Numerical investigation on the flow, combustion and NOx emission characteristics in a 500 MWe tangentially fired pulverized-coal boiler

Fuel 88 (2009) 1720–1731

Contents lists available at ScienceDirect

Fuel journal homepage: www.elsevier.com/locate/fuel

Numerical investigation on the flow, combustion and NOx emission characteristics in a 500 MWe tangentially fired pulverized-coal boiler Choeng Ryul Choi a, Chang Nyung Kim b,c,*,1 a

Department of Mechanical Engineering, College of Advanced Technology, Kyung Hee University, Seocheon-dong, Giheung-gu, Yongin-si, Gyeonggi-do 446-701, Republic of Korea College of Advanced Technology, Kyung Hee University, Seocheon-dong, Giheung-gu, Yongin-si, Gyeonggi-do 446-701, Republic of Korea c Industrial Liaison Research Institute, Kyung Hee University, Seocheon-dong, Giheung-gu, Yongin-si, Gyeonggi-do 446-701, Republic of Korea b

a r t i c l e

i n f o

Article history: Received 25 June 2008 Received in revised form 24 March 2009 Accepted 1 April 2009 Available online 18 April 2009 Keywords: Tangentially fired pulverized-coal boiler Combustion simulation Nitrogen oxide (NOx) emissions

a b s t r a c t The characteristics of the flow, combustion, temperature and NOx emissions in a 500 MWe tangentially fired pulverized-coal boiler are numerically studied using comprehensive models, with emphasis on fuel and thermal NOx formations. The comparison between the measured values and predicted results shows good agreement, which implies that the adopted combustion and NOx formation models are suitable for correctly predicting characteristics of the boiler. The relations among the predicted temperature, O2 and CO2 mass fractions are discussed based on the calculated distributions. The predicted results clearly show that NOx formation within the boiler highly depends on the combustion processes as well as the temperature and species concentrations. The results obtained from this study have shown that overfire air (OFA) operation is an efficient way to reduce the NOx emissions of the pulverized-coal fired boiler. Air staging combustion technology (OFA operation) adopted in this boiler has helped reduce fuel NOx formation as well as thermal NOx formation under the present simulated conditions. The decrease in the formation of fuel NOx is due to the decreased contact of the nitrogen from the fuel with the oxygen within the combustion air, while the decrease in thermal NOx formation is caused by a decrease in temperature. The detailed results presented in this paper may enhance the understanding of complex flow patterns, combustion processes and NOx emissions in tangentially fired pulverized-coal boilers, and may also provide a useful basis for NOx reduction and control. Ó 2009 Elsevier Ltd. All rights reserved.

1. Introduction Tangentially fired pulverized-coal boilers are the most widely used type of boiler for industrial coal combustion because of their good flame distribution and uniform heat flux to the furnace walls, but still have some problems such as large amounts of combustible matter in fly ash, combustion instabilities at low loads, heat imbalance and gas temperature deviation in super-heaters and re-heaters, and slagging in the furnaces [1]. Moreover, although nitrogen oxide (NOx) emissions from the boilers are comparatively low, NOx emission has a significant impact on the environment. The control and reduction of NOx emissions from coal combustion has become an international concern, because NOx forms acid rain and is involved in the generation of photochemical smog. As a consequence, governments around the world and international organizations, which support limiting air pollution, have established

* Corresponding author. Address: College of Advanced Technology, Kyung Hee University, Seocheon-dong, Giheung-gu, Yongin-si, Gyeonggi-do 446-701, Republic of Korea. Tel.: +82 31 201 2578; fax: +82 31 202 9715. E-mail address: [email protected] (C.N. Kim). 1 Tel.: +82 31 201 2909; fax: +82 31 202 8106. 0016-2361/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.fuel.2009.04.001

restrictive legislation that suppresses the emission of NOx into the atmosphere [2]. Several technologies such as burner design modification, air/fuel staging, over fire air (OFA) operation, flue gas recirculation and reburning have been used to reduce and control NOx emissions. The main objective of these technologies is to minimize the reaction temperature and the contact between nitrogen from the fuel and oxygen in the combustion air, while creating a fuelrich zone in which NOx can be reduced to N2 [1,3–5]. The simplest technique is to introduce the OFA operation because this requires a simple air duct modification along with straightforward management of the air streams. Since the NOx emissions are strongly related to complex physical and chemical processes such as turbulent flow, combustion, temperature, heat transfer, and NOx formation mechanisms, understanding these complex processes is a prerequisite for reducing NOx emissions, as well as resolving the previously mentioned problems including large amounts of combustible matter in fly ash. Several numerical studies have been completed in order to better understand the complicated phenomena, including gas–particle flow, combustion, heat transfer and NOx formation in boilers [1,3–18]. These studies have been focused on the investigation of

C.R. Choi, C.N. Kim / Fuel 88 (2009) 1720–1731

causes and methods for decreasing gas flow and temperature deviation in the boilers [1,6–8]. Some studies have been carried out in order to predict NOx emissions from boilers under several operation conditions [3–5,13–16,18]. Their results show that the prediction of NOx emissions is affected by the used combustion and NOx formation models, and the amount of NOx formed is highly sensitive to the temperature and oxygen concentration distributions as well as fluid flow. Despite the remarkable progress shown in literature, there are still significant uncertainties surrounding the control and reduction of NOx that necessitates further research due to the complicated phenomena in furnaces involving complex physical and chemical processes that need to be precisely modeled [1]. There are very few studies that are focused on how NOx formation is affected by changes in operating conditions. Deep insight about the influence of operating conditions on NOx formation could provide useful guidelines for controlling NOx formation. In this study, the characteristics of the flow, combustion, temperature and NOx emissions in a 500 MWe tangentially fired pulverized-coal boiler are numerically studied with emphasis on fuel and thermal NOx formations. In order to generate accurate predictions, additional attention is given to selecting the calculation domain, generating mesh, and choosing numerical models, since NOx formation is affected by fluid flow, temperature and oxygen concentration distributions. The predicted results are compared with available measured data and operating data. A calculation is

1721

also carried out for the case with OFA operation, which allows some combustion air to be supplied through OFA ports, and is compared with the case without OFA operation in association with the reduction of NOx emissions. 2. The tangentially fired pulverized-coal boiler The tangentially fired pulverized-coal boiler considered in this study is a 500 MWe unit, and is shown schematically in Fig. 1. The height up to the furnace exit is approximately 51 m, and the horizontal cross section in the furnace has both a width and depth of 16.5 m. In order to make up a concentric firing system within the furnace, as shown in Fig. 1c, twelve burner sets are installed in the four corners ranging from the lower group A to the upper group C. Each burner set consists of two coal burners with different air to fuel ratios (a fuel rich burner (CONC) and a fuel lean burner (WEAK)) and two oil burners. The coal burners are of a low NOx burner type (pollution minimum (PM) type) and their injection angles can be controlled ±30°. It is known that this type of burners help an effective mixing of pulverized coal and air prior to injection. The oil burners are only used during start-up. Therefore, this boiler is operated with 24 low NOx burners during normal operations. Each burner set is treated as one module in the present study and the burner arrangement explained above is shown in Fig. 1b. OFA ports are installed above the burners.

Fig. 1. Schematic configurations of the tangentially fired pulverized-coal boiler.

1722

C.R. Choi, C.N. Kim / Fuel 88 (2009) 1720–1731

In the upper part of the boiler, there are two re-heaters (RH), three super-heaters (SH) and one economizer (Eco.) along the main flow direction. The furnace walls from the hopper inlet to the furnace exit are spiral water walls which minimize the number of tubes as well as provide a high mass flux. The other walls above the furnace are vertical water wall types.

3. Numerical analysis procedure 3.1. Calculation domain and mesh system The calculation domain includes the combustion zone, furnace, re-heaters, super-heaters, crossover pass, economizer and rear pass, as shown in Fig. 1. The re-heaters, super-heaters and economizer are treated as porous media, which allows the tube bundle0 s influence on flow and pressure losses to be considered. Therefore, detailed geometrical modeling for the heat exchangers is unnecessary. The mesh system consists of 957,402 cells. The combustion and furnace zones contain tetrahedral cells and all other zones contain hexahedral cells. Mesh is refined in the neighborhood of the burners where the combustion processes actively take place. Grid dependence tests have been completed with the current grid system as well as two additional grid systems of approximately 500,000 and 700,000 cells. The results have shown that the current grid system sufficiently provides grid-independent solutions.

the complex phenomena in large-scale tangentially-fired boilers. Detailed formulations of the models are not given here since they can be found in the FLUENT 6.3 User’s guide [29]. 3.2.1. NOx modeling After the combustion calculation, the NOx calculation is performed based on the predicted temperature and species concentrations under the assumption that the NOx concentration is very low and, therefore, has a negligible impact on the coal combustion. NOx is formed mainly by thermal NOx, fuel NOx and prompt NOx formation mechanisms. A transport equation for the NO species (i.e., NO, NO2, N2O) is calculated, which takes into account convection, diffusion, production and consumption of NO species. For fuel NOx sources, additional transport equations for intermediate species (HCN and NH3) are solved. Thermal NOx is formed when nitrogen and oxygen within the combustion air combine at a relatively high temperature in fuellean environments. The formation rate is primarily a function of temperature as well as the residence time of nitrogen at that temperature. In this study, the formation of thermal NOx is modeled by the extended Zeldovich mechanism as follows. kf ;1

O þ N2 ¢ N þ NO

ð1Þ

kr;1

kf ;2

N þ O2 ¢ O þ NO

ð2Þ

kr;2

kf ;3

3.2. Numerical models

N þ OH ¢ H þ NO

Numerical calculations are carried out using a commercial computational fluid dynamics code (FLUENT, ANSYS) in order to predict turbulent flow, coal particle motion, turbulent combustion and NOx emissions in the boiler. The time averaged conservation equations for mass, momentum, enthalpy and species are solved using the SIMPLE algorithm for predicting the flow, temperature and concentration of gas species within the boiler. The RNG k–e model is used instead of the standard k–e model to consider the effect of swirling turbulent flow in the furnace, since the RNG k–e model is superior to the standard k–e model for flow prediction with swirl or sharp change in the calculation domain [1,19,20]. Lagrangian particle trajectories of the pulverized coal particles are calculated throughout the computational domain. The dispersion of particles due to gas turbulence is predicted using the stochastic tracking model which includes the effect of instantaneous turbulent velocity fluctuations of the gas on the particle trajectories. The interaction between gas and coal particles is considered every 25 iterations for fluid flow. The discrete ordinates (DO) radiation model is used to simulate radiation heat transfer. Absorption coefficients of the gas phase are calculated using the weighted-sum-ofgray-gases model (WSGGM) [21–23]. Coal devolatilization and char combustion take place while the coal particles are traveling through the gas. Coal devolatilization is modeled by the two-competing-rates model proposed by Kobayashi et al. [24] and the gas released during this process contributes to the gaseous reaction. The char combustion is computed according to the kinetics/diffusion-limited model of Baum and Street [25] and Field [26] where the surface reaction rate is determined either by kinetics or by a diffusion rate. Since the non-premixed model is used in this study, a mixture-fraction equation is solved instead of equations for individual species, and the individual species concentration is derived from the predicted mixture fraction concentration under the assumption of chemical equilibrium. Interaction between the turbulence and chemistry is accounted for with the b-function probability density function (PDF) [27,28]. All models used in the current study are widely used due to the fact that they have shown themselves to be efficient in capturing

where kf and kr represent the forward and reverse reaction rates, respectively, and are determined based on the evaluation of Hanson and Salimian [30]. For the above thermal NOx mechanisms, only a NO species transport equation is required:

ð3Þ

kr;3

@ v Y NO Þ ¼ r  ðqDrY NO Þ þ SNO ðqY NO Þ þ r  ðq~ @t

ð4Þ

where YNO is a mass fraction of NO species in the gas phase, D is the effective diffusion coefficient, and SNO is the source term for NO species. The net rate of formation of NO species via Eqs. (1)–(3) is given by

d½NO ¼ kf ;1 ½O½N2  þ kf ;2 ½N½O2  þ kf ;3 ½N½OH  kr;1 ½NO½N dt  kr;2 ½NO½O  kr;3 ½NO½H

ð5Þ

3

where all concentrations have units of gmol/m . The concentrations of O, H, and OH are calculated by the partial equilibrium approach [29]. Fuel NOx is formed when nitrogen bound in the coal, both in the volatile matter and in the char, combines with the excess oxygen of the combustion air. The fuel NOx mechanisms are more involved. The tracking of nitrogen-containing intermediate species is important, and it is widely accepted that HCN and NH3 are the dominant intermediate species formed including nitrogen from the volatiles. It is assumed that 90% of the nitrogen from the volatiles will be converted to HCN, and the rest will form NH3 [31]. The formed HCN and NH3 generally react to form either NO in fuel-lean regions or N2 in fuel-rich regions according to the following reaction processes [32]. K4

HCN þ O2 ! NO þ . . . K5

HCN þ NO ! N2 þ . . . K6

NH3 þ O2 ! NO þ . . . K7

NH3 þ NO ! N2 þ . . .

ð6Þ ð7Þ ð8Þ ð9Þ

1723

C.R. Choi, C.N. Kim / Fuel 88 (2009) 1720–1731

where the reaction rate constants k4–k7 proposed by DeSoete [32] are used. For the above fuel NOx mechanisms, the transport equations for HCN and NH3 are solved in addition to the NO species as shown in Eq. (4).

@ v Y HCN Þ ¼ r  ðqDrY HCN Þ þ SHCN ðqY HCN Þ þ r  ðq~ @t @ v Y NH3 Þ ¼ r  ðqDrY NH3 Þ þ SNH3 ðqY NH3 Þ þ r  ðq~ @t

ð10Þ ð11Þ

where YHCN and YNH3 are mass fractions of HCN and NH3 in the gas phase. SHCN and SNH3 are the source terms for HCN and NH3. The source terms in Eqs. (4), (10), and (11) are determined as follows:

SHCN ¼ Spv c;HCN þ SHCN1 þ SHCN2

ð12Þ

SNH3 ¼ Spv c;NH3 þ SNH3 1 þ SNH3 2

ð13Þ

SNO ¼ Schar;NO þ SNO1 þ SNO2 þ SNO3

ð14Þ

ðSpv c;HCN ¼ Sv ol;HCN þ Schar;HCN and Spv c;NH3 ¼ Sv ol;NH3 þ Schar;NH3 Þ where SHCN1, SHCN2, SNH31, SNH32, SNO1, and SNO2 represent the mass consumption rates of HCN, NH3, and NO [33]. The sources of HCN, NH3, and NO from the volatiles and char in the above Eqs. (12)–(14) are determined as follows:

Sv ol Y N;v ol M w;HCN M w;N V Sv ol Y N;v ol M w;NH3 Sv ol;NH3 ¼ Mw;N V Sc Y N;char M w;HCN Schar;HCN ¼ Mw;N V Sc Y N;char M w;NH3 Schar;NH3 ¼ Mw;N V Sc Y N;char M w;NO Schar;NO ¼ Mw;N V

Sv ol;HCN ¼

ð15Þ ð16Þ ð17Þ ð18Þ ð19Þ

where

The present calculations are carried out under practical operating conditions for 100% NR operation. The total mass flow rate of the coal introduced into the furnace through the burners is 172,310 kg/h at 355 K and the total mass flow rate of the air is 1,636,800 kg/h at 588 K including 20% excess air, with equal flow rates for all the burners. That is, the flow rates of the coal and air introduced to each burner are 1.994 and 18.944 kg/s, respectively. As air is supplied through the OFA ports during OFA operation condition, 10% of the total air is assigned to the OFA ports. The type of coal used is pulverized Bituminous coal and it has a high heating value of 28.286 MJ/kg (dried basis). The coal properties, which are presented in Table 1, include the following: proximate analysis, ultimate analysis, high heating value and coal particle diameter distribution. The coal consists of 29.5% volatile matter, 54.7% fixed carbon and 15.8% ash in terms of the dry coal base (DCB). The coal particle0 s density is 1,300 kg/m3 and the mean diameter is 44 lm. The calculations assume that the size distribution of the coal particles models the Rosin–Rammler distribution with a spread parameter of 1.15. The two re-heaters, three super-heaters and one economizer are modeled as porous media with inertial resistances in order to consider their effects on flow and pressure drops, and are treated as heat sinks in the heat transfer model. This treatment is very efficient and is considered acceptable [4,8–10]. The inertial resistance coefficients can be determined based on the tube geometry and arrangement, flow patterns and pressure losses within each part. The inertial resistance coefficients in the main flow direction (ydirection) are fitted based on the pressure loss across each part as listed in Table 2. The coefficients for the other flow directions are set to 50 in order to consider the anisotropic influence of the tube geometry and arrangement on flow direction [4,7,16]. The heat sink for each zone is assigned based on the measured value listed in Table 3. The total measured heat transfer to these

Table 1 Analysis data of the pulverized coal and coal particle diameter distribution.

Svol = source of volatiles originating from the coal particles into the gas phase (kg/s) Sc = char burnout rate (kg/s) YN,vol = mass fraction of nitrogen in the volatiles YN,char = mass fraction of nitrogen in the char Mw,i = molecular weight of the species (i : HCN and NH3) V = cell volume(m3) In this study, it is assumed that the nitrogen contained in the char will be directly converted to NO, mainly as a desorption product from oxidized char nitrogen atoms [33]. Therefore, the sources (Schar,HCN and Schar;NH3 ) of HCN and NH3 from the char are set to zero. The NO formed via the above reaction mechanisms can be reacted to form N2 in a heterogeneous reaction with char particles. The NO will be reduced by a reburning process reacting with hydrocarbons as follows. kchar

char þ NO ! N2 þ . . . CHi þ NO

3.3. Numerical analysis

k8 ;k9 ;k10

!

HCN þ productsðO; OH;H2 OÞ

ð20Þ ð21Þ

where the rate constants k8–k10 are taken from Bowman [34]. Prompt NOx is formed by the reaction of atmospheric nitrogen with hydrocarbon derived from fuel in low-temperature and/or fuel-rich conditions. In this study, thermal and fuel NOx are considered while prompt NOx is neglected since it is significant only in very fuel-rich conditions and will only has a small portion of the total NOx formed in most combustion systems.

Proximate analysis [%] (DCB: Dry Coal Base)

Volatile matters Fixed carbon Ash

29.5 54.7 15.8

Ultimate analysis [%] (DAF: Dry Ash Free)

Carbon (C) Hydro (H) Oxygen (O) Nitrogen (N) Surfer (S)

81.9 5.1 10.3 1.7 1.0

Higher heating value [MJ/kg]

Dried basis Fired basis

28.286 25.456

Particle diameter distribution

Under #200 [wt%] Min./Max. diameters [lm] Mean diameter [lm] Rosin–Rammler spread parameter [-]

80 6/132 44 1.15

Table 2 Pressure losses and heat absorptions in the heat exchangers (without OFA operation).

Primary SH Secondary SH Final RH Final SH Primary RH Economizer

Pressure loss [Pa] (design)

Inertial resistance coefficient (y-dir)

Heat absorption [MW] (measured)

5.9 21.6

0.540 0.536

89.2 106.8

5.9 12.7 396.2 486.4

0.461 0.592 12.915 6.278

71.5 75.9 143.6 57.3

1724

C.R. Choi, C.N. Kim / Fuel 88 (2009) 1720–1731

Table 3 Comparison between the design/measured values and the predicted results. Section name

Measured (without OFA)

Predicted (without OFA)

Predicted (with OFA)

Heat flux

Furnace wall

515

541

520

Temperature [K]

Furnace exit Primary superheater inlet Platen superheater inleta Final reheater inleta Final superheater inleta Primary reheater inleta Economizer inleta Economizer outlet Boiler exit

1485 1485 1400 1222 1093 962 720 621 621

1509 1509 1365 1190 1076 962 727 646 646

1526 1526 1381 1204 1089 975 741 659 659

O2

Furnace exit

–

2.61

2.63

[%, mole fraction]

Boiler exit

2.50

2.56

2.57

CO2

Furnace exit

–

13.31

13.28

[%, mole fraction] NOx [ppm]

Boiler exit Furnace exit Boiler exit

15.53 – 150

15.01 161 159

15.00 148 146

Note: The value means the mass-weighted average in each section. a Design value.

zones is approximately 544 MW and the furnace wall temperature is 700 K. Since the heat transfer in the furnace is highly dependent on the emissivity of the furnace wall, the value is set after preliminary calculations. Previous numerical studies used emissivities of 0.7 [6,15] and 0.85 [3]. In this study, the preliminary calculations have been carried out with various emissivities ranging from 0.5 to 0.9. The predicted temperatures at the furnace exit for the emissivities of 0.5, 0.6 and 0.8 were 1,554, 1,509 and 1,452 K, respectively, and the total heat fluxes to the furnace walls were 506, 541 and 576 MW, respectively. These results have clearly shown that emissivity has a strong impact on the furnace exit gas temperature. As shown in Table 3, the measured temperature at the furnace exit is 1,485 K and the total heat flux to the furnace walls is 515 MW. One of the most important factors in analyzing NOx emissions is the temperature. Therefore, the value of 0.6 has been chosen for the emissivity of the furnace wall since the predicted furnace exit temperature for this value is the closest to the measured temperature.

as well as in major cross sections, as shown in Table 3. The comparison between the measured values and predicted results shows good agreement. The predicted total heat flux to the furnace walls is 541 MW, while the measured heat flux is 515 MW. The predicted temperatures at the furnace exit and boiler exit are 1,509 and 646 K, respectively, while the measured temperatures are 1,485 and 621 K, respectively. The predicted concentrations of O2, CO2 and NOx at the boiler exit are 2.56%, 15.01% and 159 ppm, respectively. The differences between the measured and predicted values given in Table 3 are less than 5% except for the NOx concentration, which has a difference of 6.1% at the boiler exit. Features in flow patterns, temperature and species distributions, and NOx emissions, as shown in the following sections, have been qualitatively consistent with the results in the previous studies [1,4,7,12,13,16]. This means that the adopted numerical models in the present calculations are reasonable for combustion analysis and NOx emissions in the boiler.

4. Results and discussion

4.2. Flow fields and coal particle trajectories

4.1. Validation of the combustion simulation results

The velocity distribution and vectors in cross sections along the furnace height are shown in Fig. 2. The flows located near the burners show more activity than in other locations. A clockwise swirling flow formed via the air and coal particles injected through the burner ports is found in the center of the furnace, as expected. The swirling flow is stronger in the lower level (section A) than in the higher level (section C). In the upper region (sections E and F), the swirling flow is remarkably reduced. The flue gas flows into the primary super-heater with a mean velocity of 8.78 m/s and standard deviation of 2.24 m/s. The upward velocity distribution located downstream, after the furnace exit, is relatively flat and the swirling flow is very weak. This implies that super-heaters and re-heaters installed downstream help the flow to be even and the residual swirling flow to be reduced. The streamlines of the flue gas and the trajectories of the coal particles are depicted in Fig. 3. The streamlines and trajectories show very complicated three-dimensional flow characteristics which promote the mixing of the air and coal particles and enhancing heat transfer via bulk motion. The coal particle trajectories are very similar to the flue gas pathlines but not coincident due to the different densities and turbulent fluctuations. The flue gas and coal particles injected from the lower burners (group A burners), initially circulate in the bottom of the furnace and the ash hopper,

At first, only the gas flow equations are solved in order to achieve for stable calculations and fast convergences for two-phase flow, combustion, heat-transfer and chemical reactions. After the flow field converges, the trajectories of the coal particles interacting with the flue gas are calculated. Next, the chemical reaction and enthalpy equations for coal combustion are taken into account. Numerical calculations are repeated until the flow and temperature converge. Finally, the NOx transport equations are solved based on the predicted flow and temperature. A convergence criterion is that the normalized residuals for all the variables need to be less than 103. Flow and temperature field convergence is obtained after more than 20,000 iterations. The NOx calculation requires smaller iterations with a total of approximately 300 times. It took about 42 h to achieve converged solutions in an IBM p595 system (2.3 GHz * 8 CPUs, 6 GB RAM, KISI supercomputing center, Korea). The results obtained from the calculation without the OFA operation are presented first and discussed in Sections 4.1–4.5. Then, the results with the OFA operation are presented and discussed along with a comparison of the two results in Section 4.6. The results predicted from the numerical calculations are then compared with the design/measured values at the furnace exit and boiler exit

C.R. Choi, C.N. Kim / Fuel 88 (2009) 1720–1731

1725

Fig. 2. Velocity distribution and velocity vectors in different cross sections.

and coal particles injected from the higher burners are shorter in comparison to the flue gas and coal particles injected from the lower burners. The average residence time and traveling length of the flue gas within the boiler are 22.2 s and 155 m, respectively, and these values are listed in Table 4. Coal devolatilization and char combustion take place while the coal particles are traveling around the furnace, as shown in Fig. 3. The average residence time of the coal particles is 21.2 s. In general, it is known that all fuel species except char are consumed very quickly in the furnace. However, char burns at a slower rate and is consumed in the central region of the furnace. In this calculation, the conversion ratio of the combusting particles is approximately 100%. This implies that the present combustion processes offer sufficient residence time even for char conversion. 4.3. Temperature distributions

Fig. 3. Streamlines of flue gas (a) and coal particle trajectories (b) colored by burner ID.

and eventually travel up through the high-temperature and swirling-flow region (so-called fire-ball) formed in the central region of the furnace, while the flue gas and coal particles from the higher burners (group B and C burners) pass around the surface of the fire-ball region. As a result, the residence times of the flue gas

The temperature distributions in the cross sections are shown in Fig. 4. The temperature of the flue gas is relatively high in the central region of the furnace where coal combustion actively takes place. As the flue gas flows from the furnace exit to the boiler exit, the temperature gradually decreases due to the heat transfer from the flue gas to the furnace walls, re-heaters, super-heaters and economizer. The change pattern of the flue gas temperature is clearly shown along the furnace height in Fig. 4b. The temperature of the air injected from the corner burners is 355 K, and it increases up to a maximum temperature of 2,132 K in the central region of the furnace. A notable temperature deviation is seen in the lower level of the furnace (section A), but in the higher levels, the temperature deviation decreases due to stronger swirling and increased mixing. With the increased height (sections A–D), the average temperatures in each cross sections also increase because of high combustion intensity. However, as the flue gas goes upward in the furnace (sections E and F), the temperature of the flue gas decreases due to the heat transfer between the flue gas and furnace walls via convection and radiation. Finally, the flue gas leaves the furnace at

1726

C.R. Choi, C.N. Kim / Fuel 88 (2009) 1720–1731

Table 4 Summary of flue gas behavior and coal particle tracking (No. of particles tracked: 35,600).

Flue gas

Coal particles

Residence time [s] Traveling length [m] Residence time [s] Conversion ratio of combusting particles [%]

Volatile Content Char Content

an average temperature of 1,524 K. The standard temperature deviation of the flue gas leaving the furnace is 89 K, which is considered to be an acceptable level. The iso-surfaces at temperatures of 1800 and 1900 K indicating high temperature zones in the furnace are depicted in Fig. 4c. These high temperature regions are closely related to thermal NOx formation which is dependent on local temperature. The relation between temperature and thermal NOx formation will be discussed in detail in Section 4.5.

Without OFA (Avg./Std. Dev.)

With OFA (Avg./Std. Dev.)

22.16/9.05 155.35/49.05

21.82/8.98 150.84/44.30

21.15/5.76 100.00 99.97

21.13/6.40 100.00 99.99

4.4. Species distributions The mass fraction distributions of O2 and CO2 and the surfaces where the O2 and CO2 mass fractions are 0.05 and 0.19, respectively, are depicted in Fig. 5. Also, their distributions in the section C penetrating group C burners are presented in Fig. 6. The O2 concentration in the furnace is relatively higher near the burners. The O2 contained in the air injected into the furnace, where the temperature is relatively higher, is quickly consumed during the

Fig. 4. Temperature distributions.

1727

C.R. Choi, C.N. Kim / Fuel 88 (2009) 1720–1731

Fig. 5. Mass fraction distributions of O2 and CO2.

Fig. 8. NOx concentration and iso-surface of 180 ppm.

to the regions of the lower O2 mass fraction. In contrast to the O2 mass fraction, the CO2 mass fraction significantly increases as the air moves from the burners due to the active combustion processes. The temperature, O2 mass fraction and CO2 mass fraction along the furnace height, the vertical center line (line A*–B*), and the diagonal line (line B–D) are presented in Fig. 7. In Fig. 7a, three high peaks of the O2 mass fraction are shown in the middle zone of the furnace when the CO2 mass fraction is low, which suggests that the sudden variation in their values is due to the supply of combustion air through the burners. In Fig. 7c, as the temperature of the flue gas increases, the O2 mass fraction decreases steeply while the CO2 mass fraction increases because volatile combustible contents of the coal burn near the burners.

Fig. 6. Mass fraction distributions of O2 and CO2 in section C.

combustion processes. As a result, the O2 mass fraction rapidly decreases. In particular, the O2 mass fraction remarkably decreases near the burners, where combustion is more active, and the fuel volatile species are more rapidly consumed. As depicted in Fig. 4, the high temperature regions in the furnace roughly correspond Temperature [K]

Temperature [K]

1000 1200 1400 1600 1800 2000

Temperature [K]

1000 1200 1400 1600 1800 2000

600 900 1200 1500 1800 2100 D

50 Temp. O2

40

CO2

Group C burners: 33 m

Height [m]

Height [m]

CO2

30 Group B burners: 26 m

20

20

Temp. O2

Group A burners: 19 m

30 20

0 0.00 0.05 0.10 0.15 0.20 0.25

0 0.00 0.05 0.10 0.15 0.20 0.25

Mass fraction [-]

Mass fraction [-]

(a) mass-weighted average in horizontal cross sections along the furnace height

10

Temp. O2

CO2

5

10

10

15

Length [m]

40

50

(b) along the center line (A*-B*) shown in Fig. 5(a)

B 0

0.00 0.05 0.10 0.15 0.20 0.25 0.30

Mass fraction [-]

(c) along the diagonal line (B-D) shown in Fig. 6(a)

Fig. 7. Variations of temperature, O2 and CO2 mass fractions.

1728

C.R. Choi, C.N. Kim / Fuel 88 (2009) 1720–1731

Fig. 9. Fuel, thermal and total NOx formation regions; in each pair of figures, the left figure indicates the NOx formation region and the right figure shows an iso-surface (2  104 gmol/m3-s).

4.5. NOx emissions The NOx concentration contour and the iso-surface showing a NOx concentration zone of 180 ppm are depicted in Fig. 8. The predicted maximum NOx concentration is 225 ppm and the relatively high NOx concentration zones are found in the furnace center where the temperature is higher and the combustion processes are more active. The predicted total NOx emission is 161 ppm with CO2 levels at 13.31% (mole fraction) at the furnace exit, and 159 ppm with CO2 levels at 15.01% (mole fraction) at the boiler exit. These predictions agree well with the measured values listed in Table 3. Based on these calculations, the fuel NOx accounts for 89.26% of the total NOx and the thermal NOx accounts for the remainder as listed in Table 5. The results show that the NOx emission is fairly low due to the combination of lower temperature peaks, use of low NOx burners and staging combustion technology. The regions of the fuel NOx and thermal NOx and the total NOx formations are shown in Fig. 9. In each pair of figures, the left figure

D

D

20

20

15

10

15

Fuel Thermal Total

5

B 0

Length [m]

Length [m]

indicates NOx formation region and the right figure shows an isosurface of a NOx formation rate of 2  104 gmol/m3-s. Since fuel and thermal NOx formation rates primarily depend on the temperature and fuel/oxygen ratio within the combustion region, NOx is mostly formed within an envelope of flames. Near the burner region, the rapid release of the volatiles contained in the coal, along with the incoming O2 injected from the burners, lead to the rapid formation of HCN or NH3. A fraction of the N contained in the formed HCN and NH3 is converted to NO in fuel-lean regions and N2 in fuel-rich regions. The zones with a thermal NOx formation rate above 2  104 gmol/m3-s are shown in the center of the furnace in Fig. 9b. The sizes of the zones shown in the right figure of Fig. 9b are smaller in comparison to those of the fuel NOx formation rate above 2  104 gmol/m3-s shown in Fig. 9a because fuel NOx is formed more readily than thermal NOx [35,36]. The zones with a high thermal NOx formation rate are found within the high temperature regions when compared to the temperature and O2 mass fraction distributions shown in Figs. 4 and 5. Fig. 9c shows the zones of the total NOx formation. The total NOx formation region is not the same as the combined region for the thermal and fuel NOx formation. This difference is caused by reversible reactions

Burnout Devolatilization

10

5

0.000 0.003 0.006 0.009

B 0

0.0000

0.0005

0.0010

0.0015

Fig. 10. Distributions of NOx formation rates in section C (unit: gmol/m3-s).

Fig. 11. Temperature distribution in the case with OFA operation (Max.: 2152 K).

1729

C.R. Choi, C.N. Kim / Fuel 88 (2009) 1720–1731 Table 5 Comparison of NOx emissions and formation rates with and without OFA operation. Without OFA

Variance ratio [%]a

With OFA

NOx at the boiler exit [ppm]

159

146

8.21

Fuel NOx formation rateb [gmol/m3s] Thermal NOx formation rateb [gmol/ m3s] Total NOx formation rateb [gmol/ m3s]

1.58E-04 1.90E-05 1.77E-04

1.44E-04 1.79E-05 1.62E-04

8.51 5.72 8.21

Portion of fuel NOx [%] Portion of thermal NOx [%]

89.26 10.74

88.97 11.03

– –

Mean temperatureb [K]

1518

1505

b

Variance ratio = (with OFA  without OFA)/(without OFA)  100. In the furnace including the hopper zone.

where the NOx that is formed by a reaction mechanism is destroyed in other reaction mechanisms (Eqs. 20 and 21). The distributions of the fuel, thermal and total NOx formation rates described in section C and the NOx formation rates and the burnout and devolatilization rates along the diagonal line (line B–D) are shown in Fig. 10. As discussed above, these results show that majority of NOx is formed within the frame envelope. The NOx formation rates are low since the oxygen concentration is very low in the central zones, although the temperature is relatively high. A correlation between NOx formation and burnout/devolatilization rates is shown in Fig. 10d and e, where the sources of intermediates (HCN and NH3) from the volatiles and the source of NO from the char depend on the burnout and devolatilization rates, respectively, as modeled in Eqs. 15, 16, and 19. These results show that the NOx formation in the boiler depends on considerably on com-

bustion processes such as burnout and devolatilization as well as temperature and species concentrations. 4.6. NOx emissions with OFA operation The calculation under OFA operation, where 10% of the total air mass flow rate is supplied through the OFA ports, is carried out to investigate NOx reduction. The predicted temperature distribution in the vertical section and the iso-surface at 1800 K indicating a relatively high temperature zone in the furnace, are depicted in Fig. 11. Compared with the temperature distribution without OFA operation discussed in Section 4.3, a relatively high temperature region is moved upward and slightly enlarged in the upper furnace due to the occurrence of combustion consuming the air supplied through the OFA ports. The predicted total heat flux to

Temperature [K] 1000 1200 1400 1600 1800 2000

50

50

40

40

Temp. O2 CO2 Temp. (OFA)

Height [m]

Height [m]

O2 (OFA)

30 20 10

140

20 10

without OFA OFA

0 120

CO2 (OFA)

30

160

180

0 0.00 0.05 0.10 0.15 0.20 0.25

200

NOx concentration [ppm]

Mass fraction [-]

(a) NOx concentration 50

Fuel Fuel (OFA)

(b) temperature and species mass fractions

50

40

Thermal Thermal (OFA)

50

40

Total Total (OFA)

40

30 Group B burners: 26 m

20

Group A burners: 19 m

10 0 0.000

Height [m]

Height [m]

OFA ports: 37 m Group C burners: 33 m

Height [m]

a

30 20 10

0.001

0.002 3

NOx formation rate [gmol/m -s]

(c) fuel NOx formation rate

0 0.0000

30 20 10 0 0.000

0.0002 3

NOx formation rate [gmol/m -s]

(d) thermal NOx formation rate

0.001

0.002 3

NOx formation rate [gmol/m -s]

(e) total NOx formation rate

Fig. 12. Comparisons of results with and without OFA operation (average in each horizontal cross section along the furnace height).

1730

C.R. Choi, C.N. Kim / Fuel 88 (2009) 1720–1731

the furnace walls is 520 MW and the predicted temperatures at the furnace exit and boiler exit are 1,526 and 659 K, respectively. The predicted concentrations of O2, CO2 and NOx at the boiler exit are 2.57%, 15.00% and 146 ppm, respectively. In comparison to the case without OFA operation, the heat flux to the furnace walls in the case with OFA operation is slightly decreased and the temperatures at the furnace exit and boiler exit are increased due to the occurrence of combustion consuming the air supplied through the OFA ports in the upper furnace zone. However, the mean temperature of the furnace including the hopper zone is decreased by 13 K. Detailed results are listed in Tables 3 and 5. With the introduction of OFA operation, the fuel and thermal NOx formations decrease by 8.51% and 5.72%, respectively, based on the data without OFA operation. As a result, the total NOx emissions in the boiler are reduced by 8.21% as noted in Table 5. The reduction in fuel NOx might be the result of decreased contact of nitrogen from the fuel with oxygen in the combustion air, which makes a fuel-rich zone where NO could be reduced to N2, while the reduction in thermal NOx might be due to the decreased temperature in the furnace. These features are consistent with the concept of air staging, in which the reaction temperature is decreased and thus the formation of thermal NOx decreases. The variances in NOx concentrations, temperatures, species mass fractions and NOx formation rates along the furnace height with and without OFA operation are presented in Fig. 12 and the values correspond to the average in each horizontal cross section. The NOx concentrations above the OFA ports decrease in both cases, as shown in Fig. 12a. In particular, the NOx concentration significantly decreases in the case with OFA operation. Fig. 12c–e shows that the NOx formation rates above the OFA ports have negative () signs. This implies that the gradual decrease of the NOx concentration in both the cases is due to heterogeneous and reburning reactions via Eqs. 20 and 21, while the significant decrease in the case with OFA operation is due to dilution caused by air injected through the OFA ports. The O2 mass fraction in the middle of the furnace is lower in the case with OFA operation than in the case without OFA operation because 10% of the total air is supplied through the OFA ports. As a result, the N contained in the HCN or NH3 formed from the volatiles contained in the coal might be readily converted to N2 rather than NO in the local environments. Therefore, the fuel NOx formation rate decreases in the regions as shown in Fig. 12c. 5. Conclusions The characteristics of the flow, combustion, temperature and NOx emissions in the 500 MWe tangentially fired pulverized-coal boiler have been numerically investigated using comprehensive models for the combustion processes and NOx formation. In order to generate accurate predictions, additional attentions have been paid in selecting the calculation domain, generating mesh, and choosing numerical models, since NOx formation is affected by fluid flow, temperature and oxygen concentration distributions. The flow fields, flue gas and coal particle motion, temperature distributions, species distributions and NOx emissions in the boiler have been obtained and compared with the measured values. The comparison between the predicted results and measured values have shown a good agreement, which implies that the adopted combustion and NOx formation models are suitable for predicting the characteristics of the flow, combustion, temperature and NOx emissions in the boiler. The relation among the temperature, O2 mass fraction and CO2 mass fraction has been clearly demonstrated based on the calculated distributions. The predicted results have shown that the NOx formation in the boiler highly depended on the combustion

processes as well as the temperature and species concentrations. The results obtained from this study have shown that OFA operation is a good way to reduce the NOx emissions from the pulverized-coal fired boiler. Air staging combustion technology (OFA operation) adopted in this boiler has helped reduce fuel NOx formation as well as thermal NOx formation under the current simulated conditions. The decrease in the fuel NOx formation is due to the decreased contact of nitrogen from the fuel with oxygen in the combustion air, while the decrease in thermal NOx formation is caused by the decrease in temperature. It is generally accepted that at temperatures below 1800 K, thermal NOx is not significantly formed by the Zeldovich mechanism and is also not a major source of NOx in local fuel-rich zones [30,35,37]. Therefore, for more accurate control of NOx formation, it is important to control temperatures above 1,800 K and encourage fuel-rich conditions. The detailed results presented in this paper may enhance the understanding of complex flow patterns, combustion processes and NOx emissions in tangentially fired pulverized-coal boilers. This paper may also provide a useful basis for NOx reduction and control. Further research focused on the reduction and control of NOx emissions may be needed under various operating conditions such as different flow rates of over fire air, tangential/tilting angles of burners and operating loads. Acknowledgement The first author would like to gratefully acknowledge the sincere help of Dr. K.C. Kim in KOPEC. References [1] Fan J, Qian L, Ma Y, Sun P, Cen K. Computational modeling of pulverized coal combustion processes in tangentially fired furnaces. Chem Eng J 2001;81:261–9. [2] European Commission. Directive 2001/80/EC on the limitation of emissions of certain pollutants into the air from large combustion plants; 2001. [3] Kokkinos A, Wasyluk D, Brower M, Barna JJ. Reducing NOx emissions in tangentially-fired boilers – a new approach. ASME international joint power generation conference; 2000. [4] Díez L, Cortés C, Pallarés J. Numerical investigation of NOx emissions from a tangentially-fired utility boiler under conventional and overfire air operation. Fuel 2007. doi:10.1016/j.fuel.2007.07.02. [5] Hill SC, Douglas Smoot L. Modeling of nitrogen oxides formation and destruction in combustion systems. Progress Energ Combust Sci 2000;26:417–58. [6] Xu M, Yuan J, Ding S, Cao H. Simulation of the gas temperature deviation in large-scale tangential coal fired utility boilers. Comput Meth Appl Mech Eng 1998;155:369–80. [7] Yin C, Rosendahl L, Condra T. Further study of the gas temperature deviation in large-scale tangentially coal-fired boilers. Fuel 2003;82:1127–37. [8] He B, Chen M, Yu Q, Liu S, Fan L, Sun S, et al. Numerical study of the optimum counter-flow mode of air jets in a large utility furnace. Comput Fluids 2004;33:1201–23. [9] Zhou H, Lou C, Cheng Q, Jiang Z, He J, Huang B, et al. Experimental investigations on visualization of three-dimensional temperature distributions in a large-scale pulverized-coal-fired boiler furnace. In: Proceedings of the combustion institute, vol. 30; 2005. p. 1699–706. [10] Belosevic S, Sijercic M, Oka S, Tucakovic D. Three-dimensional modeling of utility boiler pulverized coal tangentially fired furnace. Int J Heat Mass Transfer 2006;49:3371–8. [11] Ma Z, Iman F, Lu P, Sears R, Kong L, Rokanuzzaman A, et al. A comprehensive slagging and fouling prediction tool for coal-fired boilers and its validation/ application. Fuel Process Technol 2007;88:1035–43. [12] Asotani T, Yamashita Y, Tominaga H, Uesugi Y, Itaya Y, Mori S. Prediction of ignition behavior in a tangentially fired pulverized coal boiler using CFD. Fuel 2008;87:482–90. [13] Fan J, Sun P, Zheng Y, Ma Y, Cen K. Numerical and experimental investigation on the reduction of NOx emission in a 600 MW utility furnace by using OFA. Fuel 1999;78:1387–94. [14] Stanmore BR, Visona SP. Prediction of NO emissions from a number of coalfired power station boilers. Fuel Process Technol 2000;64:25–46. [15] Xu M, Azevedo JLT, Carvalho MG. Modelling of the combustion process and NOx emission in a utility boiler. Fuel 2000;79:1611–9. [16] Yin C, Caillat S, Harion J, Baudoin B, Perez E. Investigation of the flow, combustion, heat-transfer and emissions from a 609 MW utility tangentially fired pulverized-coal boiler. Fuel 2002;81:997–1006.

C.R. Choi, C.N. Kim / Fuel 88 (2009) 1720–1731 [17] Dimitriou DJ, Kandamby N, Lockwood FC. A mathematical modeling technique for gaseous and solid fuel reburning in pulverized coal combustors. Fuel 2003;82:2107–14. [18] Spitz N, Saveliev R, Perelman M, Korytni E, Chudnovsky B, Talanker A, et al. Firing a sub-bituminous coal in pulverized coal boilers configured for bituminous coals. Fuel 2007. doi:10.1016/j.fuel.2007.08.02. [19] Launder BE, Spalding DB. Lectures in mathematical models of turbulence. London: Academic Press; 1972. [20] Choudhury D. Introduction to the renormalization group method and turbulence modeling. Fluent Inc. Technical Memorandum TM-107; 1993. [21] Raithby G, Chui E. A finite-volume method for predicting a radiant heat transfer in enclosures with participating media. J Heat Transfer 1990;112:415–23. [22] Chui E, Raithby G. Computation of radiant heat transfer on a non-orthogonal mesh using the finite-volume method. Numeric Heat Transfer, Part B 1993;23:269–88. [23] Murthy J, Mathur S. A finite volume method for radiative heat transfer using unstructured meshes. AIAA-98-0860, January 1998. [24] Kobayashi H, Howard JB, Sarofim AF. Coal devolatilization at high temperatures. In: 16th symposium (international) on combustion. The Combustion Institute; 1976. [25] Baum M, Street PJ. Predicting the combustion behavior of coal particles. Combust Sci Tech 1971;3(5):231–43. [26] Field MA. Rate of combustion of size-graded fractions of char from a low rank coal between 1200 K and 2000 K. Combust Flame 1969;13:237–52.

1731

[27] Jones W, Whitelaw J. Calculation methods for reacting turbulent flows: a review. Combust Flame 1982;48:1–26. [28] Sivathanu Y, Faeth G. Generalized state relationships for scalar properties in non-premixed hydrocarbon/air flames. Combust Flame 1990;82:211–30. [29] FLUENT 6.3 User’s guide. Fluent Inc.; 2006. [30] Hanson RK, Salimian S. Survey of rate constants in H/N/O systems. In: Gardiner WC, editor. Combustion chemistry; 1984. p. 361. [31] Winter F, Wartha C, Loffler G, Hofbauer H. The NO and N2O formation mechanism during devolatilization and char combustion under fluidized bed conditions. In: 26th symposium (international) on combustion. The Combustion Institute; 1996. p. 3325–334. [32] DeSoete GG. Overall reaction rates of NO and N2 formation from fuel nitrogen. In: Fifteenth symposium (international) on combustion. The Combustion Institute; 1975. p. 1093–102. [33] Lockwood FC, Romo-Millanes CA. Mathematical modeling of fuel – NO emissions from PF burners. J Int Energ 1992;65:144–52. [34] Bowman CT. Chemistry of gaseous pollutant formation and destruction. In: Bartok W, Sarofim AF, editors. Fossil fuel combustion. Canada: John Wiley and Sons; 1991. [35] Palmer HB, Seery DJ. Chemistry of pollutant formation in flames. Annu Rev Phys Chem 1973;24:235–62. [36] Smoot LD, Smith PJ. Coal combustion and gasification. New York: Springer, Plenum; 1985. [37] Malte PC, Pratt DT. The role of energy-releasing kinetics in NOx formation: fuel-lean, jet-stirred CO-air combustion. Combust Sci Technol 1974;9(5):221–31.