Available online at www.sciencedirect.com
ScienceDirect Energy Procedia 114 (2017) 481 – 489
13th International Conference on Greenhouse Gas Control Technologies, GHGT-13, 14-18 November 2016, Lausanne, Switzerland
Experimental and numerical investigations on heat transfer characteristics of a 35MW oxy-fuel combustion boiler Junjun Guo, Fan Hu, Xudong Jiang, Xiaohong Huang, Pengfei Li, Zhaohui Liu*, Chuguang Zheng State Key Laboratory of Coal Combustion, School of Energy and Power Engineering, Huazhong University of Science and Technology, Luoyu Road 1037, Wuhan 430074, China
Abstract
Oxy-fuel combustion has different combustion and heat transfer characteristics from those of air-combustion, due to the high concentration of CO2 and H2O in the flue gas. This study investigated the heat transfer characteristics in a 35 MW oxy-combustion large pilot plant. The experimental and numerical investigations are carried out with a subbituminous coal. The gas temperature, exhaust emission, and heat transfer are measured in conventional aircombustion and oxy-fuel conditions with wet and dry flue gas recycle. Detailed numerical studies of combustion and heat transfer are performed with a modified radiative property model and global reaction mechanism. The results are in good agreement with the experimental data. The effects of char and ash on radiative heat transfer are investigated. The experimental and numerical results show that although the peak temperature decreases, the heat transfer in membrane-wall and superheater in oxy-fuel combustion is slightly greater than that in air-combustion. The absorption coefficient ratios of particle to gas are approximately 2.5 and 2.4 in air-combustion and oxy-fuel combustion. The particle radiation dominates the radiative heat transfer in the pulverized coal combustion, which has a significant influence on the temperature level and distribution. Published by Elsevier Ltd. Ltd. This is an open access article under the CC BY-NC-ND license © 2017 2017The TheAuthors. Authors. Published by Elsevier (http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of the organizing committee of GHGT-13. Peer-review under responsibility of the organizing committee of GHGT-13. Keywords: Oxy-fuel combustion, radiative heat transfer, large pilot, carbon capture and storage;
* Corresponding author. Tel.: +86-27-87552151; fax: +86-27-87545526. E-mail address:
[email protected]
1876-6102 © 2017 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of the organizing committee of GHGT-13. doi:10.1016/j.egypro.2017.03.1190
482
Junjun Guo et al. / Energy Procedia 114 (2017) 481 – 489
1. Introduction The oxy-fuel combustion is one of the most promising technologies for large-scale carbon capture and storage (CCS) of power plants [1,2]. A combination of oxygen and recycled flue gas is used for combustion of the fuel. As a result, the flue gas consists mainly of CO2 and H2O. The recycled flue gas (RFG) is used to control the flame temperature and ensure the oxidant volume. The furnace temperature and heat transfer in oxy-fuel combustion differ from those of air-combustion, because of the different physical and chemical properties between CO2 and N2. In order to achieve the compatible operation of air and oxy-fuel combustion in a new boiler, or retrofit an existing boiler with this technique, the heat transfer process during oxy-fuel combustion should be similar to that during aircombustion. Experimental studies of the heat transfer process in oxy-fuel combustion have been carried out in laboratory and pilot scale furnaces. From the study in Chalmers 100 kW oxy-fuel facility, Andersson et al. [3] found the total radiation intensities is similar during oxy-fuel and air combustion as long as the temperature distributions are similar. From the experiment in 300 kW facility, Tan et al. [4] obtained the heat flux in oxy-fuel combustion can match to that in air-combustion with an O2 concentration between 28% and 35% in the oxidant. For application of oxy-coal combustion on a commercial scale, several significant developments have been made on pilot and demonstration scales equivalent to 10–30MWe [5,6,7]. The 35 MW large pilot oxy-fuel combustion power plant was constructed in 2015, and stable oxy-fuel combustion was successfully established [8]. This plant was the largest demonstration plant on oxy-coal combustion in Asia. The plant was established to investigate the impacts of air leakage, oxy-air switch combustion, flue gas recycle, and other system parameters. This paper is aimed to investigate the heat transfer process of oxy-fuel combustion in the 35 MW large pilot boiler. The flue gas temperature and heat transfer in air-combustion and oxy-combustion conditions are measured. Moreover, detailed numerical studies on the oxy-fuel combustion are conducted and the results are validated by the experiment. The heat transfer process in different combustion model is discussed in detail. The effects of char and ash particle on radiative heat transfer are investigated. 2. Experimental details 2.1. System and boiler description The experiments were performed in a 35 MW oxy-fuel power plant, which is retrofitted from an existing aircombustion power plant. Both dry and wet flue gas recycle modes are carried out. This boiler can provide steam to the steam turbine for power generation. Figure 1 shows the schematic diagram of the 35 MW oxy-fuel combustion power plant. The recycle gas in milling system (red solid line in Figure 1) can jet into the furnace via the over fire stream (OFS) nozzles or let out after the fabric filter. The flue gas processing unit can realize the dust removal, desulfurization and condensation. Both in the wet and dry recycle conditions, the primary recycled flue gas is taken after the flue gas condenser (FGC). The secondary recycled flue gas is recirculated after the electrostatic precipitator (ESP) or FGC in the wet recycle or dry recycle, respectively. The oxy-fuel combustion boiler (Figure 2) is a typical industrial front wall-fired boiler, which measures 3.533 m in depth, 4.733 m in width, and 15.150 m in height. The burner system is arranged in the front wall, which including three 12 MW pulverized coal burners and two OFS nozzles. The swirl burner geometry is shown in Figure 2. Central nozzle is for central air, which is used to protect the oil pipe. The inner annular nozzle is for the primary stream and pulverized coal, which uses the RFG after condensation during oxy-fuel combustion. This is to avoid the agglomeration of coal particles and corrosion issue by high humidity in the mill system. The outer annular nozzle is for the secondary stream, which uses RFG after condensation (dry recycle) or before condensation (wet recycle) during oxy-fuel combustion. The initial temperature of the primary and secondary stream is approximately 390 K and 500 K, respectively. The swirl number is about 1.0. Several groups of superheaters are symmetrically installed in the upper furnace.
483
Junjun Guo et al. / Energy Procedia 114 (2017) 481 – 489
Figure 1. Schematic diagram of the 35 MW large pilot oxy-fuel combustion power plant and pipe system.
Figure 2 The geometry of the boiler and burner
2.2. Commissioning tests A sub-bituminous coal with 23.77% volatiles was used as fuel for the experiments. The proximate and ultimate analysis are given in Table 1. The pulverized coal is milled to give a size distribution with mass fraction of 60% < 75 µm, 78% < 91 µm, and 99% < 200 µm. Table 1. Proximate and ultimate analysis of the pulverized coal. Proximate analysis (wt %) (ad) Ultimate analysis (wt %) (ad) Moisture
Ash
Volatiles
Fixed Carbon
C
H
O
N
S
3.38
23.94
23.77
48.91
60.40
3.65
7.30
0.48
0.85
Lower heating value (MJ/kg) 23.33
484
Junjun Guo et al. / Energy Procedia 114 (2017) 481 – 489
The operating conditions are presented in Table 2. All cases are chosen to keep the same global equivalence ratio (approximately 0.85). Both in the dry and wet recycle conditions, the initial oxygen concentration is 28 vol% (DryO28 and WetO28). The choice of the oxygen fraction is motivated by the requirements of temperature and heat transfer, as well as the flame stability at different recycle rate. In all cases, the mass flow rate of the primary stream is maintaining unchanged to ensure the transport of pulverized coal to the furnace. Approximately 13% of total O2 from ASU is introduced into the primary stream pipe, resulting in the O2 concentration of the primary stream of approximately 20 vol%, which is similar to air. The O2 concentration of the secondary stream is approximately 32 vol%. The purity of oxygen from the ASU is 98.4%. The impurities in the oxygen are simplified to N2 in the simulations. The temperature, heat transfer, and exhaust emission were measured in-situ. The temperature was measured by the hand-held radiation thermometer (OS3753). The temperature calculated from the radiation thermometer should be interpreted as a relative average along the line-of-sight. An in-house-designed gas sampling probe is used to collect gas samples in the furnace. The gas is dried and filtered after leaving the probe, and then gas is sent to an on-line gas analyzer. In order to maintain stable test conditions during the entire experimental period, the on-line gas analyzers are used to monitor the flue gas compositions (i.e., O2 and CO2). To investigate the heat transfer properties, a number of thermocouples are arranged in the wall and tubes to monitor the temperature of water and steam, including boiler feed water, drum, membrane-wall, superheater, and main steam, among others. Table 2. Operating conditions. Parameter Heat input (MW) Global equivalence ratio Global oxygen concentration in oxidant
[MW] [-] [vol%]
Air 31 0.85 21
DryO28 29 0.85 28
WetO28 28 0.85 28
Mass flow rate of primary stream O2 fraction of primary stream CO2 fraction of primary stream Temperature of primary stream
[t/h] [vol%] [vol%] [K]
8.5 20.52 0.03 393
8.7 20.14 60.91 383
8.5 20.19 53.39 385
Mass flow rate of secondary stream O2 fraction of secondary stream CO2 fraction of secondary stream Swirl number of secondary stream Temperature of secondary stream
[t/h] [vol%] [vol%] [-] [K]
32.3 20.52 0.03 1.0 547
30.8 31.15 52.18 1.0 495
30.3 30.30 41.72 1.0 507
Mass flow rate of over fire stream O2 fraction of over fire stream CO2 fraction of over fire stream Temperature of over fire stream
[t/h] [vol%] [vol%] [K]
5.6 20.52 0.03 342
-----
-----
3. CFD modeling The modeling schemes for the simulations are summarized in Table 3. The realizable two-equation k-H model with the standard wall functions is implemented to model the gas phase flow. The chemical percolation devolatilization (CPD) model [9] is considered to model the devolatilization behavior. The coefficients for the CPD model are estimated from the proximate and ultimate analysis of the pulverized coal. The eddy-dissipation concept model with the JL mechanism [10] is applied to consider the chemical reactions of gas volatiles in air-combustion. However, large amounts of CO2 and H2O participate in the reactions and influence the reaction pathway in oxy-combustion. An inhouse developed global mechanism [11] based on the JL mechanism is adopted for oxy-combustion. The radiative transfer equation was solved by the discrete ordinates method with an angular discretization of 4 divisions and 4 pixels both in the polar and azimuthal directions. The WSGG model expanded for oxy-fuel condition by Guo et al. [12] was programmed into the code using a user defined function (UDF) for calculation of the gas emissivity. The radiative property of the particle is mainly determined by the particle concentration and size. The particle absorption coefficient D p and particle scattering coefficient V p are evaluated as follows, respectively [13]:
485
Junjun Guo et al. / Energy Procedia 114 (2017) 481 – 489 N
D p lim ¦ (H pn Apn V )
(1)
V p lim ¦ ª¬ 1 f pn 1 H pn Apn V º¼ V o0 n 1
(2)
V o0
n 1 N
where H pn , Apn , and f pn are the emissivity, projected area, and scattering factor of particle n. The particle emissivity and scattering factor in the present study are 0.9 and 0.6, respectively. The model equations are discretized and solved by the finite-volume method and the implicit method. The SIMPLE algorithm is used to solve pressure−velocity coupling. In order to improve the simulation accuracy, the second order upwind scheme is employed. The mass flow inlet, pressure outlet, and temperature wall boundary are set as boundary conditions. The wall temperature is considered of 615 K, which obtained according to the saturation temperature of steam. The value of 0.45 is considered to the wall emissivity, because the boiler is new, and little ash deposition is observed on the wall. The Ansys Fluent v16.0 code is used for the present simulation, and the new models are implemented by the user defined function (UDF). In all cases, a fine hexahedral structured mesh system with 3,543,786 cells is employed, which is verified by comparing the results (e.g., velocity, temperature, and species) with those obtained using a finer grid with 5,462,764 cells. Table 3 Summary of computational modeling Model
Details
Gas radiation
Improved WSGG model for oxy-fuel combustion by Guo et al. [12]
Particle radiation
D p lim ¦ (H pn Apn V )
N
V o0
n 1
V p lim ¦ ª¬ 1 f pn 1 H pn Apn V º¼ V o0 n 1 N
Radiation resolve
Discrete ordinates method with 64 angular divisions
Coal devolatilization
Chemical percolation devolatilization (CPD) model, the coefficients are estimated from the proximate and ultimate analysis of the pulverized coal
Gas-phase reactions
An in-house developed global reaction mechanism for oxy-fuel combustion[11]
Char combustion
Kinetic/diffusion-limited rate model
Coal particle
Lagrangian tracking of 123,200 particles with turbulent dispersion
Turbulence
Realizable two-equation k-H model with the standard wall functions
Reaction rate: eddy dissipation concept model
4. Results and discussion 4.1. Temperature distribution The flame temperature is the most important parameters, because of its 4th order dependence for radiative heat transfer. Therefore, in order to match the radiative heat transfer between air- and oxy-combustion, gas temperatures in the latter should be close to those found in the former. Figure 3 presents the effects of different operating conditions on the temperature distribution from the CFD results. It can be seen that a similar temperature distribution is achieved in these three cases, which indicates the radiative heat transfer in all cases is similar and the stable air- and oxycombustion can be achieved in the same burner system. The gas temperature in the burner zone slightly decreases under dry recycle condition, compared to the other cases. The temperature in wet recycle is higher than dry recycle at the same oxygen concentration in oxidant. Moreover, peak temperatures in the whole furnace of the Air, WetO28, and DryO28 cases are 1957 K, 1824 K, and 1712 K, respectively. The lower peak temperatures conducive to repress the formation of NOx. The change of the temperature distribution is caused by the physical and chemical effect of CO2. The physical effect refers to the increase in volumetric heat capacity by CO2 addition. The chemical effect refers to the increase in ignition delay and decrease in reaction rate caused by CO2 addition.
486
Junjun Guo et al. / Energy Procedia 114 (2017) 481 – 489
Air
DryO28
WetO28
Figure 3. Predicted distributions of the flue gas temperature (K) in the furnace.
4.2. Comparison of heat transfer The heat transfer parameters between air-combustion and oxy-fuel combustion are listed in Table 4. The boiler inputs in this three cases slightly differ in the realistic operation. In order to comparison, the transferred heat to membrane-wall and superheaters is normalized by the thermal input. As shown in Table 4, the predicted results agree well with the experimental data, in terms of the heat transfer, peak temperature, and species concentration in the exhaust. Due to the increase of the volumetric heat capacity of the flue gas, peak temperature decreases under oxycombustion in the measurement area, which obtained by radiation thermometer. Although the peak flame temperature of the oxy-combustion decreases, the heat transfer in membrane-wall and superheaters increases under the wet and dry recycle oxy-fuel conditions. Due to the increased triatomic gas concentration and particle concentration, both the gas and particle absorption coefficient increase. Although the flue gas mass flow rate and velocity decrease, the increase in the heat conductivity coefficient leads to the slight increase in convective heat transfer coefficient. Considering the increased heat transfer coefficient, both the convective and radiative heat transfers increase under the oxy-fuel conditions, relative to that of air-combustion.
487
Junjun Guo et al. / Energy Procedia 114 (2017) 481 – 489 Table 4 The comparison of heat transfer parameters between air-fired and oxy-fuel boilers Air Boiler input Membrane-wall Superheaters Peak temperature O2 in oxidant O2 in flue gas CO2 in flue gas
[MW] [-] [-] [K] [vol%] [vol%] [vol%]
Furnace exit temperature Oxidant flow Volumetric heat capacity Average gas absorption coefficient Average particle absorption coefficient Flue gas velocity Flue gas density Heat transfer coefficient
[K] [kg/s] [kJ/kg·K] [m-1] [m-1] [m/s] [kg/m3] [W/m2·K]
Exp. a 31 0.566 0.132 1803 21 3.7 14.6
Dry Simu. b 31 0.568 0.125 1783 21 3.8 15.3 1266 46.4 1.266 0.140 0.325 7.9 0.37 53.88
Exp. a 29 0.591 0.152 -28 3.5 80.9
Wet Simu. b 29 0.589 0.142 1701 28 3.2 81.4 1258 38.8 1.357 0.177 0.429 5.9 0.47 57.47
Exp. a 28 0.612 0.149 1720 28 3.8 70.2
Simu. b 28 0.610 0.139 1735 28 4.1 69.0 1235 39.5 1.371 0.193 0.426 6.0 0.45 59.04
a Exp. = Experiment value; b Simu. =Simulation value.
4.3. The role of particle radiation In order to investigate the influence of the particle radiation (unburn char and fly ash) on temperature distribution and heat transfer, four cases are added to calculated, including the cases with fuel of dry-ash-free pulverized coal and cases without the particle radiation under air-combustion and oxy-fuel combustion, respectively. The Air_with ash and Dry_with ash cases represent the realistic condition, which include the particle radiation of char and fly ash. The Air_w/o ash and Dry_w/o ash cases represent a given condition, which remove the fly ash. The Air_w/o particle and Dry_w/o particle cases represent a given condition, which ignore the particle radiation. The sub-bituminous coal with ash content of 23.94 wt% is used in experiment. Figure 4 illustrates the impact of particle radiation in radiation temperature under air- and oxy-combustion. It shows that the radiation temperature drops with the removing of the particle radiation interaction. The peak radiation temperature decreases from 1598 K to 1518 K to 1448 K, and 1584 K to 1487 K to 1403 K for air-combustion and oxy-combustion, respectively. It is can be seen in Figure 4, neglecting the ash radiation, although the radiation temperature decreases, the trend of the temperature distribution keeps unchanged. Removing the particle radiation will reset the radiation temperature distribution. The particle radiation has a significant influence on the temperature level and distribution. Figure 5 shows the area-averaged total absorption coefficient (including gas and particle) along the furnace height, including the air-combustion and oxy-fuel combustion. In oxy-fuel combustion, both the gas and particle absorption coefficient are larger than those in air-combustion. The contribution of unburn char radiation is mainly in the burner zone. While the contribution of ash in radiation is in all furnace. In the upper furnace, the total absorption coefficient slightly increases due to the increase of the particle concentration, which caused by the decrease of the gas velocity. The absorption coefficient ratio of particle to gas is about 2.5 and 2.4 in air-combustion and oxy-fuel combustion, and the total absorption coefficient is about 0.5 and 0.6 in air-fired and oxy-fuel combustion. So, particle radiation dominates the radiation heat transfer in the pulverized coal combustion. In order to predict the heat transfer in the furnace, an accurate particle radiative model is very important.
488
Junjun Guo et al. / Energy Procedia 114 (2017) 481 – 489
(a)
Air_with ash
(b) Air_w/o ash
(c) Air_w/o particle
(d) Dry_with ash (e) Dry_w/o ash (f) Dry_w/o particle Figure 4. The impact of particle radiation in radiation temperature under air-combustion and oxy-combustion. (a\d) cases including particle radiation with dry-basis pulverized coal; (b\e) cases including particle radiation with dry-ash-free pulverized coal; (c\f) cases without particle radiation.
Dry_with ash Dry_w/o ash Dry_w/o particle
Air_with ash Air_w/o ash Air_w/o particle
-1
Total absorption coefficient (m )
1.0
0.8
0.6
0.4
0.2
Burner A/B 0.0 2
3
4
OFA
Burner C 5
6
7
8
9
10
11
Height along the furnace (m) Figure 5. The area-averaged total absorption coefficient (including gas and particle) profile along the furnace height.
Junjun Guo et al. / Energy Procedia 114 (2017) 481 – 489
5. Conclusion In this paper, an experimental and numerical investigation on the heat transfer process is presented in an up-todate large pilot oxy-combustion plant in China. Different operation conditions (air-combustion, wet and dry flue gas recycle oxy-combustion) are examined. A stable air- and oxy-combustion can be achieved in the same burner system. The measured temperature and heat transfer are presented. Moreover, detailed numerical studies on oxy-fuel combustion are performed. The predicted results agree well with the experimental data. The initial oxygen concentration in both dry and wet recycle condition is 28 vol%. Although the peak temperature decreases, both the convective and radiative heat transfers increase under the oxy-fuel conditions than that of air-combustion because of the large amount of CO2 that participates in radiative heat transfer. Moreover, the role of the particle radiation is investigated. In this study, with a sub-bituminous coal of 23.94 wt% ash content, the particle radiation (especially of fly ash) which dominates the radiation heat transfer in the pulverized coal combustion, has a significant influence on the temperature level and distribution. Acknowledgements The authors gratefully acknowledge financial support by grants from the National Natural Science Foundation of China (51506065, 51406001), Ministry of Science and Technology of China (2013DFB60140), National Key Research and Development Program of China (S2016G9005, 2016YFB0600801), Hubei Province (2015ACA051), the EPSRC and the UKCCSRC References [1] Ghoniem AF. Needs, resources and climate change: clean and efficient conversion technologies. Prog Energy Combust Sci 2011; 37: 15-51. [2] Wall TF. Combustion processes for carbon capture. Proc Combust Inst 2007; 31: 31-47. [3] Andersson K, Johansson R, Hjärtstam S, Johnsson F, Leckner B, Exp Therm Fluid Sci 2008; 33: 67-76. [4] Tan Y, Croiset E, Douglas MA, Thambimuthu KV. Combustion characteristics of coal in a mixture of oxygen and recycled flue gas. Fuel 2006; 85: 507-12. [5] Fujimoria T, Yamadab T. Realization of oxyfuel combustion for near zero emission power generation. Proc Combust Inst 2013; 34: 2111-30. [6] Lupion M, Alvarez I, Otero P, Kuivalainen R, Lantto J, Hotta A, et al. 30 MWth CIUDEN Oxy-CFB Boiler-First Experiences. Energy Proc 2013; 37: 6179-88. [7] Anheden M, Burchhardt U, Ecke H, Faber R, Jidinger O, Giering R, et al. Overview of operational experience and results from test activities in Vattenfall's 30 MWth oxyfuel pilot plant in Schwarze Pumpe. Energy Proc 2011; 4: 941-50. [8] Zhang T. Milestone oxyfuel plant going into operation in Hubei China, http://www.globalccsinstitute.com/insights/authors//2016/05/05/milestone-oxyfuel-plant-going-operation-hubei-china?author=MTY4OTg%3D; 2016. [accessed 05.05.16] [9] Fletcher TH, Kerstein AR, Pugmire RJ, Solum M, Grant DM. Chemical percolation model for devolatilization. 3. Direct use of carbon-13 NMR data to predict effects of coal type. Energy Fuels 1992; 6: 414-31. [10] Jones WP, Lindstedt RP. Global reaction schemes for hydrocarbon combustion. Combust Flame 1988; 73: 233-49. [11] Guo J, Liu Z, Wang P, Huang X, Li J, Xu P, et al. Numerical investigation on oxy-combustion characteristics of a 200 MWe tangentially fired boiler. Fuel 2015; 140: 660-8. [12] Guo J, Li X, Huang X, Liu Z, Zheng C. A full spectrum k-distribution based weighted-sum-of-gray-gases model for oxy-fuel combustion. Int J Heat Mass Transfer 2015; 90: 218-26. [13] Chui EH, Hughes PMJ, Raithby GD. Implementation of the finite volume method for calculating radiative transfer in a pulverized fuel flame. Combust Sci Technology 1993; 92: 225-42.
489
