- Email: [email protected]
Numerical investigation of air-staged combustion emphasizing char gasification and gas temperature deviation in a large-scale, tangentially fired pulverized-coal boiler
Applied Energy 177 (2016) 323–334
Contents lists available at ScienceDirect
Applied Energy journal homepage: www.elsevier.com/locate/apenergy
Numerical investigation of air-staged combustion emphasizing char gasification and gas temperature deviation in a large-scale, tangentially fired pulverized-coal boiler Yacheng Liu, Weidong Fan ⇑, Yu Li School of Mechanical and Power Engineering, Shanghai Jiao Tong University, No. 800, Dongchuan Road, Minhang District, Shanghai 200240, PR China
h i g h l i g h t s Systematic comparison of various models on simulation under deep air staging. Refined char gasification model reasonably predicts the combustion process. Significantly higher CO profile in furnace under deep air staging. Horse-saddle type distribution of the thermal load for the final super-heater.
a r t i c l e
i n f o
Article history: Received 12 February 2016 Received in revised form 5 May 2016 Accepted 21 May 2016
Keywords: Char gasification model Air-staged combustion Gas temperature deviation Residual swirling flow Tangentially fired boiler
a b s t r a c t A refined char gasification model, successfully validated in a pilot-scale 20 kW down-fired furnace, is now applied to a numerical investigation of the characteristics of the flow, temperature, and species distribution under various air-staged levels of combustion in a 600 MWe tangentially fired (T-fired) pulverizedcoal (PC) boiler. The simulation results with char gasification show that the CO concentration profile in both the primary combustion zone and the reduction zone is much higher than the corresponding case without the gasification model for deep (burnout air rate, fS = 0.42), middle (fS = 0.30), and shallow (fS = 0.17) air-staged cases. Moreover, this result is in accordance with the tests from an industrial pulverized-coal-fired furnace. It can be concluded that the char gasification mechanism should be considered in the numerical simulation of large-scale air-staged T-fired PC boilers. On the basis of a reasonable prediction of combustion characteristics, the gas temperature deviation in the crossover pass was also depicted under conditions of various air-staged levels. The result of the thermal load curve of the final super-heater panels clearly presents a saddle-type distribution for the existing two peak values. These inherent deviations originate from the residual swirling flow at the furnace exit. More specifically, parameters of swirling momentum intensity (d) in the furnace and heat flow intensity (W) at the entry of the final super-heater were employed to identify the temperature deviation in degrees. Ó 2016 Elsevier Ltd. All rights reserved.
1. Introduction Tangentially fired pulverized-coal (T-fired PC) boilers are widely employed in the thermal power generation industry. This combustion system can guarantee stable combustion of fuel, wide coal adaptability, and high combustion efficiency. However, some inherent problems related to this combustion mode always exist, especially gas temperature deviation in super-heaters and reheaters [1–3]. The latest policies on restricting NOx emissions from coal-fired power boilers are stringent. To comply with these poli⇑ Corresponding author. E-mail address: [email protected] (W. Fan). http://dx.doi.org/10.1016/j.apenergy.2016.05.135 0306-2619/Ó 2016 Elsevier Ltd. All rights reserved.
cies the overall deep air-staged combustion system (also called a separated over-fire air system) has been adopted for the updating of many old burners in active service. This system was chosen due to its high ability to abate NOx, higher reliability, greater suitability, and lower investment compared to various other low-NOx emission technologies. Previous studies indicated that over-fire air (OFA) operation is an effective way to reduce the NOx emissions of pulverized-coal (PC) fired boilers [4–6]. Zhang et al. [7] developed a numerical model to investigate the influence of horizontal bias combustion (HBC) and air staging combustion (OFA) technologies on NOx emission in a 200 MWe tangentially fired pulverizedcoal boiler. Their results showed that the air-staged combustion plays a dominant role in comparison with HBC in terms of NOx
324
Y. Liu et al. / Applied Energy 177 (2016) 323–334
reduction. Zhang et al. [8] investigated the effect of OFA on NOx emission in a 3 MW pilot-scale T-fired boiler. Simultaneously, the corresponding simulations on the boiler were conducted. The experimental and simulated results were in good agreement and showed that NOx emission decreased as the excess air ratio in the primary combustion zone decreased and residence time in the reduction zone increased, but the carbon content in the fly ash increased at the furnace exit. Recent studies elucidated the chemical reaction mechanisms of coal combustion and nitric oxide formation mechanisms. Specifically, Al-Abbas et al. [9] described the main sources of NOx formation with a model (thermal NO and fuel NO) and performed the experiments in air-firing and oxy-firing conditions on a lab-scale 100 kW Chalmers furnace. Perhaps NOx simulation and measurements under air-staged combustion conditions were not the main concern in the article, and therefore, in the case of air-staged combustion was not arranged to investigate NOx reduction. Hodzˇic´ et al. [10] firstly tested the influence on NOx reduction of reburning various fuels in parallel during co-firing of coal with sawdust and natural gas in a 20 kW PC furnace, and in their measurements, NOx reduction efficiency was higher in natural gas reburning. Though overall air-staged combustion technology has been widely applied in reducing NOx emissions in the last ten years, there is still insufficient theoretical research on this combustion mode. Few studies are focused on the effect of char gasification on the characteristics of the flow, temperature, and species distribution in tangentially fired pulverized-coal boilers under airstaged conditions. These characteristics would exert a significant impact on reduction of NOx [11]. It is widely accepted that char gasification reactions should be considered in oxy-fuel combustion of pulverized fuels [12], but fewer scholars pay major attention to applying gasification reactions in tangentially fired pulverized-coal boilers under air-staged conditions. In conventional simulations of pulverized-coal combustion, the gasification reactions between char and CO2 or H2O are not taken into consideration because these reactions are thought to be negligible compared with the primary reaction of char oxidized by O2. Actually, in the case of air-staged combustion, the partial pressures of CO2 and H2O become higher than that of O2 in the primary combustion zone of the furnace [13]. Especially in the case of deep air-staged combustion, the char gasification reaction with CO2 or H2O is more intensive. Taniguchi et al. [14] reported the experimental data and numerical calculations when considering char gasification reactions in a hightemperature tandem-type staged drop-tube where a higher temperature drop occurred in the middle due to its structure. However, CO concentrations were not obtained in their results, which significantly impacted the combustion performance and emission of pollutants. As the capacity of the T-fired PC boiler increases, the problem of the gas temperature deviation in the crossover pass tends to be more serious. This can result in tube explosions in super-heaters and re-heaters. Many researchers generally agree that the velocity deviation at the exit of furnace plays a dominant role on temperature deviation in the crossover pass where super-heaters and reheaters are arranged. Xu et al. [15] reported, by simulations and experiments in two utility boilers, that gas temperature deviation originates from the after-twirl and the platen SH, and it could be decreased by decreasing the tangential circle diameter of the secondary air if the flow field was not disturbed. Moreover, other effective ways to alleviate deviation in a horizontal pass for large-scale T-fired boilers were proposed by Yin et al. [16], such as a deeper, higher furnace arch and different arrangements of division platen SH. This paper presents a numerical study of the importance of a refined char gasification model on combustion characteristics in a 600 MWe tangentially fired pulverized-coal boiler under differ-
ent degrees of air staging. The numerical models used have been validated against experimental data from a pilot-scale 20 kW down-fired furnace. Meanwhile, the conditions without the gasification model were also compared to present the impact of considering gasification on the accuracy of the prediction of combustion characteristics. Major attention is paid to the CO concentration profile due to a more important role of CO in the combustion reaction and NOx reduction, which significantly impacts the prediction of NOx formation. On the basis of a reasonable prediction of combustion characteristics in the furnace, the gas temperature deviation of the final super-heaters in the horizontal flue was analyzed. 2. Design and operating conditions of the boiler 2.1. Boiler specifications The boiler considered in this study is a tangentially fired furnace, ultra-supercritical, once-through, single reheat, 600 MWe boiler and is shown schematically in Fig. 1. The height to the furnace exit is approximately 57.5 m, and the horizontal cross section of the furnace has a width of 18.816 m and depth of 18.144 m. The utility boiler has six layers of low NOx concentric firing system (LNCFS) burners with two layers of close-coupled over-fire air (CCOFA) and five layers of separate over-fire air (SOFA) to provide burnout air. The primary air and secondary air from the nozzles installed at the four corners are injected into the furnace center to form two imaginary circles in the furnace center which rotate clockwise. Moreover, as shown in Fig. 2, Concentric Firing Secondary (CFS) air nozzles are biased 22° with respect to the jet direction of the primary air, which generates a stronger oxidizing atmosphere in the area close to the furnace wall and consequently reduces the risk of slagging. Yaw angles of the SOFA and CCOFA can be horizontally varied in the range of ±15°, the function of which is to alleviate the residual swirling flow at the furnace exit. In this paper, the yaw angle is fixed at 0°. Although twenty-four pulverized-coal burners are installed at the four corners ranging from the lower layer A to the upper layer F, only twenty burners from the upper five layers are on duty under a full-load condition. The idle burners still need to be protected by passing a little cooling air through them. That is to say, this boiler is operated with twenty low-NOx burners working during normal operations. The arrangement of the burners explained above is shown in Fig. 1; CCOFA and SOFA ports are installed above the burners. In the crossover pass of the boiler, there are division super-heaters (SH), platen SH, final SH and two re-heaters (RH) along the gas flow direction. The furnace walls from the hopper inlet to the furnace exit comprise spiral water walls, which minimize the number of tubes as well as ensuring a higher mass flux in the water-wall tubes. The other walls above the furnace are of the vertical water wall type. 2.2. Operating conditions of the boiler This boiler is designed to use Huainan (HN) bituminous coal. Its basic properties are listed in Table 1. The present calculations are carried out under the practical operating conditions with a boiler load of 100% MCR. The total mass flow rate of coal fired is 267,610 kg/h, and the total air mass flow rate is 2,345,940 kg/h at 617 K with equal flow rates distributed among all burners. In all, 17%, 30%, and 42% of the total air mass are assigned to the OFA ports, respectively. These just represent three cases of different air-staged levels. In each case the CCOFA makes up 10% of the total amount of OFA. The primary air ratio (marked by fM) means the ratio of all air injected into the primary combustion zone except CCOFA and the total air amount injected into the fur-
325
Y. Liu et al. / Applied Energy 177 (2016) 323–334
Fig. 1. Schematic of 600 MWe tangentially fired pulverized-coal boiler (unit, mm).
3. Numerical analysis procedure 3.1. Calculation domain and mesh system
Fig. 2. Z–X section of the boiler.
nace. The burnout air ratio (marked by fS) means the ratio of OFA amount to the total air. Thus, fM + fS = 1. In addition, all cases have the total excess air ratio set at 1.2 (approximately 3.5% O2 in the flue gas, if burned out). During air-staged trials, 0.17, 0.30, and 0.42 were adopted for fS to arrange trials of different staged levels. Here, fS = 0.42, fS = 0.30, and fS = 0.17 represents deep, middle, and shallow air-staged cases, respectively. The corresponding stoichiometric ratios (S.R.) in the primary combustion zone (PCZ), are 0.696, 0.840, and 0.996, respectively for deep, middle, and shallow air-staged cases.
The calculation domain includes the combustion zone, furnace, re-heaters, super-heaters, and rear pass, as shown in Fig. 1. The porous medium model has been widely used to simulate the flow through perforated plates and packed beds [17]. The re-heaters and super-heaters are treated as porous media to allow for the tube bundle’s influence on flow and pressure losses. In the simulation of such a large-scale boiler, the mesh quality of the boiler model has a great influence on the accuracy and effectiveness of the results. As observed in Fig. 3, the mesh system of the 600 MWe boiler was created using ICEM—a FLUENT pre-processor, and mesh formations in the burner regions and the cross-section throughout the burners were shown in detail. All the zones contain only hexahedral cells, and meshes are refined in the burner zones where various parameters would change dramatically because of the rapid combustion of pulverized coal. As observed in Fig. 4, a mesh-independent test was conducted to determine the appropriate mesh number and the mesh system used in the present work was 1,182,493 cells. To guarantee better entry conditions, a channel was arranged before each burner nozzle to cause the air flow and pulverized coal particles to form a fully developed flow.
3.2. Numerical models and validation 3.2.1. Numerical models Numerical calculations are carried out using a commercial computational fluid dynamics code. As frequently observed in other lit-
Table 1 Proximate analysis and ultimate analysis. Coal species
HN
Proximate analysis (wt%, as received)
Ultimate analysis (wt%, as received)
Q net;ar (MJ/kg)
V
A
M
FC
Car
Har
Oar
Nar
Sar
28.5
26
7
38.5
54.73
3.74
7.12
1.04
0.37
21.3
326
Y. Liu et al. / Applied Energy 177 (2016) 323–334
Fig. 3. Grid system of the 600 MWe boiler.
model was proposed for exactly reproducing the experimental results of the concentrations of the main species, such as CO and O2. The coal particle undergoes decomposition into char and volatile matter. Then, the volatile matter is rapidly consumed to form CO and H2, and subsequently CO2 and H2O are formed, according to the following kinetic mechanism.
x 1 y m O2 ! xCO þ H2 þ nSO2 þ N2 þn 2 2 2 2
Cx Hy Ol Sn Nm þ
ð1Þ
CO þ 0:5O2 ! CO2
ð2Þ
H2 þ 0:5O2 ! H2 O
ð3Þ
For the heterogeneous oxidation of char, the char combustion process was modeled by the multiple-surface reaction (MSR) model consisting of two-step combustion reactions (reaction (4) and (2)) and two char gasification reactions (reaction (5) and (6)).
Fig. 4. Mesh-independent test based on temperature distribution along the furnace height.
erature [18,19], a species transport model, the standard k–epsilon two-equation model, a discrete ordinates model involving radiation heat transfer, and the Lagrangian discrete phase model are applied in the calculation. The weighted sum of gray gases model (WSGGM) is adopted for the absorption coefficient, and an isotropic model is employed for the scattering function in the radiation modeling. All models used in the current study are widely used because they have been confirmed to be efficient in capturing the complex phenomena in large-scale tangentially fired boilers. The modeling of pulverized-coal combustion process includes water evaporation, devolatilization, the homogeneous reaction of volatiles, the heterogeneous reaction of char, and the formation of pollutants. In our previous works [20], a refined char gasification
Cchar þ 0:5O2 ! CO
ð4Þ
Cchar þ CO2 ! 2CO
ð5Þ
Cchar þ H2 O ! CO þ H2
ð6Þ
All of the pre-exponential factors Ar and activation energies Er are derived from the conventional model and the published literature [21–24]. The variable Rj is the reaction rate of the jth reaction. The coefficients of all reactions involved in the refined char gasification model are summarized in Table 2. The refined model proposed here is referred to as ‘‘refined” in the following discussion. To accurately present the effects of different char combustion models on the CO concentration, a systematic comparison of the different models has been conducted. There are
Table 2 The kinetic coefficients of the refined char gasification model. Reactions
Ar
Er (J/kmol)
Rate exponent
Ref.
(1) (2) (3) (4) (5) (6)
2.119e+11 2.240e+12 5.690e+11 0.005 0.00635 0.00192
2.027e+08 4.18e+07 1.465e+08 7.396e+07 1.620e+08 1.469e+08
[vol]:0.2; [O2]:1.3 [CO]:1;[O2]:0.25 [H2]:1; [O2]:0.5 [O2]:1 [CO2]:1.3 [H2O]:1
Normal [32] [33] [13] [13] [13]
Y. Liu et al. / Applied Energy 177 (2016) 323–334
two other models that are studied here for the purpose of discussing the significance of the refined gasification reactions. The first is the conventional model without gasification but consists of two-step char combustion reactions (R2 and R4), referred to as ‘‘common”, and this model, in which the dominant primary gas product from char oxidation is CO, prevails in pulverized-coal combustion [25]. The second is the model with gasification (all reactions are activated), except for the rate exponents of the species concentrations all being set as 1, referred to as ‘‘reference” [13]. In the NOx modeling, only the thermal and fuel NOx reaction mechanisms are chosen. For the thermal NOx reaction mechanism, the partial equilibrium of [O] and [OH] is chosen. For the fuel NOx reaction mechanism, the HCN/NH3/NO/N2 reaction system is adopted for the conversion of volatile-N and char-N. The reduction mechanism is chosen with the ‘‘partial equilibrium” method to calculate the re-burn model with CO as the re-burn species. Normally, the equivalent fuel type is chosen as [CH3] [13]. Specifically, the fraction of N in the char is set to 0.3 according to the experiments obtained from former experiment [20]. The conversion fractions of volatile-N and char-N are all set to 1 due to the high combustion temperature, and it is assumed that 90% of the nitrogen is released via the intermediate HCN and 10% via NH3 for volatile-N, and 30% of the nitrogen is released via the intermediate HCN and 20% via the NH3 for char-N. The remaining nitrogen is released directly as NO. The reactions considered for NO reduction are the following:
HCN þ NO ! N2 þ
ð7Þ
NH3 þ NO ! N2 þ
ð8Þ
CO þ NO ! N2 þ
ð9Þ
3.2.2. Validation of the numerical models Firstly, the model was validated by tests from a 20 kW downfired furnace as shown in Fig. 5 and 6. A series of staged airinjection ports and gas-sampling ports are set along the furnace height with constant spacing. In addition, the thermocouples are
327
set along the height of furnace to monitor the temperature profile. The completely self-sustained combustion of coal can be ensured in the furnace without the addition of any heat from electrical heaters or gas fuel. The previous tests confirmed that the temperature in the primary combustion zone was close to that under industrial conditions. The details of this pilot-scale down-fired furnace were described in previous work [26]. The final results between experimental and simulated data are presented in Fig. 7. Two important features that can be inferred from the experimental and simulated results are the differences of the peak value of the CO concentration and the actual peak position of CO concentration compared to the model with and without considering the char gasification. It can be observed that the CO concentration profile of the refined model agrees well with the experimental data for various air-staging degrees. However, the simulations of two other models could not fit the CO concentration profile at all, despite the fact that the O2 profile fits well with the experimental results. Furthermore, Fig. 8 presents a comparison of the NOx profile along the furnace between the experimental data and the simulated results at various air-staging conditions. It can also be observed that the NOx profile along the furnace of the refined model fits well with the measured values at various air-staging conditions, and it may be concluded that the refined model is reasonable applied in air-staging combustion condition. More detailed validation of the refined char gasification model was described in previous work [20]. In this paper, the refined model with gasification is expected tobe capable of predicting the performances of the large scale, tangentially fired boiler because combustion air distribution along vertical direction is similar to this down-fired furnace. Thus, tentative simulations without application of CCOFA (i.e., the total amount of OFA was injected through the SOFA nozzles) were first to be performed to compare species profile in Fig. 7. The predicted results in Fig. 9 have exhibited some similar characteristics with the down-fired furnace, such as higher CO concentration and lower O2 concentration in reduction zone (especially for deep air staging). Moreover, Costa and Azevedo [27] measured the CO concentration in an actual industrial pulverized-coal-fired furnace under the value of fS = 0.22. In their measurement, the highest CO concentration is approximately 5 vol.% in the primary combustion zone at the highest elevation of burners as a consequence of a low stoichiometric ratio. In conclusion, the refined char gasification model is a reasonable model to simulate the combustion process in the large-scale T-fired PC boiler. Then, simulations of three different air-staging degrees with or without gasification were performed when both CCOFA and SOFA were in service (CCOFA makes up 10% of the total amount of OFA). The calculated results of all cases and the design case (fS = 0.40) were summarized in Table 3. According to the comparison between the design values and the simulation values, the results with gasification were clearly in good agreement with the design case.
4. Results and discussion 4.1. Combustion process in the furnace
Fig. 5. Geometry of refractory-lined down-fired furnace. Left: cross section of the furnace; middle: the pilot-scale furnace.
4.1.1. Temperature distribution A reasonable temperature profile along the furnace height is a significant foundation from which to obtain reasonable species profiles and analyze a real reaction mechanism in the furnace. Fig. 10 presents the profiles of the average temperature and char burnout values at the horizontal planes along the furnace height. By comparison of three air-staged cases with and without gasification (the ‘‘refined” and ‘‘common” models are taken to be with and without gasification, respectively, and the same below), it is found
328
Y. Liu et al. / Applied Energy 177 (2016) 323–334
Fig. 6. Computational domain of the down-fired furnace: 1, primary air with pulverized coal; 2, secondary air; 3, upper furnace boundary; 4–7, idealized staged air ports; 8, down furnace boundary; 9, outflow.
Fig. 7. The O2 and CO profile along the furnace resulting from the refined gasification model based on the experiment (first column). sim: simulation result; exp: experimental results obtained from the down-fired furnace, the reference model with coal gasification (second column), and the conventional model without coal gasification (third column) [20].
that with gasification the char burnout values in the cross-section of the primary air nozzles are much higher than those values without gasification. Thus, the average values of gas temperature in the primary combustion zone are also higher due to the consumption of more char with gasification. In the primary combustion zone, the temperature peak would form because of the violent combustion of coal. However, before the burnout air is injected into the furnace from the SOFA nozzles, the temperature in the reduction zone would clearly decrease due to combustion stagnation. With the addition of burnout air, the temperature rises again as unburned char and combustible gas burn vigorously again. The great release of heat from this vigorous secondary burning of fuel in the burnout zone cannot maintain a monotonous increasing trend of temperature because of the considerable heat transfer from the hot flue gas to the water wall in the upper furnace. The results also show that application of deep air staging generates a lower temperature-level profile in the primary combustion zone, and the flame core moves up. Moreover, the simulated results
are in reasonable accordance with the measurements by Dale and Brent [28] in a 160 MWe pulverized-coal corner-fired utility boiler. In their measurement, the temperatures in the primary combustion zone were 1450 °C, 1380 °C, and 1300 °C, respectively for various percentage of SOFA (0%, 37%, 100%). Consequently, more unburned reactants from the primary combustion zone lead to a continual increase in gas temperature in the burnout zone. Table 3 indicates the heat flux values to the furnace water wall for various air-staging cases, and Fig. 11 shows the contours of heat flux on the right-side furnace wall. By comparison of the heat flux values from these cases, it can be observed that more heat released from the coal combustion is absorbed by the water wall during shallow air staging. In addition, the deep air-staging combustion presents approximately 10% lower heat flux in the furnace water wall than that in the case of shallow air staging. For the cases with gasification, more char would be consumed in the central area of the primary combustion zone and less char is burned out in the burnout zone. Furthermore, once the coal par-
329
Y. Liu et al. / Applied Energy 177 (2016) 323–334
Fig. 8. The NOx profile along the furnace between experimental data and simulation results from the three models at various air-staging conditions.
account for better char consumption performance in the primary combustion zone.
Fig. 9. The O2 and CO profile along the furnace height resulting from the refined gasification model under various degree of air staging (no staged air for CCOFA).
ticles are injected into the furnace, the char will be quickly consumed to the lowest level, which contributes to a lower unburned carbon mass fraction in the fly ash at the furnace exit. The char gasification reactions with CO2 and H2O from R5 and R6 could consume an additional fraction of char in the primary combustion zone compared to the results without gasification, which could
4.1.2. CO distribution A reasonable CO concentration along the furnace height plays a key role in accurately predicting the NOx concentration distribution. According to the refined char gasification model employed in this study, CO is not only a product of volatile matter and char combustion but also a product of the gasification of char with CO2 and H2O. Therefore, both the devolatilization and char consumption rates can affect the CO concentration. Fig. 12 shows the trends of the O2 and CO concentration profiles along the furnace height. A better devolatilization performance and a lower stoichiometry would result in more CO formation in the primary combustion zone, but CO would react with the O2 from the OFA injection and be converted into CO2. That is, the CO concentration rapidly rises to its peak due to the lack of oxygen in the reduction zone, and decreases rapidly to nearly zero once the burnout air is injected from the SOFA nozzles. This is in agreement with the view of Liu [11]. For the case of shallow air staging, a relatively higher O2 concentration and lower CO concentration are observed in the primary combustion zone and reduction zone due to a higher combustion air ratio (fM) in the primary combustion zone. Moreover, it can be observed that the CO concentration in the primary combustion zone is much higher than that of corresponding cases without gasification. Without considering char gasification, the CO concentration is not more than 200 ppm. This case is far removed from the actual values from tests in utility boilers, despite the fact that the
Table 3 Comparison of calculated results of all cases. Design
Furnace exit temperature, °C Heat to water wall, MW Furnace exit O2 concentration, vol% CO2 concentration, vol% CO concentration, ppm NOx concentration, mg/N m3@6%O2 Carbon content in fly ash, % Entrance temperature of final SH, °C Heat to final SH, MW
1330 660 3.5
<300 1058 78
With gasification
Without gasification
fS = 0.17
fS = 0.30
fS = 0.42
fS = 0.17
fS = 0.30
fS = 0.42
1320 708 3.43 14.3 1.47 454 0.4 1027 73.3
1326 668 3.39 14.4 648 300 0.9 1051 76
1341 648 3.46 13.5 845 256 1.6 1060.5 80.2
1146 897 3.56 14.2 0.02 – 0.5 925 50.2
1148 890 3.54 14.25 0.12 – 2.1 940 54
1217 860 3.25 14.6 13.7 – 5.6 961 58.5
330
Y. Liu et al. / Applied Energy 177 (2016) 323–334
Fig. 10. Comparison of temperature and char burnout profiles along the furnace height with or without gasification for various air-staging levels. Left column: results without gasification; right column: results with gasification.
Fig. 11. Comparison of heat flux (W/m2) in the side wall for various degrees of air staging.
O2 concentration does not result in differences. As observed in Fig. 12, in the case with gasification, the highest CO concentration in the furnace is approximately 65,000 ppm, 35,000 ppm, and 20,000 ppm, respectively, for deep, middle, and shallow air staging. However, the CO concentration decreases drastically after the
burnout air is injected from the SOFA nozzles. Thus, the simulation results of the CO concentration profile along the height of furnace without gasification is unreasonable despite the fact that the CO concentrations at the furnace exit show relative agreement for two cases with gasification and without gasification.
Y. Liu et al. / Applied Energy 177 (2016) 323–334
331
Fig. 12. Comparison of CO and O2 profiles along the furnace height with or without gasification. Left column: results without gasification; right column: results with gasification.
4.1.3. NOx distribution In the current NOx reduction model, two NOx reduction paths are taken into consideration. One is a homogeneous reduction reaction (such as CO); the other one is a heterogeneous reduction reaction (such as char). Some scholars believe that the heterogeneous reduction reaction may contribute little to the reduction of
Fig. 13. NOx profile along the furnace height for various degrees of air staging.
NOx because the homogeneous reaction is faster than the heterogeneous reaction. On the basis of a reasonable prediction of the CO concentration profile along the furnace height, Fig. 13 shows the NOx profiles along the furnace height for various degrees of air staging. The data displayed in Fig. 13 showed clearly that application of OFA has a significant positive effect on NOx reduction. More specifically, when the staged air ratio fS ranges from 0.17 to 0.42, the NOx concentration at the furnace exit decreases from 454 mg/N m3 at 6% excess O2 to 256 mg/N m3 at 6%O2, which means an NOx reduction of approximately 45%. The fact that the lower stoichiometry in the primary combustion zone leads to a lower NOx concentration at the furnace exit can be explained by the following analysis. The lower stoichiometry produces a lower-temperature, oxygen-deficient, and fuel-rich primary zone, which suppresses the fuel nitrogen being converted to NOx. It can be found that three curves show a similar trend for the change in NOx concentration along the furnace height. It fits a typical trend of the NOx concentration profile along the combustion process in the case of overall air-staged combustion [29,30]. Firstly, as the primary combustion zone is a place where most of the coal combusts, most of the coal nitrogen would be released in the primary combustion zone, and part would be oxidized into NO (the initial product type of NOx). An NO peak appeared and it is the highest NO value in the whole furnace. Next, after the gas enters the reduction zone, the NO value drops gradually and approaches a low value. Finally, the NO value tends to be stable or to change only slightly after the gas enters the burnout zone. Then, the decrease in NOx concentration in the reduction zone is attributed to the existence of abundant reducing media including a high CO
332
Y. Liu et al. / Applied Energy 177 (2016) 323–334
concentration, while the slight increase in NOx concentration above the SOFA nozzles is due to the residual nitrogen release from char combustion [8].
4.2. Gas deviation in crossover pass 4.2.1. Gas temperature deviation of the final SH In the recent years, the examples from many power plants tell us that tube ruptures usually occur in the final SH and the final RH where tube overheating occasionally takes place during the operation of boilers. The gas temperature deviation becomes uniform and flattened as the flue gas travels along the main flow direction. This means that the gas temperature deviation in the final SH is more serious than that in the final RH. Thus, in this section, the greatest attention was paid to the region of the final SH. Fig. 14 shows the gas temperature and velocity distribution of different cross-sections under middle air staging (fS = 0.30). The gas temperature distribution at the middle section of Z = 0 is displayed in Fig. 14(a). The highest temperature obviously occurs in the primary combustion zone and the maximum value can be observed in the vicinity of the exit from the burners. Then, it decreases along the main flow direction by mixing with OFA and exchanging heat with the division SH. The gas temperature in front of the final SH is shown in Fig. 14(d), and it can be observed that the gas temperature is biased to the lower-left region of the plane. That is, the temperature in the left side is higher than that in the right side. Simultaneously, the velocity vector in the front of the final SH is presented, in which the higher velocity is on the lower-left region of the plane, similar to the gas-temperature distribution. However, the particle concentration distribution seems inclined to the right side wall, which is in opposite sense to the temperature and velocity profiles. The gas velocity distributions at the furnace exit and CFS (AII) horizontal section are shown in Fig. 14(b) and (c). As the primary air from the B to F burner layers was injected into the furnace, the clockwise fire-ball formed in the center of the furnace. The velocity distribution in the CFS horizontal section exhibited a larger tangential circle diameter, which is beneficial in that it yields an oxidizing atmosphere with higher O2 concentration in the region close to the water wall and consequently reduces the risk of slagging. Moreover, the residence time for pulverized coal in
the furnace is longer when the tangential circle diameter becomes larger, which is conducive to decreasing the carbon content in the fly ash. However, a larger tangential circle would lead to an intensive residual swirling flow at the furnace exit and this would result in increasing gas temperature deviation in the crossover pass [14]. The residual swirling flow can be observed in Fig. 14(b) and the velocity directed toward the crossover pass of the residual swirl flow is much higher on the left side than on the right side due to the induction of rear pass. When flue gas enters the platen SH area in a clockwise direction, the gas on the left side is led by the platen SH and flows directly into the rear pass with a short residence time, and thus less heat is exchanged to the platen SH. In contrast, the route of the right side gas is in a ‘‘C” shape and shows a long residence time and a greater heat flux in the platen SH area [19]. Thus, the gas temperature and velocity in front of the final SH are higher on the left side than on the right side. Moreover, the gas temperature and velocity deviation nearly disappear with the increasing height in the crossover pass. 4.2.2. Effect of various degrees of air staging on deviation As discussed above, the rotation direction of the boiler is clockwise in this simulation result and the residual swirling flow at the furnace exit would cause the gas temperature and velocity deviation in the lower-left region of the final SH. It was believed that uneven gas temperature and velocity distribution are related to steam flow deviation among panels at the steam side. The main heat absorption process in the final SH is convective heat transfer between the flue gas and the tube bundle in the crossover pass. Moreover, the velocity of flue gas has a greater influence on the convective heat transfer coefficient, and the temperature of flue gas has an impact on the temperature gradient of convection heat transfer too. Thus, a new parameter was introduced to integrating the influence of flue gas flow velocity and temperature on the heat transfer. The heat flow intensity W was commonly used by ALSTOM POWER Corporation to represent the deviation of the heating surface in the crossover pass, which integrates the influence of flue gas flow velocity and temperature on the heat transfer. So this parameter can fully reflect the deviation in trends of heat transfer, defined as follows:
W ¼ qC p T v
ð10Þ
Fig. 14. Temperature and velocity profiles of different cross-sections under the condition of fS = 0.30. (a) Temperature profile in the middle section of Z = 0; (b) velocity profile at the furnace exit; (c) velocity profile in the CFS horizontal section (AII); (d) velocity, temperature, and coal particle concentration profile at the entrance of the final SH.
Y. Liu et al. / Applied Energy 177 (2016) 323–334
333
where C p is specific heat capacity, J/(kg K); T is the temperature of flue gas, K; v is the velocity of flue gas, m/s. Fig. 15 shows the comparison of the heat flow intensity in the front of the final SH for various degrees of air staging. The swirl flow in the furnace was somewhat strengthened by the addition of CCOFA and SOFA, and the flue gas mixed with the OFA air flowed into the furnace exit. That is, application of deep air staging would promote the amount of CCOFA and SOFA air and at the same time the residual swirling flow becomes more intensive at the furnace exit. The parameter of swirling momentum intensity d is employed to present the swirl intension at the furnace exit, and it is calculated as follows:
Z d ¼ 2p
ZZ
R 0
qw2 r2 dr ¼
A
qw2 rd
ð11Þ
where W is the tangential velocity of flue gas, m/s; q is the density of flue gas, kg/m3; r is the rotating radius, m; R is the equivalent radius, m. The swirling momentum intensity profile along the furnace height under three air-staged cases is shown in Fig. 16. The calculated results at the furnace exit are 3510.5 N m, 4117.6 N m, and 5685.3 N m, respectively for shallow, middle, and deep air staging. So it can be concluded that deeper air staging would lead to more intensive residual swirl flow. It can be observed from Fig. 16 that the distributions of high heat flow intensity of flue gas in all cases are biased to the lower left region, but the deviation of the shallow air staging is not more serious than others. Specifically, for the deep air-staging case, the highest value of gas heat flow intensity on the left side is approximately 1.5e+07 W/m2, and the lowest value on the right side is approximately 4.5e+06 W/m2. So the heat flow intensity deviation between the left and right sides is 1.05e+07 W/m2. In addition, the thermal load curves of the final SH panels along the Z-direction from the left side wall to the right side wall were calculated for various air-staging levels, as observed in Fig. 17. The total absorbed energy by each final SH panel consists of the front gas room radiation energy, the rear gas room radiation energy, the tube-bundle radiation energy, and the tube-bundle convection energy [31]. The simulated results of every panel of the final SH also confirmed that the deviation on the left side is higher, while another peak value can be observed on the right side. As observed in the figure, the heat duty clearly presents a saddletype distribution. Two peaks of each heat duty curve are close to two side walls in various air-staging cases. The larger the degree of air staging, the higher the peak seems to be. Actually, higher gas temperature is biased to the left side wall and higher particle concentration is biased to the right side wall. Consequently, the higher gas heat flow intensity is close to the left side wall and the higher radiation intensity from the unburned particles to the final SH is close to the right side wall. Thus, there exist two peak values in each thermal load curve. For shallow air staging, the heat
Fig. 16. Swirling momentum intensity profile along the furnace height for various degrees of air staging.
Fig. 17. Comparison of thermal load distribution of the final SH panels along the Zdirection from the left side wall to the right side wall for various degrees of air staging.
duty curve of the final SH panel is relatively steady in the zone far from side walls.
Fig. 15. Comparison of heat flow intensity W (W/m2) in the final SH for various degrees of air staging.
334
Y. Liu et al. / Applied Energy 177 (2016) 323–334
5. Conclusion The complex process involving combustion, temperature, species concentration and NOx emissions in a 600 MWe tangentially fired pulverized-coal boiler has been numerically investigated using CFD codes with a refined char gasification model. On the basis of reasonable results of these characteristics in the furnace, the deviation of the final super-heater in the crossover pass was then discussed. Three cases (shallow, middle, and deep airstaged) have been studied to compare these characteristics with or without gasification, by which we can understand the significance of employing the char gasification model in a reasonable simulation of a large boiler. The good agreement among the simulated results and the design values implies that this modeling method is capable of predicting the performance of the selected boiler. The results show that CO concentrations in the furnace are much higher than those in corresponding cases without gasification in three air-staged levels. Moreover, the char consumption rate is higher in the primary combustion zone due to introduction of gasification reactions with CO2 and H2O. The application of OFA results in lower stoichiometry in the primary combustion zone, which makes an effective contribution to the reduction of NOx. Concerning the gas temperature deviation, the results revealed that a higher gas temperature and velocity appear in the lower left region of the final SH where tube ruptures occasionally take place. Furthermore, it was found that the heat duty clearly presents a saddle-type distribution. In addition, it was discovered that a larger degree of air staging can lead to more intensive residual swirl flow at the furnace exit, which eventually results in a higher deviation of the final SH. Thus, the refined char gasification model is proposed to achieve a more reasonable prediction of the coal combustion process under air-staged combustion conditions, and in turn, the pulverized-coal combustion theory under deep air-staged combustion can provide reasonable direction to engineering applications such as burner retrofit and boiler design. Acknowledgments This work was supported by the Shanghai Economic and Information Technology Commission (Grant No. 15XI-1-25). References [1] Park HY, Baek SH, Kim YJ, Kim TH, Kang DS, Kim DW. Numerical and experimental investigations on the gas temperature deviation in a large scale, advanced low NOx, tangentially fired pulverized coal boiler. Fuel 2013;104:641–6. [2] Modlinski N. Computational modeling of a utility boiler tangentially-fired furnace retrofitted with swirl burners. Fuel Process Technol 2010;91 (11):1601–8. [3] Karampinis E, Nikolopoulos N, Nikolopoulos A, Grammelis P, Kakaras E. Numerical investigation Greek lignite/cardoon co-firing in a tangentially fired furnace. Appl Energy 2012;97:514–24. [4] Choi CR, Kim CN. Numerical investigation on the flow, combustion and NOx emission characteristics in a 500MWe tangentially fired pulverized-coal boiler. Fuel 2009;88(9):1720–31. [5] Munir S, Nimmo W, Gibbs BM. The effect of air staged, co-combustion of pulverised coal and biomass blends on NOx emissions and combustion efficiency. Fuel 2011;90(1):126–35.
[6] Fan WD, Lin ZC, Li YY, Kuang JG, Zhang MC. Effect of air-staging on anthracite combustion and NOx formation. Energy Fuels 2009;23:111–20. [7] Zhang X, Zhou J, Sun S, Sun R, Qin M. Numerical investigation of low NOx combustion strategies in tangentially-fired coal boilers. Fuel 2015;142:215–21. [8] Zhang J, Chen K, Wang CA, Xiao K, Xu X, Che D. Optimization of separated overfire air system for a utility boiler from a 3-MW pilot-scale facility. Energy Fuels 2013;27(2):1131–40. [9] Al-Abbas AH, Naser J. Effect of chemical reaction mechanisms and NOx modeling on air-fired and oxy-fuel combustion of lignite in a 100-kW furnace. Energy Fuels 2012;26(6):3329–48. [10] Hodzˇic´ N, Kazagic´ A, Smajevic´ I. Influence of multiple air staging and reburning on NOx emissions during co-firing of low rank brown coal with woody biomass and natural gas. Appl Energy 2016;168:38–47. [11] Liu H, Liu Y, Yi G, Nie L, Che D. Effects of air staging conditions on the combustion and NOx emission characteristics in a 600 MW wall fired utility boiler using lean coal. Energy Fuels 2013;27(10):5831–40. [12] Yin C, Yan J. Oxy-fuel combustion of pulverized fuels: combustion fundamentals and modeling. Appl Energy 2016;162:742–62. [13] Stadler H, Toporov D, Förster M, Kneer R. On the influence of the char gasification reactions on NO formation in flameless coal combustion. Combust Flame 2009;156(9):1755–63. [14] Taniguchi M, Kamikawa Y, Tatsumi T, Yamamoto K. Staged combustion properties for pulverized coals at high temperature. Combust Flame 2011;158 (11):2261–71. [15] Xu Mh, Yuan Jw, Ding Sf, Cao Hd. Simulation of the gas temperature deviation in large-scale tangential coal fired utility boilers. Comput Methods Appl Mech Eng 1998;155:369–80. [16] Yin C, Rosendahl L, Condra TJ. Further study of the gas temperature deviation in large-scale tangentially coal-fired boilers. Fuel 2003;82(9):1127–37. [17] Karim MR, Naser J. Progress in numerical modelling of packed bed biomass combustion. In: 19th Australasian fluid mechanics conference, Melbourne, Australia; 2014. p. 8–11. [18] Yang J-H, Kim J-EA, Hong J, Kim M, Ryu C, Kim YJ, et al. Effects of detailed operating parameters on combustion in two 500-MWe coal-fired boilers of an identical design. Fuel 2015;144:145–56. [19] Dong C, Yang Y, Yang R, Zhang J. Numerical modeling of the gasification based biomass co-firing in a 600MW pulverized coal boiler. Appl Energy 2010;87 (9):2834–8. [20] Li Y, Fan W. Effect of char gasification on NOx formation process in the deep air-staged combustion in a 20kW down flame furnace. Appl Energy 2016;164:258–67. [21] Silaen A, Wang T. Effect of turbulence and devolatilization models on coal gasification simulation in an entrained-flow gasifier. Int J Heat Mass Transf 2010;53(9–10):2074–91. [22] Irfan MF, Usman MR, Kusakabe K. Coal gasification in CO2 atmosphere and its kinetics since 1948: a brief review. Energy 2011;36(1):12–40. [23] Kim R-G, Hwang C-W, Jeon C-H. Kinetics of coal char gasification with CO2: impact of internal/external diffusion at high temperature and elevated pressure. Appl Energy 2014;129:299–307. [24] Jeong HJ, Seo DK, Hwang J. CFD modeling for coal size effect on coal gasification in a two-stage commercial entrained-bed gasifier with an improved char gasification model. Appl Energy 2014;123:29–36. [25] Zhang M, Yu J, Xu X. A new flame sheet model to reflect the influence of the oxidation of CO on the combustion of a carbon particle. Combust Flame 2005;143(3):150–8. [26] Li Y, Fan WD, Guo QH, Zhang XW. Numerical modeling of oxy-fuel combustion in a refractory-lined down flame furnace. Energy Fuels 2014;28(1):155–62. [27] Costa M, Azevedo JLT. Experimental characterization of an industrial pulverized coal-fired furnace under deep staging conditions. Combust Sci Technol 2007;179(9):1923–35. [28] Dale RT, Brent WW. Local temperature measurements in a full-scale utility boiler with overfire air. Fuel 1997;76(11):1057–66. [29] Coda B, Kluger F, Förtsch D, Spliethoff H, Hein KRG. Coal-nitrogen release and NOx evolution in air-staged combustion. Energy Fuels 1998;12:1322–7. [30] Mackrory AJ, Tree DR. Measurement of nitrogen evolution in a staged oxycombustion coal flame. Fuel 2012;93:298–304. [31] Xu L, Khan JA, Chen Z. Thermal load deviation model for superheater and reheater of a utility boiler. Appl Therm Eng 2000;20:545–58. [32] Andersen J, Rasmussen CL, Giselsson T, Glarborg P. Global combustion mechanisms for use in CFD modeling under oxy-fuel conditions. Energy Fuels 2009;23:1379–89. [33] Jones WP, Lindstedt RP. Global reaction schemes for hydrocarbon combustion. Combust Flame 1988;73:233–49.
Contents lists available at ScienceDirect
Applied Energy journal homepage: www.elsevier.com/locate/apenergy
Numerical investigation of air-staged combustion emphasizing char gasification and gas temperature deviation in a large-scale, tangentially fired pulverized-coal boiler Yacheng Liu, Weidong Fan ⇑, Yu Li School of Mechanical and Power Engineering, Shanghai Jiao Tong University, No. 800, Dongchuan Road, Minhang District, Shanghai 200240, PR China
h i g h l i g h t s Systematic comparison of various models on simulation under deep air staging. Refined char gasification model reasonably predicts the combustion process. Significantly higher CO profile in furnace under deep air staging. Horse-saddle type distribution of the thermal load for the final super-heater.
a r t i c l e
i n f o
Article history: Received 12 February 2016 Received in revised form 5 May 2016 Accepted 21 May 2016
Keywords: Char gasification model Air-staged combustion Gas temperature deviation Residual swirling flow Tangentially fired boiler
a b s t r a c t A refined char gasification model, successfully validated in a pilot-scale 20 kW down-fired furnace, is now applied to a numerical investigation of the characteristics of the flow, temperature, and species distribution under various air-staged levels of combustion in a 600 MWe tangentially fired (T-fired) pulverizedcoal (PC) boiler. The simulation results with char gasification show that the CO concentration profile in both the primary combustion zone and the reduction zone is much higher than the corresponding case without the gasification model for deep (burnout air rate, fS = 0.42), middle (fS = 0.30), and shallow (fS = 0.17) air-staged cases. Moreover, this result is in accordance with the tests from an industrial pulverized-coal-fired furnace. It can be concluded that the char gasification mechanism should be considered in the numerical simulation of large-scale air-staged T-fired PC boilers. On the basis of a reasonable prediction of combustion characteristics, the gas temperature deviation in the crossover pass was also depicted under conditions of various air-staged levels. The result of the thermal load curve of the final super-heater panels clearly presents a saddle-type distribution for the existing two peak values. These inherent deviations originate from the residual swirling flow at the furnace exit. More specifically, parameters of swirling momentum intensity (d) in the furnace and heat flow intensity (W) at the entry of the final super-heater were employed to identify the temperature deviation in degrees. Ó 2016 Elsevier Ltd. All rights reserved.
1. Introduction Tangentially fired pulverized-coal (T-fired PC) boilers are widely employed in the thermal power generation industry. This combustion system can guarantee stable combustion of fuel, wide coal adaptability, and high combustion efficiency. However, some inherent problems related to this combustion mode always exist, especially gas temperature deviation in super-heaters and reheaters [1–3]. The latest policies on restricting NOx emissions from coal-fired power boilers are stringent. To comply with these poli⇑ Corresponding author. E-mail address: [email protected] (W. Fan). http://dx.doi.org/10.1016/j.apenergy.2016.05.135 0306-2619/Ó 2016 Elsevier Ltd. All rights reserved.
cies the overall deep air-staged combustion system (also called a separated over-fire air system) has been adopted for the updating of many old burners in active service. This system was chosen due to its high ability to abate NOx, higher reliability, greater suitability, and lower investment compared to various other low-NOx emission technologies. Previous studies indicated that over-fire air (OFA) operation is an effective way to reduce the NOx emissions of pulverized-coal (PC) fired boilers [4–6]. Zhang et al. [7] developed a numerical model to investigate the influence of horizontal bias combustion (HBC) and air staging combustion (OFA) technologies on NOx emission in a 200 MWe tangentially fired pulverizedcoal boiler. Their results showed that the air-staged combustion plays a dominant role in comparison with HBC in terms of NOx
324
Y. Liu et al. / Applied Energy 177 (2016) 323–334
reduction. Zhang et al. [8] investigated the effect of OFA on NOx emission in a 3 MW pilot-scale T-fired boiler. Simultaneously, the corresponding simulations on the boiler were conducted. The experimental and simulated results were in good agreement and showed that NOx emission decreased as the excess air ratio in the primary combustion zone decreased and residence time in the reduction zone increased, but the carbon content in the fly ash increased at the furnace exit. Recent studies elucidated the chemical reaction mechanisms of coal combustion and nitric oxide formation mechanisms. Specifically, Al-Abbas et al. [9] described the main sources of NOx formation with a model (thermal NO and fuel NO) and performed the experiments in air-firing and oxy-firing conditions on a lab-scale 100 kW Chalmers furnace. Perhaps NOx simulation and measurements under air-staged combustion conditions were not the main concern in the article, and therefore, in the case of air-staged combustion was not arranged to investigate NOx reduction. Hodzˇic´ et al. [10] firstly tested the influence on NOx reduction of reburning various fuels in parallel during co-firing of coal with sawdust and natural gas in a 20 kW PC furnace, and in their measurements, NOx reduction efficiency was higher in natural gas reburning. Though overall air-staged combustion technology has been widely applied in reducing NOx emissions in the last ten years, there is still insufficient theoretical research on this combustion mode. Few studies are focused on the effect of char gasification on the characteristics of the flow, temperature, and species distribution in tangentially fired pulverized-coal boilers under airstaged conditions. These characteristics would exert a significant impact on reduction of NOx [11]. It is widely accepted that char gasification reactions should be considered in oxy-fuel combustion of pulverized fuels [12], but fewer scholars pay major attention to applying gasification reactions in tangentially fired pulverized-coal boilers under air-staged conditions. In conventional simulations of pulverized-coal combustion, the gasification reactions between char and CO2 or H2O are not taken into consideration because these reactions are thought to be negligible compared with the primary reaction of char oxidized by O2. Actually, in the case of air-staged combustion, the partial pressures of CO2 and H2O become higher than that of O2 in the primary combustion zone of the furnace [13]. Especially in the case of deep air-staged combustion, the char gasification reaction with CO2 or H2O is more intensive. Taniguchi et al. [14] reported the experimental data and numerical calculations when considering char gasification reactions in a hightemperature tandem-type staged drop-tube where a higher temperature drop occurred in the middle due to its structure. However, CO concentrations were not obtained in their results, which significantly impacted the combustion performance and emission of pollutants. As the capacity of the T-fired PC boiler increases, the problem of the gas temperature deviation in the crossover pass tends to be more serious. This can result in tube explosions in super-heaters and re-heaters. Many researchers generally agree that the velocity deviation at the exit of furnace plays a dominant role on temperature deviation in the crossover pass where super-heaters and reheaters are arranged. Xu et al. [15] reported, by simulations and experiments in two utility boilers, that gas temperature deviation originates from the after-twirl and the platen SH, and it could be decreased by decreasing the tangential circle diameter of the secondary air if the flow field was not disturbed. Moreover, other effective ways to alleviate deviation in a horizontal pass for large-scale T-fired boilers were proposed by Yin et al. [16], such as a deeper, higher furnace arch and different arrangements of division platen SH. This paper presents a numerical study of the importance of a refined char gasification model on combustion characteristics in a 600 MWe tangentially fired pulverized-coal boiler under differ-
ent degrees of air staging. The numerical models used have been validated against experimental data from a pilot-scale 20 kW down-fired furnace. Meanwhile, the conditions without the gasification model were also compared to present the impact of considering gasification on the accuracy of the prediction of combustion characteristics. Major attention is paid to the CO concentration profile due to a more important role of CO in the combustion reaction and NOx reduction, which significantly impacts the prediction of NOx formation. On the basis of a reasonable prediction of combustion characteristics in the furnace, the gas temperature deviation of the final super-heaters in the horizontal flue was analyzed. 2. Design and operating conditions of the boiler 2.1. Boiler specifications The boiler considered in this study is a tangentially fired furnace, ultra-supercritical, once-through, single reheat, 600 MWe boiler and is shown schematically in Fig. 1. The height to the furnace exit is approximately 57.5 m, and the horizontal cross section of the furnace has a width of 18.816 m and depth of 18.144 m. The utility boiler has six layers of low NOx concentric firing system (LNCFS) burners with two layers of close-coupled over-fire air (CCOFA) and five layers of separate over-fire air (SOFA) to provide burnout air. The primary air and secondary air from the nozzles installed at the four corners are injected into the furnace center to form two imaginary circles in the furnace center which rotate clockwise. Moreover, as shown in Fig. 2, Concentric Firing Secondary (CFS) air nozzles are biased 22° with respect to the jet direction of the primary air, which generates a stronger oxidizing atmosphere in the area close to the furnace wall and consequently reduces the risk of slagging. Yaw angles of the SOFA and CCOFA can be horizontally varied in the range of ±15°, the function of which is to alleviate the residual swirling flow at the furnace exit. In this paper, the yaw angle is fixed at 0°. Although twenty-four pulverized-coal burners are installed at the four corners ranging from the lower layer A to the upper layer F, only twenty burners from the upper five layers are on duty under a full-load condition. The idle burners still need to be protected by passing a little cooling air through them. That is to say, this boiler is operated with twenty low-NOx burners working during normal operations. The arrangement of the burners explained above is shown in Fig. 1; CCOFA and SOFA ports are installed above the burners. In the crossover pass of the boiler, there are division super-heaters (SH), platen SH, final SH and two re-heaters (RH) along the gas flow direction. The furnace walls from the hopper inlet to the furnace exit comprise spiral water walls, which minimize the number of tubes as well as ensuring a higher mass flux in the water-wall tubes. The other walls above the furnace are of the vertical water wall type. 2.2. Operating conditions of the boiler This boiler is designed to use Huainan (HN) bituminous coal. Its basic properties are listed in Table 1. The present calculations are carried out under the practical operating conditions with a boiler load of 100% MCR. The total mass flow rate of coal fired is 267,610 kg/h, and the total air mass flow rate is 2,345,940 kg/h at 617 K with equal flow rates distributed among all burners. In all, 17%, 30%, and 42% of the total air mass are assigned to the OFA ports, respectively. These just represent three cases of different air-staged levels. In each case the CCOFA makes up 10% of the total amount of OFA. The primary air ratio (marked by fM) means the ratio of all air injected into the primary combustion zone except CCOFA and the total air amount injected into the fur-
325
Y. Liu et al. / Applied Energy 177 (2016) 323–334
Fig. 1. Schematic of 600 MWe tangentially fired pulverized-coal boiler (unit, mm).
3. Numerical analysis procedure 3.1. Calculation domain and mesh system
Fig. 2. Z–X section of the boiler.
nace. The burnout air ratio (marked by fS) means the ratio of OFA amount to the total air. Thus, fM + fS = 1. In addition, all cases have the total excess air ratio set at 1.2 (approximately 3.5% O2 in the flue gas, if burned out). During air-staged trials, 0.17, 0.30, and 0.42 were adopted for fS to arrange trials of different staged levels. Here, fS = 0.42, fS = 0.30, and fS = 0.17 represents deep, middle, and shallow air-staged cases, respectively. The corresponding stoichiometric ratios (S.R.) in the primary combustion zone (PCZ), are 0.696, 0.840, and 0.996, respectively for deep, middle, and shallow air-staged cases.
The calculation domain includes the combustion zone, furnace, re-heaters, super-heaters, and rear pass, as shown in Fig. 1. The porous medium model has been widely used to simulate the flow through perforated plates and packed beds [17]. The re-heaters and super-heaters are treated as porous media to allow for the tube bundle’s influence on flow and pressure losses. In the simulation of such a large-scale boiler, the mesh quality of the boiler model has a great influence on the accuracy and effectiveness of the results. As observed in Fig. 3, the mesh system of the 600 MWe boiler was created using ICEM—a FLUENT pre-processor, and mesh formations in the burner regions and the cross-section throughout the burners were shown in detail. All the zones contain only hexahedral cells, and meshes are refined in the burner zones where various parameters would change dramatically because of the rapid combustion of pulverized coal. As observed in Fig. 4, a mesh-independent test was conducted to determine the appropriate mesh number and the mesh system used in the present work was 1,182,493 cells. To guarantee better entry conditions, a channel was arranged before each burner nozzle to cause the air flow and pulverized coal particles to form a fully developed flow.
3.2. Numerical models and validation 3.2.1. Numerical models Numerical calculations are carried out using a commercial computational fluid dynamics code. As frequently observed in other lit-
Table 1 Proximate analysis and ultimate analysis. Coal species
HN
Proximate analysis (wt%, as received)
Ultimate analysis (wt%, as received)
Q net;ar (MJ/kg)
V
A
M
FC
Car
Har
Oar
Nar
Sar
28.5
26
7
38.5
54.73
3.74
7.12
1.04
0.37
21.3
326
Y. Liu et al. / Applied Energy 177 (2016) 323–334
Fig. 3. Grid system of the 600 MWe boiler.
model was proposed for exactly reproducing the experimental results of the concentrations of the main species, such as CO and O2. The coal particle undergoes decomposition into char and volatile matter. Then, the volatile matter is rapidly consumed to form CO and H2, and subsequently CO2 and H2O are formed, according to the following kinetic mechanism.
x 1 y m O2 ! xCO þ H2 þ nSO2 þ N2 þn 2 2 2 2
Cx Hy Ol Sn Nm þ
ð1Þ
CO þ 0:5O2 ! CO2
ð2Þ
H2 þ 0:5O2 ! H2 O
ð3Þ
For the heterogeneous oxidation of char, the char combustion process was modeled by the multiple-surface reaction (MSR) model consisting of two-step combustion reactions (reaction (4) and (2)) and two char gasification reactions (reaction (5) and (6)).
Fig. 4. Mesh-independent test based on temperature distribution along the furnace height.
erature [18,19], a species transport model, the standard k–epsilon two-equation model, a discrete ordinates model involving radiation heat transfer, and the Lagrangian discrete phase model are applied in the calculation. The weighted sum of gray gases model (WSGGM) is adopted for the absorption coefficient, and an isotropic model is employed for the scattering function in the radiation modeling. All models used in the current study are widely used because they have been confirmed to be efficient in capturing the complex phenomena in large-scale tangentially fired boilers. The modeling of pulverized-coal combustion process includes water evaporation, devolatilization, the homogeneous reaction of volatiles, the heterogeneous reaction of char, and the formation of pollutants. In our previous works [20], a refined char gasification
Cchar þ 0:5O2 ! CO
ð4Þ
Cchar þ CO2 ! 2CO
ð5Þ
Cchar þ H2 O ! CO þ H2
ð6Þ
All of the pre-exponential factors Ar and activation energies Er are derived from the conventional model and the published literature [21–24]. The variable Rj is the reaction rate of the jth reaction. The coefficients of all reactions involved in the refined char gasification model are summarized in Table 2. The refined model proposed here is referred to as ‘‘refined” in the following discussion. To accurately present the effects of different char combustion models on the CO concentration, a systematic comparison of the different models has been conducted. There are
Table 2 The kinetic coefficients of the refined char gasification model. Reactions
Ar
Er (J/kmol)
Rate exponent
Ref.
(1) (2) (3) (4) (5) (6)
2.119e+11 2.240e+12 5.690e+11 0.005 0.00635 0.00192
2.027e+08 4.18e+07 1.465e+08 7.396e+07 1.620e+08 1.469e+08
[vol]:0.2; [O2]:1.3 [CO]:1;[O2]:0.25 [H2]:1; [O2]:0.5 [O2]:1 [CO2]:1.3 [H2O]:1
Normal [32] [33] [13] [13] [13]
Y. Liu et al. / Applied Energy 177 (2016) 323–334
two other models that are studied here for the purpose of discussing the significance of the refined gasification reactions. The first is the conventional model without gasification but consists of two-step char combustion reactions (R2 and R4), referred to as ‘‘common”, and this model, in which the dominant primary gas product from char oxidation is CO, prevails in pulverized-coal combustion [25]. The second is the model with gasification (all reactions are activated), except for the rate exponents of the species concentrations all being set as 1, referred to as ‘‘reference” [13]. In the NOx modeling, only the thermal and fuel NOx reaction mechanisms are chosen. For the thermal NOx reaction mechanism, the partial equilibrium of [O] and [OH] is chosen. For the fuel NOx reaction mechanism, the HCN/NH3/NO/N2 reaction system is adopted for the conversion of volatile-N and char-N. The reduction mechanism is chosen with the ‘‘partial equilibrium” method to calculate the re-burn model with CO as the re-burn species. Normally, the equivalent fuel type is chosen as [CH3] [13]. Specifically, the fraction of N in the char is set to 0.3 according to the experiments obtained from former experiment [20]. The conversion fractions of volatile-N and char-N are all set to 1 due to the high combustion temperature, and it is assumed that 90% of the nitrogen is released via the intermediate HCN and 10% via NH3 for volatile-N, and 30% of the nitrogen is released via the intermediate HCN and 20% via the NH3 for char-N. The remaining nitrogen is released directly as NO. The reactions considered for NO reduction are the following:
HCN þ NO ! N2 þ
ð7Þ
NH3 þ NO ! N2 þ
ð8Þ
CO þ NO ! N2 þ
ð9Þ
3.2.2. Validation of the numerical models Firstly, the model was validated by tests from a 20 kW downfired furnace as shown in Fig. 5 and 6. A series of staged airinjection ports and gas-sampling ports are set along the furnace height with constant spacing. In addition, the thermocouples are
327
set along the height of furnace to monitor the temperature profile. The completely self-sustained combustion of coal can be ensured in the furnace without the addition of any heat from electrical heaters or gas fuel. The previous tests confirmed that the temperature in the primary combustion zone was close to that under industrial conditions. The details of this pilot-scale down-fired furnace were described in previous work [26]. The final results between experimental and simulated data are presented in Fig. 7. Two important features that can be inferred from the experimental and simulated results are the differences of the peak value of the CO concentration and the actual peak position of CO concentration compared to the model with and without considering the char gasification. It can be observed that the CO concentration profile of the refined model agrees well with the experimental data for various air-staging degrees. However, the simulations of two other models could not fit the CO concentration profile at all, despite the fact that the O2 profile fits well with the experimental results. Furthermore, Fig. 8 presents a comparison of the NOx profile along the furnace between the experimental data and the simulated results at various air-staging conditions. It can also be observed that the NOx profile along the furnace of the refined model fits well with the measured values at various air-staging conditions, and it may be concluded that the refined model is reasonable applied in air-staging combustion condition. More detailed validation of the refined char gasification model was described in previous work [20]. In this paper, the refined model with gasification is expected tobe capable of predicting the performances of the large scale, tangentially fired boiler because combustion air distribution along vertical direction is similar to this down-fired furnace. Thus, tentative simulations without application of CCOFA (i.e., the total amount of OFA was injected through the SOFA nozzles) were first to be performed to compare species profile in Fig. 7. The predicted results in Fig. 9 have exhibited some similar characteristics with the down-fired furnace, such as higher CO concentration and lower O2 concentration in reduction zone (especially for deep air staging). Moreover, Costa and Azevedo [27] measured the CO concentration in an actual industrial pulverized-coal-fired furnace under the value of fS = 0.22. In their measurement, the highest CO concentration is approximately 5 vol.% in the primary combustion zone at the highest elevation of burners as a consequence of a low stoichiometric ratio. In conclusion, the refined char gasification model is a reasonable model to simulate the combustion process in the large-scale T-fired PC boiler. Then, simulations of three different air-staging degrees with or without gasification were performed when both CCOFA and SOFA were in service (CCOFA makes up 10% of the total amount of OFA). The calculated results of all cases and the design case (fS = 0.40) were summarized in Table 3. According to the comparison between the design values and the simulation values, the results with gasification were clearly in good agreement with the design case.
4. Results and discussion 4.1. Combustion process in the furnace
Fig. 5. Geometry of refractory-lined down-fired furnace. Left: cross section of the furnace; middle: the pilot-scale furnace.
4.1.1. Temperature distribution A reasonable temperature profile along the furnace height is a significant foundation from which to obtain reasonable species profiles and analyze a real reaction mechanism in the furnace. Fig. 10 presents the profiles of the average temperature and char burnout values at the horizontal planes along the furnace height. By comparison of three air-staged cases with and without gasification (the ‘‘refined” and ‘‘common” models are taken to be with and without gasification, respectively, and the same below), it is found
328
Y. Liu et al. / Applied Energy 177 (2016) 323–334
Fig. 6. Computational domain of the down-fired furnace: 1, primary air with pulverized coal; 2, secondary air; 3, upper furnace boundary; 4–7, idealized staged air ports; 8, down furnace boundary; 9, outflow.
Fig. 7. The O2 and CO profile along the furnace resulting from the refined gasification model based on the experiment (first column). sim: simulation result; exp: experimental results obtained from the down-fired furnace, the reference model with coal gasification (second column), and the conventional model without coal gasification (third column) [20].
that with gasification the char burnout values in the cross-section of the primary air nozzles are much higher than those values without gasification. Thus, the average values of gas temperature in the primary combustion zone are also higher due to the consumption of more char with gasification. In the primary combustion zone, the temperature peak would form because of the violent combustion of coal. However, before the burnout air is injected into the furnace from the SOFA nozzles, the temperature in the reduction zone would clearly decrease due to combustion stagnation. With the addition of burnout air, the temperature rises again as unburned char and combustible gas burn vigorously again. The great release of heat from this vigorous secondary burning of fuel in the burnout zone cannot maintain a monotonous increasing trend of temperature because of the considerable heat transfer from the hot flue gas to the water wall in the upper furnace. The results also show that application of deep air staging generates a lower temperature-level profile in the primary combustion zone, and the flame core moves up. Moreover, the simulated results
are in reasonable accordance with the measurements by Dale and Brent [28] in a 160 MWe pulverized-coal corner-fired utility boiler. In their measurement, the temperatures in the primary combustion zone were 1450 °C, 1380 °C, and 1300 °C, respectively for various percentage of SOFA (0%, 37%, 100%). Consequently, more unburned reactants from the primary combustion zone lead to a continual increase in gas temperature in the burnout zone. Table 3 indicates the heat flux values to the furnace water wall for various air-staging cases, and Fig. 11 shows the contours of heat flux on the right-side furnace wall. By comparison of the heat flux values from these cases, it can be observed that more heat released from the coal combustion is absorbed by the water wall during shallow air staging. In addition, the deep air-staging combustion presents approximately 10% lower heat flux in the furnace water wall than that in the case of shallow air staging. For the cases with gasification, more char would be consumed in the central area of the primary combustion zone and less char is burned out in the burnout zone. Furthermore, once the coal par-
329
Y. Liu et al. / Applied Energy 177 (2016) 323–334
Fig. 8. The NOx profile along the furnace between experimental data and simulation results from the three models at various air-staging conditions.
account for better char consumption performance in the primary combustion zone.
Fig. 9. The O2 and CO profile along the furnace height resulting from the refined gasification model under various degree of air staging (no staged air for CCOFA).
ticles are injected into the furnace, the char will be quickly consumed to the lowest level, which contributes to a lower unburned carbon mass fraction in the fly ash at the furnace exit. The char gasification reactions with CO2 and H2O from R5 and R6 could consume an additional fraction of char in the primary combustion zone compared to the results without gasification, which could
4.1.2. CO distribution A reasonable CO concentration along the furnace height plays a key role in accurately predicting the NOx concentration distribution. According to the refined char gasification model employed in this study, CO is not only a product of volatile matter and char combustion but also a product of the gasification of char with CO2 and H2O. Therefore, both the devolatilization and char consumption rates can affect the CO concentration. Fig. 12 shows the trends of the O2 and CO concentration profiles along the furnace height. A better devolatilization performance and a lower stoichiometry would result in more CO formation in the primary combustion zone, but CO would react with the O2 from the OFA injection and be converted into CO2. That is, the CO concentration rapidly rises to its peak due to the lack of oxygen in the reduction zone, and decreases rapidly to nearly zero once the burnout air is injected from the SOFA nozzles. This is in agreement with the view of Liu [11]. For the case of shallow air staging, a relatively higher O2 concentration and lower CO concentration are observed in the primary combustion zone and reduction zone due to a higher combustion air ratio (fM) in the primary combustion zone. Moreover, it can be observed that the CO concentration in the primary combustion zone is much higher than that of corresponding cases without gasification. Without considering char gasification, the CO concentration is not more than 200 ppm. This case is far removed from the actual values from tests in utility boilers, despite the fact that the
Table 3 Comparison of calculated results of all cases. Design
Furnace exit temperature, °C Heat to water wall, MW Furnace exit O2 concentration, vol% CO2 concentration, vol% CO concentration, ppm NOx concentration, mg/N m3@6%O2 Carbon content in fly ash, % Entrance temperature of final SH, °C Heat to final SH, MW
1330 660 3.5
<300 1058 78
With gasification
Without gasification
fS = 0.17
fS = 0.30
fS = 0.42
fS = 0.17
fS = 0.30
fS = 0.42
1320 708 3.43 14.3 1.47 454 0.4 1027 73.3
1326 668 3.39 14.4 648 300 0.9 1051 76
1341 648 3.46 13.5 845 256 1.6 1060.5 80.2
1146 897 3.56 14.2 0.02 – 0.5 925 50.2
1148 890 3.54 14.25 0.12 – 2.1 940 54
1217 860 3.25 14.6 13.7 – 5.6 961 58.5
330
Y. Liu et al. / Applied Energy 177 (2016) 323–334
Fig. 10. Comparison of temperature and char burnout profiles along the furnace height with or without gasification for various air-staging levels. Left column: results without gasification; right column: results with gasification.
Fig. 11. Comparison of heat flux (W/m2) in the side wall for various degrees of air staging.
O2 concentration does not result in differences. As observed in Fig. 12, in the case with gasification, the highest CO concentration in the furnace is approximately 65,000 ppm, 35,000 ppm, and 20,000 ppm, respectively, for deep, middle, and shallow air staging. However, the CO concentration decreases drastically after the
burnout air is injected from the SOFA nozzles. Thus, the simulation results of the CO concentration profile along the height of furnace without gasification is unreasonable despite the fact that the CO concentrations at the furnace exit show relative agreement for two cases with gasification and without gasification.
Y. Liu et al. / Applied Energy 177 (2016) 323–334
331
Fig. 12. Comparison of CO and O2 profiles along the furnace height with or without gasification. Left column: results without gasification; right column: results with gasification.
4.1.3. NOx distribution In the current NOx reduction model, two NOx reduction paths are taken into consideration. One is a homogeneous reduction reaction (such as CO); the other one is a heterogeneous reduction reaction (such as char). Some scholars believe that the heterogeneous reduction reaction may contribute little to the reduction of
Fig. 13. NOx profile along the furnace height for various degrees of air staging.
NOx because the homogeneous reaction is faster than the heterogeneous reaction. On the basis of a reasonable prediction of the CO concentration profile along the furnace height, Fig. 13 shows the NOx profiles along the furnace height for various degrees of air staging. The data displayed in Fig. 13 showed clearly that application of OFA has a significant positive effect on NOx reduction. More specifically, when the staged air ratio fS ranges from 0.17 to 0.42, the NOx concentration at the furnace exit decreases from 454 mg/N m3 at 6% excess O2 to 256 mg/N m3 at 6%O2, which means an NOx reduction of approximately 45%. The fact that the lower stoichiometry in the primary combustion zone leads to a lower NOx concentration at the furnace exit can be explained by the following analysis. The lower stoichiometry produces a lower-temperature, oxygen-deficient, and fuel-rich primary zone, which suppresses the fuel nitrogen being converted to NOx. It can be found that three curves show a similar trend for the change in NOx concentration along the furnace height. It fits a typical trend of the NOx concentration profile along the combustion process in the case of overall air-staged combustion [29,30]. Firstly, as the primary combustion zone is a place where most of the coal combusts, most of the coal nitrogen would be released in the primary combustion zone, and part would be oxidized into NO (the initial product type of NOx). An NO peak appeared and it is the highest NO value in the whole furnace. Next, after the gas enters the reduction zone, the NO value drops gradually and approaches a low value. Finally, the NO value tends to be stable or to change only slightly after the gas enters the burnout zone. Then, the decrease in NOx concentration in the reduction zone is attributed to the existence of abundant reducing media including a high CO
332
Y. Liu et al. / Applied Energy 177 (2016) 323–334
concentration, while the slight increase in NOx concentration above the SOFA nozzles is due to the residual nitrogen release from char combustion [8].
4.2. Gas deviation in crossover pass 4.2.1. Gas temperature deviation of the final SH In the recent years, the examples from many power plants tell us that tube ruptures usually occur in the final SH and the final RH where tube overheating occasionally takes place during the operation of boilers. The gas temperature deviation becomes uniform and flattened as the flue gas travels along the main flow direction. This means that the gas temperature deviation in the final SH is more serious than that in the final RH. Thus, in this section, the greatest attention was paid to the region of the final SH. Fig. 14 shows the gas temperature and velocity distribution of different cross-sections under middle air staging (fS = 0.30). The gas temperature distribution at the middle section of Z = 0 is displayed in Fig. 14(a). The highest temperature obviously occurs in the primary combustion zone and the maximum value can be observed in the vicinity of the exit from the burners. Then, it decreases along the main flow direction by mixing with OFA and exchanging heat with the division SH. The gas temperature in front of the final SH is shown in Fig. 14(d), and it can be observed that the gas temperature is biased to the lower-left region of the plane. That is, the temperature in the left side is higher than that in the right side. Simultaneously, the velocity vector in the front of the final SH is presented, in which the higher velocity is on the lower-left region of the plane, similar to the gas-temperature distribution. However, the particle concentration distribution seems inclined to the right side wall, which is in opposite sense to the temperature and velocity profiles. The gas velocity distributions at the furnace exit and CFS (AII) horizontal section are shown in Fig. 14(b) and (c). As the primary air from the B to F burner layers was injected into the furnace, the clockwise fire-ball formed in the center of the furnace. The velocity distribution in the CFS horizontal section exhibited a larger tangential circle diameter, which is beneficial in that it yields an oxidizing atmosphere with higher O2 concentration in the region close to the water wall and consequently reduces the risk of slagging. Moreover, the residence time for pulverized coal in
the furnace is longer when the tangential circle diameter becomes larger, which is conducive to decreasing the carbon content in the fly ash. However, a larger tangential circle would lead to an intensive residual swirling flow at the furnace exit and this would result in increasing gas temperature deviation in the crossover pass [14]. The residual swirling flow can be observed in Fig. 14(b) and the velocity directed toward the crossover pass of the residual swirl flow is much higher on the left side than on the right side due to the induction of rear pass. When flue gas enters the platen SH area in a clockwise direction, the gas on the left side is led by the platen SH and flows directly into the rear pass with a short residence time, and thus less heat is exchanged to the platen SH. In contrast, the route of the right side gas is in a ‘‘C” shape and shows a long residence time and a greater heat flux in the platen SH area [19]. Thus, the gas temperature and velocity in front of the final SH are higher on the left side than on the right side. Moreover, the gas temperature and velocity deviation nearly disappear with the increasing height in the crossover pass. 4.2.2. Effect of various degrees of air staging on deviation As discussed above, the rotation direction of the boiler is clockwise in this simulation result and the residual swirling flow at the furnace exit would cause the gas temperature and velocity deviation in the lower-left region of the final SH. It was believed that uneven gas temperature and velocity distribution are related to steam flow deviation among panels at the steam side. The main heat absorption process in the final SH is convective heat transfer between the flue gas and the tube bundle in the crossover pass. Moreover, the velocity of flue gas has a greater influence on the convective heat transfer coefficient, and the temperature of flue gas has an impact on the temperature gradient of convection heat transfer too. Thus, a new parameter was introduced to integrating the influence of flue gas flow velocity and temperature on the heat transfer. The heat flow intensity W was commonly used by ALSTOM POWER Corporation to represent the deviation of the heating surface in the crossover pass, which integrates the influence of flue gas flow velocity and temperature on the heat transfer. So this parameter can fully reflect the deviation in trends of heat transfer, defined as follows:
W ¼ qC p T v
ð10Þ
Fig. 14. Temperature and velocity profiles of different cross-sections under the condition of fS = 0.30. (a) Temperature profile in the middle section of Z = 0; (b) velocity profile at the furnace exit; (c) velocity profile in the CFS horizontal section (AII); (d) velocity, temperature, and coal particle concentration profile at the entrance of the final SH.
Y. Liu et al. / Applied Energy 177 (2016) 323–334
333
where C p is specific heat capacity, J/(kg K); T is the temperature of flue gas, K; v is the velocity of flue gas, m/s. Fig. 15 shows the comparison of the heat flow intensity in the front of the final SH for various degrees of air staging. The swirl flow in the furnace was somewhat strengthened by the addition of CCOFA and SOFA, and the flue gas mixed with the OFA air flowed into the furnace exit. That is, application of deep air staging would promote the amount of CCOFA and SOFA air and at the same time the residual swirling flow becomes more intensive at the furnace exit. The parameter of swirling momentum intensity d is employed to present the swirl intension at the furnace exit, and it is calculated as follows:
Z d ¼ 2p
ZZ
R 0
qw2 r2 dr ¼
A
qw2 rd
ð11Þ
where W is the tangential velocity of flue gas, m/s; q is the density of flue gas, kg/m3; r is the rotating radius, m; R is the equivalent radius, m. The swirling momentum intensity profile along the furnace height under three air-staged cases is shown in Fig. 16. The calculated results at the furnace exit are 3510.5 N m, 4117.6 N m, and 5685.3 N m, respectively for shallow, middle, and deep air staging. So it can be concluded that deeper air staging would lead to more intensive residual swirl flow. It can be observed from Fig. 16 that the distributions of high heat flow intensity of flue gas in all cases are biased to the lower left region, but the deviation of the shallow air staging is not more serious than others. Specifically, for the deep air-staging case, the highest value of gas heat flow intensity on the left side is approximately 1.5e+07 W/m2, and the lowest value on the right side is approximately 4.5e+06 W/m2. So the heat flow intensity deviation between the left and right sides is 1.05e+07 W/m2. In addition, the thermal load curves of the final SH panels along the Z-direction from the left side wall to the right side wall were calculated for various air-staging levels, as observed in Fig. 17. The total absorbed energy by each final SH panel consists of the front gas room radiation energy, the rear gas room radiation energy, the tube-bundle radiation energy, and the tube-bundle convection energy [31]. The simulated results of every panel of the final SH also confirmed that the deviation on the left side is higher, while another peak value can be observed on the right side. As observed in the figure, the heat duty clearly presents a saddletype distribution. Two peaks of each heat duty curve are close to two side walls in various air-staging cases. The larger the degree of air staging, the higher the peak seems to be. Actually, higher gas temperature is biased to the left side wall and higher particle concentration is biased to the right side wall. Consequently, the higher gas heat flow intensity is close to the left side wall and the higher radiation intensity from the unburned particles to the final SH is close to the right side wall. Thus, there exist two peak values in each thermal load curve. For shallow air staging, the heat
Fig. 16. Swirling momentum intensity profile along the furnace height for various degrees of air staging.
Fig. 17. Comparison of thermal load distribution of the final SH panels along the Zdirection from the left side wall to the right side wall for various degrees of air staging.
duty curve of the final SH panel is relatively steady in the zone far from side walls.
Fig. 15. Comparison of heat flow intensity W (W/m2) in the final SH for various degrees of air staging.
334
Y. Liu et al. / Applied Energy 177 (2016) 323–334
5. Conclusion The complex process involving combustion, temperature, species concentration and NOx emissions in a 600 MWe tangentially fired pulverized-coal boiler has been numerically investigated using CFD codes with a refined char gasification model. On the basis of reasonable results of these characteristics in the furnace, the deviation of the final super-heater in the crossover pass was then discussed. Three cases (shallow, middle, and deep airstaged) have been studied to compare these characteristics with or without gasification, by which we can understand the significance of employing the char gasification model in a reasonable simulation of a large boiler. The good agreement among the simulated results and the design values implies that this modeling method is capable of predicting the performance of the selected boiler. The results show that CO concentrations in the furnace are much higher than those in corresponding cases without gasification in three air-staged levels. Moreover, the char consumption rate is higher in the primary combustion zone due to introduction of gasification reactions with CO2 and H2O. The application of OFA results in lower stoichiometry in the primary combustion zone, which makes an effective contribution to the reduction of NOx. Concerning the gas temperature deviation, the results revealed that a higher gas temperature and velocity appear in the lower left region of the final SH where tube ruptures occasionally take place. Furthermore, it was found that the heat duty clearly presents a saddle-type distribution. In addition, it was discovered that a larger degree of air staging can lead to more intensive residual swirl flow at the furnace exit, which eventually results in a higher deviation of the final SH. Thus, the refined char gasification model is proposed to achieve a more reasonable prediction of the coal combustion process under air-staged combustion conditions, and in turn, the pulverized-coal combustion theory under deep air-staged combustion can provide reasonable direction to engineering applications such as burner retrofit and boiler design. Acknowledgments This work was supported by the Shanghai Economic and Information Technology Commission (Grant No. 15XI-1-25). References [1] Park HY, Baek SH, Kim YJ, Kim TH, Kang DS, Kim DW. Numerical and experimental investigations on the gas temperature deviation in a large scale, advanced low NOx, tangentially fired pulverized coal boiler. Fuel 2013;104:641–6. [2] Modlinski N. Computational modeling of a utility boiler tangentially-fired furnace retrofitted with swirl burners. Fuel Process Technol 2010;91 (11):1601–8. [3] Karampinis E, Nikolopoulos N, Nikolopoulos A, Grammelis P, Kakaras E. Numerical investigation Greek lignite/cardoon co-firing in a tangentially fired furnace. Appl Energy 2012;97:514–24. [4] Choi CR, Kim CN. Numerical investigation on the flow, combustion and NOx emission characteristics in a 500MWe tangentially fired pulverized-coal boiler. Fuel 2009;88(9):1720–31. [5] Munir S, Nimmo W, Gibbs BM. The effect of air staged, co-combustion of pulverised coal and biomass blends on NOx emissions and combustion efficiency. Fuel 2011;90(1):126–35.
[6] Fan WD, Lin ZC, Li YY, Kuang JG, Zhang MC. Effect of air-staging on anthracite combustion and NOx formation. Energy Fuels 2009;23:111–20. [7] Zhang X, Zhou J, Sun S, Sun R, Qin M. Numerical investigation of low NOx combustion strategies in tangentially-fired coal boilers. Fuel 2015;142:215–21. [8] Zhang J, Chen K, Wang CA, Xiao K, Xu X, Che D. Optimization of separated overfire air system for a utility boiler from a 3-MW pilot-scale facility. Energy Fuels 2013;27(2):1131–40. [9] Al-Abbas AH, Naser J. Effect of chemical reaction mechanisms and NOx modeling on air-fired and oxy-fuel combustion of lignite in a 100-kW furnace. Energy Fuels 2012;26(6):3329–48. [10] Hodzˇic´ N, Kazagic´ A, Smajevic´ I. Influence of multiple air staging and reburning on NOx emissions during co-firing of low rank brown coal with woody biomass and natural gas. Appl Energy 2016;168:38–47. [11] Liu H, Liu Y, Yi G, Nie L, Che D. Effects of air staging conditions on the combustion and NOx emission characteristics in a 600 MW wall fired utility boiler using lean coal. Energy Fuels 2013;27(10):5831–40. [12] Yin C, Yan J. Oxy-fuel combustion of pulverized fuels: combustion fundamentals and modeling. Appl Energy 2016;162:742–62. [13] Stadler H, Toporov D, Förster M, Kneer R. On the influence of the char gasification reactions on NO formation in flameless coal combustion. Combust Flame 2009;156(9):1755–63. [14] Taniguchi M, Kamikawa Y, Tatsumi T, Yamamoto K. Staged combustion properties for pulverized coals at high temperature. Combust Flame 2011;158 (11):2261–71. [15] Xu Mh, Yuan Jw, Ding Sf, Cao Hd. Simulation of the gas temperature deviation in large-scale tangential coal fired utility boilers. Comput Methods Appl Mech Eng 1998;155:369–80. [16] Yin C, Rosendahl L, Condra TJ. Further study of the gas temperature deviation in large-scale tangentially coal-fired boilers. Fuel 2003;82(9):1127–37. [17] Karim MR, Naser J. Progress in numerical modelling of packed bed biomass combustion. In: 19th Australasian fluid mechanics conference, Melbourne, Australia; 2014. p. 8–11. [18] Yang J-H, Kim J-EA, Hong J, Kim M, Ryu C, Kim YJ, et al. Effects of detailed operating parameters on combustion in two 500-MWe coal-fired boilers of an identical design. Fuel 2015;144:145–56. [19] Dong C, Yang Y, Yang R, Zhang J. Numerical modeling of the gasification based biomass co-firing in a 600MW pulverized coal boiler. Appl Energy 2010;87 (9):2834–8. [20] Li Y, Fan W. Effect of char gasification on NOx formation process in the deep air-staged combustion in a 20kW down flame furnace. Appl Energy 2016;164:258–67. [21] Silaen A, Wang T. Effect of turbulence and devolatilization models on coal gasification simulation in an entrained-flow gasifier. Int J Heat Mass Transf 2010;53(9–10):2074–91. [22] Irfan MF, Usman MR, Kusakabe K. Coal gasification in CO2 atmosphere and its kinetics since 1948: a brief review. Energy 2011;36(1):12–40. [23] Kim R-G, Hwang C-W, Jeon C-H. Kinetics of coal char gasification with CO2: impact of internal/external diffusion at high temperature and elevated pressure. Appl Energy 2014;129:299–307. [24] Jeong HJ, Seo DK, Hwang J. CFD modeling for coal size effect on coal gasification in a two-stage commercial entrained-bed gasifier with an improved char gasification model. Appl Energy 2014;123:29–36. [25] Zhang M, Yu J, Xu X. A new flame sheet model to reflect the influence of the oxidation of CO on the combustion of a carbon particle. Combust Flame 2005;143(3):150–8. [26] Li Y, Fan WD, Guo QH, Zhang XW. Numerical modeling of oxy-fuel combustion in a refractory-lined down flame furnace. Energy Fuels 2014;28(1):155–62. [27] Costa M, Azevedo JLT. Experimental characterization of an industrial pulverized coal-fired furnace under deep staging conditions. Combust Sci Technol 2007;179(9):1923–35. [28] Dale RT, Brent WW. Local temperature measurements in a full-scale utility boiler with overfire air. Fuel 1997;76(11):1057–66. [29] Coda B, Kluger F, Förtsch D, Spliethoff H, Hein KRG. Coal-nitrogen release and NOx evolution in air-staged combustion. Energy Fuels 1998;12:1322–7. [30] Mackrory AJ, Tree DR. Measurement of nitrogen evolution in a staged oxycombustion coal flame. Fuel 2012;93:298–304. [31] Xu L, Khan JA, Chen Z. Thermal load deviation model for superheater and reheater of a utility boiler. Appl Therm Eng 2000;20:545–58. [32] Andersen J, Rasmussen CL, Giselsson T, Glarborg P. Global combustion mechanisms for use in CFD modeling under oxy-fuel conditions. Energy Fuels 2009;23:1379–89. [33] Jones WP, Lindstedt RP. Global reaction schemes for hydrocarbon combustion. Combust Flame 1988;73:233–49.