section under different SOFA ratios of 15% (case 1), 20% (case 2), 25% (case 3), and 30% (case 4). As shown in Figure 7, the penetration level of SOFA airflow toward the furnace center becomes much deeper as the SOFA ratio increases. This change occurs because more SOFA results in a larger moment to make the decay velocity of SOFA lower. At the same time, the recirculation zones below the arches decrease slightly as the SOFA ratio increases. This difference can be explained by the reduced arch secondary airflow weakening the rigidity, leading to a shallower penetration depth of arch airflow. As shown in Figure 8, temperature distributions display similar characteristics under different SOFA ratios. In the region near the fuel-rich/lean nozzles, high-temperature zones form because of the rapid ignition of coal volatile and intense combustion of char. The NO concentration distributions under different SOFA ratios are shown in Figure 9. Similar to the temperature distribution, two high NO concentration zones are presented in the lower furnace: one is below the arches but not far from the fuel-rich burner outlet, and the other is located near the fuel-lean burner

Table 5. Calculated and Measured Performances of the Boiler with the Original and Novel Combustion Systems 600 MW (original)

600 MW (novel, SOFA of 20% and 2.0 m above the arches)

measured calculateda measured calculated

4.13 3.96 2.25 2.22

4.30 4.05 2.31 2.30

measured calculated

1501 1351

751 729

measured measured

45 400.2

76 395.4

measured

5.62

4.73

measured

1875

1905

measured

818

817

816

815

91.05

91.31

item carbon content in the fly ash (%) O2 content (vol %, at the furnace outlet, excluding in-leakage in the air preheater) NOx (mg/m3, 6% O2, before the flue gas denitrification system) CO (ppm) exhaust gas temperature (K) carbon content in the bottom slag (%) flow rate of main steam (tons/h) main steam temperature (K) reheat steam temperature (K) boiler efficiency (%)

measured measured

b

The carbon content in the fly ash by CFD was obtained using the following equation: CCFA = ((1 − B)FC)/((1 − B)FC + A), where CCFA is the carbon content in the fly ash, B is the char burnout ratio from the statistical results of CFD at the furnace outlet, FC is the char content of coal (dried, at a high-temperature produced ratio), and A is the ash content of coal (dried, at a high-temperature produced ratio). b The boiler efficiency is calculated by the heat-loss method: η = 100 − (q2 + q3 + q4 + q5 + q6), where η is the boiler efficiency (%), q2 is the flue gas loss (%), q3 is the heat loss as a result of unburned gas (%), q4 is the loss as a result of unburned carbon (%), q5 is the radiation loss (%), and q6 is the heat loss of ash and slag (%). a

3.2. Flow, Combustion, and NOx Emission Characteristics in the Furnace. Figure 6a shows the typical predicted flow field (after the modification, SOFA of 20% and SOFA location at 2.0 m above the arches). The results indicate that the airflows from the front and rear arches converge at the center of the furnace and two obvious recirculation zones form below the arches. The F-layer secondary air lifts the arch airflows to prevent the flame from rushing directly into the ash hopper. SOFA injects into the furnace with high speed and strong rigidity, and then it redirects its flow vertically upward after the collision with upward flue gas. Figure 6b shows the 7634

DOI: 10.1021/acs.energyfuels.5b01569 Energy Fuels 2015, 29, 7630−7640

Article

Energy & Fuels

Figure 6. Typical predicted results of one fuel-rich nozzle cross-section (after the modification): (a) flow field, (b) trajectories of coal particles from fuel-rich nozzles, (c) temperature distribution, (d) O2 concentration distribution, and (e) NO concentration distribution.

tendency for cases 1−4. Initially, the average temperature increases rapidly with an increasing furnace height and reaches a relative maximum at a height of approximately 15.0 m. Then, the average temperature gradually decreases along the furnace height. The gas temperature curves show four “valleys” at approximately 13.0, 16.0, 19.0, and 23.0 m, respectively. The first valley is caused by the amount of low-temperature F-layer secondary air that is sprayed into the furnace (“FSA” in Figure 10a), and the vent air and D-secondary air (“VA + DSA” in Figure 10a) result in the secondary valley below the arches. The height of the third valley is associated with the position of arch airflow (“arch air” in Figure 10a), and the fourth valley results from the large amount of SOFA injected into the upper furnace region (“SOFA” in Figure 10a). With increasing SOFA ratios, more SOFA is fed into the upper furnace to cool the upward gas, so that the gas temperature in the upper furnace (near the SOFA ports) continues to decrease. However, the gas temperature level enhances at the upper part of the upper furnace with increasing the SOFA ratio from 15 to 30%. This can be explained by the combustion level of pulverized-coal particles in the lower furnace decreasing, and consequently, the amounts of char continually combust and burn out at the upper part of the upper furnace. Figure 10b shows the change trend of the average O2 concentration along the furnace height under the different SOFA ratios of 15, 20, 25, and 30% (cases 1−4). In general, the

Figure 7. Flow fields under different SOFA ratios.

outlet. This behavior may be explained by, in these zones, the atmosphere with high temperature and relatively high O2 concentration favoring production of a large amount of fuel NOx and thermal NOx. Figure 10a shows the change trend of the average gas temperature along the furnace height under different SOFA ratios. The gas temperature curves present similar change

Figure 8. Temperature distributions under different SOFA ratios. 7635

DOI: 10.1021/acs.energyfuels.5b01569 Energy Fuels 2015, 29, 7630−7640

Article

Energy & Fuels

Figure 9. NO concentration distributions under different SOFA ratios.

consumed as coal combustion occurs in the lower furnace region and the average O2 concentration gradually decreases because less secondary air is sprayed into the lower furnace. In the upper furnace region, as a result of more SOFA injected into the furnace, the O2 concentration generally increases as the SOFA ratio increases. The O2 concentration in the lower furnace is too low to sufficiently provide intense burning of the pulverized coal. As a result, the burnout will be reduced, which affects the economic operation of the boiler. The O 2 concentration in the lower furnace is too large to restrain the NO formation in the lower furnace. Therefore, the selected SOFA ratio should be neither too large nor too small. Figure 10c shows the change trend of the average CO concentration along the furnace height with different SOFA ratios of 15, 20, 25, and 30%. The variations of the CO levels generally follow a trend in which low O2 concentrations are associated with high CO concentrations in the lower furnace. The low air stoichiometry in the lower furnace forces combustion to occur in a lean-oxygen atmosphere in the lower furnace, resulting in high CO production. In the upper furnace, SOFA provides sufficient O2 to supply the continual combustion of char and CO. Thus, the CO concentration curves gradually fall as the furnace height increases. With increasing SOFA ratios, a higher CO concentration develops in the lower furnace, which makes the reducing atmosphere stronger in the lower furnace. This behavior can be explained by, after a certain amount of secondary air is removed from the lower furnace to the upper furnace, oxygen-deficient combustion becoming stronger in the lower furnace to produce more CO in this region. Particularly, in the most intense combustion region ranging from 14.0 to 17.0 m, the difference in the CO concentration is more obvious for cases 1−4. In the upper furnace, the CO concentration level becomes lower as the SOFA ratio increases because of more SOFA making CO combustion more complete. Figure 10d shows the change trend of the average NO concentration along the furnace height under different SOFA ratios of 15, 20, 25, and 30%. A large proportion of NO is produced in the intense combustion zone from 14.0 to 17.0 m, and with an increasing furnace height, the NO concentration curves fall gradually because the NO present in the gas is reduced. The gas atmosphere and coal combustion process control NO production, and the NO concentration is largely

Figure 10. Different average parameters along the furnace height under different SOFA ratios (SOFA, separated overfire air; arch air, including all air on the arches; VA + DSA, vent air and D-layer secondary air; and FSA, F-layer secondary air).

curve trend presents a similar change law for the four settings. In the lower furnace region, the intense combustion of pulverized coal consumes a large amount of O2, especially from 14.0 to 17.0 m. As the height of the furnace increases further, the char/CO mixtures combust and the O 2 concentration continuously decreases in the upper furnace. With increasing SOFA ratios, the curves show that more O2 is 7636

DOI: 10.1021/acs.energyfuels.5b01569 Energy Fuels 2015, 29, 7630−7640

Article

Energy & Fuels

significant reduction in NOx emissions can be achieved by adopting SOFA technology. With a further increase of the SOFA ratio from 20 to 25%, NOx emissions decrease at a smaller rate from 751 to 709 mg/m3 (O2 at 6%, before the flue gas denitrification system). However, the other factors change at a clearly enhanced rate: the O2 concentration increases from 3.42 to 3.71%, and the carbon content in the fly ash increases from 4.30 to 4.88%. Therefore, when the measured and simulated results are combined and NOx emissions and the carbon content in the fly ash are considered, a SOFA ratio of about 20% is advisible. Experimental results of the steam temperature and desuperheating water mass flow rate of the superheater under different SOFA ratios are shown in Figure 12. The designed

influenced by the air-staging level. As the SOFA ratio increases, the average NO concentration decreases. This result occurs for the following reasons: (i) the lower O2 concentration reduces the oxidizing atmosphere in the lower furnace, restricting NO production, and (ii) the high-concentration CO reduces NO present in the gas, so that more of the generated NO is restored to N2. A larger SOFA ratio was observed to correspond with less NO production. Therefore, to minimize NOx emissions, the SOFA ratio should not be too small. The calculated results of the parameters at the furnace outlet are given in Table 6 under different SOFA ratios (i.e., 15, 20, Table 6. Calculated Results of the Parameters at the Furnace Outlet under Different SOFA Ratios item

case 1

case 2

case 3

case 4

SOFA ratio (%) O2 (%) NOx (mg/m3, 6% O2, before the flue gas denitrification system) CO (ppm) carbon content in the fly ash (%) average temperature (K, after the superheater)

15 2.23 980

20 2.30 729

25 2.54 699

30 2.57 682

284 4.02 1321

322 4.05 1339

415 4.46 1342

483 5.23 1351

25, and 30%). Upon increasing the SOFA ratio from 15 to 20%, the O2 concentration increases from 2.23 to 2.30%, the carbon content in the fly ash increases slightly from 4.02 to 4.05%, and NOx emissions decrease considerably from 980 to 729 mg/m3 (O2 at 6%, at the furnace outlet). When the SOFA ratio is increased further from 20 to 30%, NOx emissions decrease slightly from 729 to 682 mg/m3 (O2 at 6%, at the furnace outlet), the O2 concentration increases from 2.30 to 2.57%, and the carbon content in the fly ash increases substantially from 4.05 to 5.23%. Figure 11 shows the experimental results of the carbon content in the fly ash and O2 and NOx emissions under

Figure 12. Experimental results of the steam temperature and desuperheating water mass flow rate of the superheater under different SOFA ratios.

values of main and reheat steam temperatures are both 814 K. With an increasing SOFA ratio from 5 to 20%, the desuperheating water mass flow rate of the superheater increases from 1.5 to 27 tons/h, the main steam temperature increases from 806 to 816 K, and the reheat steam temperature increases from 806 to 814 K. With a further increase of the SOFA ratio from 20 to 25%, the de-superheating water mass flow rate of the superheater increases from 27 to 85 tons/h, the main steam temperature increases from 816 to 817 K, and the reheat steam temperature increases from 814 to 815 K. When the SOFA ratio is about 20%, the de-superheating water mass flow rate of the superheater is at the low level (below 30 tons/h) and the main/reheat steam temperatures achieve the designed values, which also indicate that a SOFA ratio of 20% is optimal. 3.4. Combustion and NOx Emission Characteristics under Different SOFA Locations in the Upper Furnace. Figures 13−15 show the velocity field, temperature distribution, and NOx concentration distribution, respectively, under three different SOFA locations (case 5, 1.0 m above the arches; case 2, 2.0 m above the arches; and case 6, 3.0 m above the arches) in the upper furnace. The SOFA location impacts the mixing between SOFA and upward flowing hot gas. Figure 13 shows that relatively symmetric flow field form for the three locations. On one hand, the prior mixing between SOFA and upward flowing hot gas brings the unburned char into contact with SOFA ahead of time to obtain more time for the combustion of char in the upper furnace. Therefore, the lower location favors coal burnout. On the other hand, it should be realized that the lower location shrinks the reducing zone size in the lower

Figure 11. Experimental results of the carbon content in the fly ash, O2 concentration (after the air preheater, including in-leakage of the air preheater), and NOx emissions under different SOFA ratios.

different SOFA ratios (i.e., 5, 15, 20, and 25%). Upon increasing the SOFA ratio from 5 to 20%, the O2 concentration (after the preheater) increases from 3.05 to 3.42% and the change in the carbon content in the fly ash is small (only 4.17 to 4.30%). However, NOx emissions decrease from 1275 to 751 mg/m3 (O2 at 6%, before the flue gas denitrification system), a decrease of approximately 41%. This result indicates that a 7637

DOI: 10.1021/acs.energyfuels.5b01569 Energy Fuels 2015, 29, 7630−7640

Article

Energy & Fuels

to location 3 (case 6, 3.0 m above the arches), the difference in the high-temperature flame is not obvious. Figure 15 displays the similar NO concentration distribution. A large amount of NO is produced in the region not far from the burner outlet. The NOx concentrations gradually decrease as the furnace height increases because the present NO is gradually reduced. Figure 16 shows the comparison of the calculated data (average gas temperature and O2, CO, and NO concentrations)

Figure 13. Flow fields under different SOFA locations.

Figure 14. Temperature distributions under different SOFA locations.

Figure 16. Different parameters along the furnace height under different SOFA locations (SOFA, separated overfire air; arch air, including all air on the arches; VA + DSA, vent air and D-layer secondary air; and FSA, F-layer secondary air).

along the furnace height under different SOFA locations (case 5, 1.0 m above the arches; case 2, 2.0 m above the arches; and case 6, 3.0 m above the arches). As shown in Figure 16a, the difference in the average temperature is not obvious in the lower furnace. As the SOFA location height increases, the gas temperature in the upper furnace (about 22.0−34.0 m) decreases because SOFA is fed into this zone to cool the hot flue gas. However, according to the statistical result (as listed in Table 7), the gas temperature after the superheater increases to a certain degree with enhancing the SOFA location. This is because the higher location of SOFA delays the mixing of upward hot gas with SOFA, which results in the release of more heat from char combustion at the upper part of the upper furnace. When SOFA is set at 2.0 m above the arches, the experimental results (as shown in Figure 12 of section 3.3, with SOFA of 20%) show that the de-superheating water mass flow rate of the superheater is about 27 tons/h and main steam and reheat steam temperatures are 816 and 814 K, respectively.

Figure 15. NO concentration distributions under different SOFA locations.

furnace, which does not favor reducing the generated NO into N2. Thus, considering the degree of coal burnout and NOx emissions, the SOFA location should be set at an appropriate location, i.e., case 2 at 2.0 m above the arches. As shown in Figure 14, the temperature distribution displays similar behaviors for the three SOFA locations. However, some difference also exists. From location 1 (case 5, 1.0 m above the arches) to location 2 (case 2, 2.0 m above the arches), the hightemperature flame (approximately 1800−1900 K) moves upward, and from location 2 (case 2, 2.0 m above the arches) 7638

DOI: 10.1021/acs.energyfuels.5b01569 Energy Fuels 2015, 29, 7630−7640

Article

Energy & Fuels

There are smaller fluctuations of the in-furnace negative pressures under different cases for a long period of operation, which reveals more stable combustion. It can be seen that NOx emissions decrease by 50%, from 1501 to 751 mg/m3 (O2 at 6%, before the flue gas denitrification system), while the carbon content in the fly ash increases only slightly from 4.13 to 4.30%. The overall evaluation of the combustion and NOx emissions show that the performance of the studied boiler is successfully improved at the optimal SOFA ratio and location of the novel combustion system.

Table 7. Calculated Results of the Parameters at the Furnace Outlet under Different SOFA Locations

item

O2 (%)

NOx emission (mg/m3, 6% O2, before the flue gas denitrification system)

case 5 case 2 case 6

2.26 2.30 2.44

795 729 713

CO concentration (ppm)

carbon content in the fly ash (%)

average temperature (K, after the superheater)

247 322 375

3.98 4.05 4.69

1332 1339 1348

4. CONCLUSION Numerical simulations and industrial measurements were performed on a 600 MWe FW down-fired boiler to evaluate the overall performance of the novel combustion system relative to different SOFA ratios and different SOFA locations in the upper furnace. The following main conclusions can be drawn: (1) The NOx emissions and carbon content in the fly ash at the furnace outlet varied with different SOFA ratios. From 15 to 20%, NOx emissions were substantially reduced but the carbon content in the fly ash slightly increased. Upon further increasing SOFA from 20 to 30%, NOx emissions slightly increased but the carbon content in the fly ash substantially increased. Considering both the environmental and economic effects, 20% was chosen as the optimal SOFA ratio. (2) With increasing the SOFA location height from 1.0 to 3.0 m above the arches, the average gas temperature and the carbon content in the fly ash at the furnace outlet somewhat increased but NOx emissions decreased. Considering both the environmental and economic effects and the safety of the furnace, the SOFA location at 2.0 m above the arches was chosen as the optimal SOFA location. (3) After the modification with the optimal SOFA ratio and location, the boiler showed good performance that NOx emissions at full load decreased significantly by 50%, from 1501 to 751 mg/m3 (O2 at 6%, before the flue gas denitrification system), while the carbon content in the fly ash increased only slightly, from 4.13 to 4.30%.

This SOFA location makes the steam temperature achieve the designed value, meanwhile, avoiding the overheat of the platen superheater. Figure 16b shows the average O2 concentration along the furnace height under different SOFA locations. In the lower furnace region, the O2 curves show the same pattern and the differences between the different cases are not obvious. However, in the upper furnace region, the O2 concentration increases as the SOFA location height increases because the higher location of SOFA results in a lower degree of coal burnout and less O2 is consumed in the upper furnace region. The differences in the O2 concentration between cases 2 and 6 are much more obvious than those for cases 2 and 5. As the location moves downward, the CO concentration level in the upper region becomes low (shown in Figure 16c). The lower location of SOFA makes the reducing atmosphere region below the SOFA smaller, and CO mixes with the SOFA in advance. The SOFA location also has a significant influence on the NOx emissions. As demonstrated in Figure 16d, when the location of SOFA is higher, the NOx emission level is lower because the higher location of SOFA increases the size of the reducing atmosphere region below the SOFA; consequently, more of the present NO is reduced. Table 7 displays the calculation results for the parameters at the furnace outlet under different SOFA locations. The O2 and CO concentrations at the furnace outlet increase gradually as the SOFA location height increases from 1.0 to 3.0 m above the arches. With the increase of the SOFA location height from 1.0 to 2.0 m above the arches, NOx emissions decrease from 795 to 729 mg/m3 (O2 at 6%, at the furnace outlet), the carbon content in the fly ash increases from 3.98 to 4.05%, and the average temperature (after the superheater) increased somewhat from 1332 to 1339 K. With the further increase of the SOFA location height from 2.0 to 3.0 m above the arches, NOx emissions decrease from 729 to 713 mg/m3 (O2 at 6%, at the furnace outlet), the carbon content in the fly ash increases from 4.05 to 4.69%, and the average temperature (after the superheater) increased somewhat from 1339 to 1348 K. The higher SOFA location will make the de-superheating water mass flow rate of the superheater maintain the relatively high level. Considering both the carbon content in the fly ash and the NOx emissions, the optimal SOFA location should be location 2 at 2.0 m above the arches. This location not only provides a relatively larger reducing atmosphere zone to reduce NOx emissions but also favors the coal burnout degree. 3.5. Performance of the Boiler with the Optimal SOFA Ratio and Location. Table 5 summarizes the performance parameters before and after the modification. After the modification with the optimal SOFA ratio and location, the steam parameter values reach the designed values. The desuperheating water rate greatly decreases from 100 tons/h (before the modification) to 30 tons/h (after the modification).



AUTHOR INFORMATION

Corresponding Authors

*Telephone/Fax: +86-27-87540249. E-mail: [email protected]. cn. *Telephone/Fax: +86-27-87541908. E-mail: gangchen@hust. edu.cn. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This work was sponsored by the Project on the Foundation of State Key Laboratory of Coal Combustion, the Youth Foundation of Huazhong University of Science and Technology (2014QN185), and the National Natural Science Foundation of China (51390494).

■ 7639

NOMENCLATURE MBEL = Mitsui Babcock Energy Limited B&W = Babcock & Wilcox FW = Foster Wheeler CHELNO = combined high-efficiency and low-NO x technique OFA = overfire air DOI: 10.1021/acs.energyfuels.5b01569 Energy Fuels 2015, 29, 7630−7640

Article

Energy & Fuels SOFA = separated overfire air WSGG = weighted sum of gray gases arch air = all air on the arches VA + DSA = vent air and D-layer secondary air FSA = F-layer secondary air



REFERENCES

(1) Staiger, B.; Unterberger, S.; Berger, R.; Hein, K. R. G. Energy 2005, 30, 1429−1438. (2) Arenillas, A.; Rubiera, F.; Pis, J. J. Environ. Sci. Technol. 2002, 36, 5498−5503. (3) Eberle, J. S.; Garcia-Mallol, J. A.; Simmerman, R. N. Proceedings of the Pittsburgh Coal Conference; Pittsburgh, PA, Sept 23−27, 2002. (4) Blas, J. G. Combustion 1970, 42, 6−13. (5) Fan, S. B.; Li, Z. Q.; Yang, X. H.; Liu, G. K.; Chen, Z. C. Fuel 2010, 89, 1525−1533. (6) Wang, H. J.; Huang, Z. F.; Wang, D. D.; Luo, Z. X.; Sun, Y. P.; Fang, Q. Y.; Lou, C.; Zhou, H. C. Meas. Sci. Technol. 2009, 20, 4006− 4017. (7) Fang, Q. Y.; Wang, H. J.; Wei, Y.; Lei, L.; Duan, X. L.; Zhou, H. C. Fuel Process. Technol. 2010, 91, 88−96. (8) Fang, Q. Y.; Wang, H. J.; Zhou, H. C.; Lei, L.; Duan, X. L. Energy Fuels 2010, 24, 4857−4865. (9) Garcia-Mallol, J. A.; Steitz, T.; Chu, C. Y.; Jiang, P. Z. Proceeding of the 2nd U.S.−China NOx and SO2 Control Workshop; Dalian, China, Aug 2−5, 2005. (10) Yang, W. J.; Yang, W. C.; Zhou, Z. J.; Zhou, J. H.; Huang, Z. Y.; Liu, J. Z.; Cen, K. F. Fuel Process. Technol. 2014, 118, 90−97. (11) Li, Z. Q.; Ren, F.; Chen, Z. C.; Liu, G. K.; Xu, Z. X. Environ. Sci. Technol. 2010, 44, 3926−3931. (12) Shih, T. H.; Liou, W. W.; Shabbir, A.; Yang, Z.; Zhu, J. Comput. Fluids 1995, 24, 227−238. (13) Gosman, A. D.; Loannides, E. Proceedings of the AIAA 19th Aerospace Science Meeting; St. Louis, MO, Jan 12−15, 1981; pp 323− 333. (14) Cheng, P. AIAA J. 1964, 2, 1662−1664. (15) Smith, T. F.; Shen, Z. F.; Friedman, J. N. J. Heat Transfer 1982, 104, 602−608. (16) Kobayashi, H.; Howard, J. B.; Sarofim, A. F. Proceedings of the 16th International Symposium on Combustion; The Combustion Institute, Pittsburgh, PA, 1976. (17) Zhou, L. X. Theory and Numerical Modeling of Turbulent GasParticle Flows and Combustion; CRC Press, Inc.: Boca Raton, FL, 1993. (18) Sivathanu, Y.; Faeth, G. Combust. Flame 1990, 82, 211−230. (19) De Soete, G. G. Fifteenth Symposium (International) on Combustion; The Combustion Institute: Pittsburgh, PA, 1975; pp 991−1195. (20) Hill, S. C.; Smoot, L. D. Prog. Energy Combust. Sci. 2000, 26, 417−458. (21) Díez, L. I.; Cortés, C.; Pallarés, J. Fuel 2008, 87, 1259−1269. (22) Choi, C. R.; Kim, C. N. Fuel 2009, 88, 1720−1731. (23) Liu, H.; Liu, Y. H.; Yi, G. Z.; Nie, L.; Che, D. F. Energy Fuels 2013, 27, 5831−5840. (24) Kuang, M.; Li, Z. Q.; Xu, S. T.; Zhu, Q. Y. Environ. Sci. Technol. 2011, 45, 3803−3811. (25) Liu, G. K.; Chen, Z. C.; Li, Z. Q.; et al. Appl. Therm. Eng. 2015, 75, 1034−1045.

7640

DOI: 10.1021/acs.energyfuels.5b01569 Energy Fuels 2015, 29, 7630−7640