Experimental Research and Mathematical Modelling for Coal Reburning in Furnace

L. Chinaa Univ of Mining Tech (English Edition)VoL 16 No. 3Experimental Research and mathematicalModelling for Coal reburning in furnaceXIANG Jun, SU Sheng, SUN Lu-shi, SUN Xue-xin, ZHANG Zhong-xiao, ZHU Ji-muState Key laboratory of Coal Combustion, Huazhong University of Sci. Tech, Wuhan, Hubei 430074, China2University of Shanghai for Sci.& Tech, Shanghai 200093, chinaSHanghai Bao Steel Power Plant, Shanghai 200971, ChinaAbstract: Reburning technology is one of the most cost-effective NOx reduction strategies for coal combustion systemsIn this paper, a nitric oxide submodel incorporated into a comprehensive coal combustion model was developed forpredicting NOx reduction in a 93 kw laboratory-scale coal combustion furnace by reburning. This NO submodelincluding reburning mechanism, requires the solution of only two transport equations to model the behavior of Noreduction in the reburning process. A number of experiments have been performed in the same furnace, and theexperimental data obtained from the optimized reburn configuration was used to validate the model. Measurements andpredictions both show above 50% reduction of NO emissions for the optimized reburning process. Profile comparisonsshow that the predicted temperature and oxygen concentration match well with the measurements, and the general trendof predicted NO concentration is very similar to that measured. The results of this study show that the present nitricoxide submodel depicts quite well the observed behaviour of NO annihilation in the reburning process. It is expectedthat this usable and computationally economic model represents a useful tool to simulate the gaseous fuel reburningprocess for the researchers concerned with practical combustorsKey words: coal combustion; nitric oxide; reburning; experimental research; mathematical modelingCLC number: TK 2231 Introductionof nitrogen oxides or other key intermediates withhydrocarbons or reburn fuel compounds areCoal combustion is one of the main sources ofmportant for reburning predictions. The approach toforming NOx in the atmosphere. To reduce NOcalculate the concentration of important radicals inemissions, various NOx reduction strategies have reburning reactions is also critical for predicting NObeen investigated. Among recent developments forreduction. For the design and analysis of large andreducing NOx emissions, reburning technologyindustrial scale reburning technique applications,considered to be one of the most cost-effective onessignificant further studies for simplifying modelsfor coal combustion systems, capable of reducing and improving predictions for reburning processes50%-70%ofNO30are requiredIn the past several years, many investigatorsIn this study, a number of experiments werehave worked on laboratory-scale measurementsperformed in a laboratory-scale furnace withkinetic mechanisms and numerical simulation ofreburning and an optimized reburning configurationreburning processes!-2. However, because of thewas obtained with maximum No, reduction. Acomplex chemistry of NOx in reburning reactions, acompact NO submodel, including a reburninglarge number of computational transport equationsmechanism, was incorporated into a comprehensiveneeds to be solved. A compact NO submodel, whichcoal combustion model for predicting thissimulates the vital chemical pathways in reburningoptimized reburn process in the furnacereactions, is required to be incorporated into acomprehensive coal combustion model forsimulating the reburning process. In recent years2 Experimental Proceduresome of the reburning process predictions, making2.1 Experimental system and conditionuse of reduced kinetic schemes or global rate2.1.1 Single Burner Furnace(SBFmodelling results show that the predictions are stillFig. I represents a schematic presentation ofinaccurate. It is found that those possible reactions the experimental system. The 93 kw single-burnerReceived 12 December 2005; accepted 05 January 2006中国煤化工Project 2004CB217704-4 supported by the Special Funds for Major State Basic Research Projects ofHCNMH Roject of ChineseCorrespondingauthorTel:+86-27-87542417-8509:E-mailaddresssusheng2003@163.comJ. China Univ, of Mining Tech(English EditionVoL16No.3furnace is 0.5 m in height, 0.35 m in width and 4 mTable 1 distance of test hole distributin length. From the top to the bottom, the burner isTest hole Distance from Test hole Distance fromarranged in the order of upper secondary air(USA),burner exitburner exit(middle secondary air(MSA), primary air(PA)andlower secondary air(LSA), as shown in Fig. 2. All0.35the secondary air ports have thcross-section40.95of 20 mmx20 mm and the primary air port has adiameter of 40 mm. Detailed descriptions of this1.65facility can be found in reference [4]2. 1.2 One-dimensional Furnace(ODFThe scheme of the electrically heatedo磁combustion reactor is presented in Fig. 3. Theceramic reaction tube is made up of 7 sections with atotal heating lenth of 2.0 m and an internal diameterof 175 mm. The pulverized coal mixed with theprimary air fired down along with top, upper andr. the coal feedepdepulverized coal with feeding rates of 2.5 kg/h. Thereburning fuel port is located at 0.65m below theFig. 1 Schematic presentation of experimental systemprimary air port and 0.37 m above the secondary air1. centrifugal blower; 2. air preheater; 3. air distributor; 4. feed hopper; 5port. Methane is used as the reburn fuel Coal qualityIcking auger; 6. primary air valve; 7. direct current motor; 8. auxiliaranalysis is also showed in Table 2valve(three): 9, bumer; 10. reburning fuel tank; 11. float type flohamber; 15. test hole(eleven); 16. spy hole1. D. C. motor2. pulverized coal hopper3. pulverized coal feederprimary air port6.P Secondary air portMSA-曰-combustion chambehole0. BlowerLSA-日-￥11. cyclone separator12. induced fan3. reburning fuel14. flowmeterFig 2 Burner scheme(dimensionsFig 3 The schematic diagram of the ODFAs shown in Fig. 1, 1l test holes are positionedhorizontally on one side of the furnace wall forTable 2 Characteristics of coal(as receivedmeasuring temperature level andcomposition in the furnace. The distances betweenParticle size(wt%)the burner exit and each test hole are summarized inMoisture(wt%) 1. 48 C 61.89 Under 200 um 100Table 1. a water-cooled sampling probe is used toAsh(wt%)26.30H1Under 150 um 80extract gaseous combustion products for analyzinVolatiles(wt%) 12.O, NO and so on(wt%)69.66 N 0.98 Average 134 uma typical low-volatile coal serves ashe247151.06primary fuel. Methane is used as the reburn fuelproximate and ultimate analyses, as well as the2.2 The optimized reburning caseparticle size distributions of coal are given in Table 2For the reburning experiments, the reburn fuel andIn order to obtain the optimized reburnburnout air are introduced in the furnace from theconfiguration, a number of experiments had beentest holes. The methane is injected horizontally in previously performed against various experimentalthe furnace from a 12 mm diameter tube. Theconditions in SBF. The base and reburningburnout air is supplied by a centrifugal blower. Theexperimental conditions are shown in Table 3. Theburnout air injector tube of 40 mm diameter is alsoamount of methaned for the rebuplaced horizontally. In all the reburn experiments中国煤化工and0tthe axes of the reburn fuel and burnout air injectionthe total thee optimtubes, as well as the axis of the primary air port, areow rate odYHCNMHGd injectionpositioned in the same horizontal planepositions for the reburn fuel and burnout air wereXIANG Jun et alExperimental Research and Mathematical Modelling for Coal Reburning in Furnaceobtained by introducing methane and burnout air inTable 4 Operation conditions and NO reductionthe furnace from different test holes and measurinrate with reburning in Odfthe lowest NO exhaust emissions simultaneouslyBase case operation conditionsTable 3 Base and reburning operation conditions in SBFMass flow rate of coal (kg/h)Primary stream(base case operation conditionsReburning operation conditionsMass flow rate of coal (kg/h)CHa% total thermal input)Mass flow of primary air(kg/h)Reburn zone srFurnace outlet sraass flow of secondary air(kg/h)78.44Secondary air temperature(K)SR. stoichiometric ratio.Primary zone SrReburn streamIn additional, the position of reburning fuel portCHa(% total thermal input)should be consideration in the furnace design If it isVolume flow rate of CH4(m/h)2.173.08located too close to the primary zone the reburningCHA temperature(Kgas would“rolReburn zone srhe oxygen during coaldevolatilization and combustion. It will have a greatBurnout streamMass flow of burnout air (kg/h)19.3630.7543.56influence on the performance of pulverized coalBurnout air temperature(K)burnout. Meanwhile. if it is located too close to theover fired air port this would eliminate a fuel richzone for reburning. both two cases would affect theoichiometric ratNO reductionThe results of the experiments show that aAccording to the optimizedmaximum of 53% No reduction can be achievedconfiguration described above, the model developedwhen methaneected from the no 4 test holein this study is applied to simulate the optimizedand the burnout air is injected from the NO 8 test reburning case. The predicted results are comparedhole simultaneously. As listed in Table 1, theand analyzed with the experimental data to evaluateoptimized injection positions for methane andthe modelburnout air are respectively 0.95 m and 2.35 mdownstream from the burner exit plane for the3 Mathematical modelcurrent furnace configuration. The effects of themethane flow rate on no exhaust emission werealso investigated in the experiments. Increasing th3.1 Turbulence and combustion modellingflow rate of methane from 1.37 m/h to 2. 17 m/hresults in an obvious no reduction but a furtherturbulence. combustion and heat transfer in theincrease of the flow rate to 3. 08 m/h does not showsingle-burner coal combustion furnace. The modela markedly larger NO reduction. Under the present based on the finite volume discretization technique,xperimental conditions, the appropriate flow rate of solves the mass and momentum equations using themethane is 2.17 m/h(15% of the total thermal input SIMPLER algorithm. The basic predictionas shown in Table 3). A more detail description procedure involves a numerical solution of theabout the results of the base and reburninconservation equations for theexperiments can be found in reference [5]gas-phase and particle-phase, the former beingCoal was fired to test the performance of No treated in Eulerian fashion and the latter inreduction with reburning in ODF. Table 4 shows the Lagrangian fashion. The gas-phase turbulence isrelationship of reburning zone equivalence rationsimulated by the standard k-8 model. Aand no reduction rate. The results indicate thatfast-chemistry multi-mixture fraction model is usedreburning zone stoichiometric become more for combustion modelling. For the particle-turblimportant factor for NO reduction. An increase in lence interaction, the Stochastic Particle Trajectoryreburning fuel improves NO reduction. A maximum (SPT)model is used to simulate the two-phase flowNO reduction rate is about 58%. The residence timeof pulverized coal in the furnace. Coalof flow in reburning zone is another importantdevolatilization is simulated by a first-order singleparameter to reduce the No production. Choosinreaction model and char particle combustion isapproprexpressed by a kinetic and diffusion model. Forreaction under reduction atmosphere. In the case ofradiation heat transfer. the Discrete Ordinatesexperiments the below secondary air port is locatedMethod ( dom) is used to simulate thermal radiationat 0.37 m downstream of the reburning fueln the furnac中国煤化工 tions of theAssuming plug flow for the furnace, this would be aCNMHG boileer canresidence time of approximately 1.0sbe found in reJ. China Univ, of Mining Tech (English EditionVol 16 No. 33. 2 NO formation and destruction modellingreduction pathway(4)and (5) to the global NOIn this model, the amount of No produced inmodel, as shown in Fig. 4. The additional reductioncombustion is characterized by reaction kinetics andpath accounts for the no destruction in the reburnhydrokinetics using the following steady-statezone by Ch radicaltransport equatiox),CH+H,Ovolatiles evolves as hCn or NH3 intermediates andCH2+H<>CH+H2(9)the fuel bound nitrogen in the char converts to NOdirectly. The following steady-state transportCH+H H,O+HIn the present NO model, an assumed shape ofX,=LoH=k-16the probability density function(Beta PDF)in terms(17) of a normalized temperature and oxygen speciesmass fraction is used to predict the No emission inThe forward and reverse rate constants, k16 andturbulent flows for the temperature and compositionk-16, for reaction(16) can be obtained from ref. [11]The values for the reaction rate constants k4, ks andfluctuations. The PDF is used for weightingbe taken from reference [12] and the rateagainst the instantaneous rates of production of Noconstants ka, kb, and ke for different fuel types can beand subsequent integration over suitable ranges tocalculated according to ref. [13]. In this study, theobtain the mean turbulent reaction raterate constants ka, kb, and ke are calculated by uk=2.12×105T-1.4e277Rr4 Results and discussionsk=1.32×1007331-15090/7k.=1.22×10T264-7707874.1 Profile comparisonsAll the rate constants have units of m/(mol- s)For the reburning case shown in Fig. 5, theTherefore, the source terms for equations(1)and(2) profiles of predicted temperature, O2 concentrationdue to reburning reactions are given band No concentration along the furnace centrelineburn.HCN=MR,MIHCN(18)are compared with the experimental data. In thesefigures, the burner exit plane is at axial distance=4.0d INom and the exit of the furnace is at axial distance=0 mrebum. No=-(R4+R3)M(19)for keeping the same direction with Fig. I1600100000.51.01.5202.53.03.54.000.51.01.52.02.53.03.54.00.51.015202.53.03.54.0Axial distance(m)(a)Temperature(b)O2(c)NOFig 5 Comparisons of measured and predicted temperature, O? and NoAs shown in Fig 5a and Fig. 5b, the predictedFig. 5c shows the comparisons between thtemperature and O2 concentration match well withpredicted and measured NO concentration along thethe experimental data, although a small discrepancyfurnace centreline. The predicted NO concentrationis observed between the predicted and measuredshows overall good agreement with the measuredvalues. The explanation of this disparity is probablyconcentration: all the same. it must be stated that thethat the radiation heat transfer model used in ourcomputation has somewhat under-predicted the Nocomputation is not sufficiently accurate. Incorrectconcentration in the reburn zone. The accurateflow field predictionshave caused the smallprediction of NO concentration, which are simulateddifferences between the measurements andin a post-processing fashion once a fully convergedpredictions. Further, relative errors in measurementflow-field solution is obtained, are dependent on theexist in experimental data because of an unstablequality of the flow-field solution. Incorrect upstreamcoal feed rate and an equipment measurement error.flame structure predicted by the CFD-basedThese experimental errors sometimes may accountcombustion model may have caused the differencebetween measurements and predictions. Further, lessfor a large part of the observed differenceO2 than observed is predicted near the CHa injectionNevertheless, the consistency of predictions with theposition(as shown in Fig. 5b), which leads to themeasurements shown in Fig 5a and Fig 5b indicatesdecomposition of more CH4 than observed tothat the modelling methodology deployed in this hydrocarbonrather than reaction with O2 tostudy is adequate to predict the overall flow-fieldgenerate CO中国煤化工 bon radicalsproperties and combustion behaviour for thepurposes of the reburning process study.a lower predicCNMHGthen lead toobserved inJ. China Univ, of Mining Tech (English Editionl.16No.3the reburn zone However, the trend of predicted noby the computationconcentration is very similar to that measured andFig. 6 shows that the methane injection positionthe predicted NO exhaust emission is quiteis located in the region of the maximum NOconsistent with the experimental data, indicating that concentration. which is beneficial to the Nothe present computational submodel for NOreduction reactions. This predicted result provesdll the behaviour of nothat the present reburn fuel injectiItioneduction in the reburning processIn spite of the small difference between theresulting from many experiments, is optimal for thecurrent furnace configuration to obtain theomputed and measured values, the results indicatemaximum NO reduction. Fig. 6(b)also show thatthat the present reburn reaction scheme is capable ofsmall quantity of No regenerates in the vicinity ofpredicting quantitatively the no reduction levelthe burnout air admission station That is due to the4.2 Discussion on the modelling resultsoxidation of the HCn transported from the reburnzone. This result indicates that the amount ofFor the base and reburning case, Fig. 6 showsthe predicted profiles of No concentration at theinjected methane(15% of the total thermal input)issufficient for the no reduction in the reburn zonehorizontal section in the furnace. That horizontaland too much methane (20% of the total thermalsection is through the axes of reburn fuel andinput) is unfavourable to total NO reduction, whichburnout air injection tube, as well as the axis ofaccords with the results of experiments describedprimary air portpreviously. This conclusion is also consistent withFig. 6(b) shows the predicted NOexperimental observations by Cha- gger et al Lconcentration profiles for the reburning case. Thiswho reported an upper limit in hydrocarbon gasfigure, compared with Fig. 6(a), clearly reveals thatpercentage for the reburning reactiona large NO reduction is obtained in the reburn zone2Axial distance(m)(a) Base case: No(mg/m)Axial distance(m)(b)Reburning case: No (mg/m)Fig 6 Predicted NO concentration profiles for base and reburning caseComparisons of predicted NO concentration reaction has been incorporated into a comprehensiveprofiles in Fig. 6 reveal that the reduction of No coal combustion model to predict the No reductionoccurs above the reburn fuel admission station. They the reburning process. This usable andexperimental measurements shown in Fig. 5c alsocomputationally economIc submodel requirestrack this earlier NO reduction. It is a fact that a solution of only two transport equations(eq. (1)small quantity of hydrocarbon radicals is transported Eq(2)to simulate the complicated physical andupstream because of the entrainment of thechemical process inherent in the reburn technologymainstream flow and thus the no reductionProfile comparisons show that the predicmechanism is activated earlier in the computatitemperature and oxygen concentration match wellThis earlier NO reduction is also predicted bywith the measurements and the general trend ofDimitriou et al o and observed by Xu et al 9predicted NO concentration is very similar to thatThe results of mathematical modelling shown measured. The results imply that the reburn -NOin Fig 8 reveal that the accurate simulations for thesubmodel depicts quite well the observed behaviourflame structure and the turbulent mixing of theof No annihilation in the reburning processreburn fuel with mainstream flow are very importantInevitably, there are some predictive disparities dueto the no concentration predictionto the partly inaccurate prediction flame structureand to experior. but in view of the5 Conclusionscomplexity of中国煤化工 dicted, theseare small. TheCNMHGeement relyThe NO submodel including the reburningon ImprovingIcLuru we laue structure andXIANG Jun et alExperimental Research and Mathematical Modelling for Coal Reburning in Furnaceobtaining the more accurate kinetic parameters for submodel in this study provides a useful tool forthe NO submodel. We can only reiterate our opinion investigators concerned with reburning techniquethat the computationally economic reburn-NOapplications in practical combustorseferences[1] Smoot L D, Hill S C, Xu H NOx control through reburning. Prog. Energy Combust. Sci, 1998, 24: 385-408[2] Dagaut P, Lecomte F, Chevailler S, et al. Experimental and detailed kinetic modeling of nitric oxide reduction by a natural gblend in simulated reburning conditions. Combus. Sci. TechnoL., 1998, 139: 329-336[3] Chen W, Smoot L D, Hill S C, et al. Global rate expression for nitric oxide reburning. Energy Fuels, 1996, 10: 1046-1052[4] Xiang J, Sun XX, Hu S, et al. An experimental research on boiler combustion performance. Fuel Processing Technology2000,68:139-151[5 Su S, Xiang J, Hu S, et al. Experimental research on denitration by use of gas reburn technology. Journal of PowerEngineering, 2004, 24:884-888 (In Chinese)[6 Xiang J, Xiong Y H, Zheng C G, et al. Using PDF-ARRhENIUS to simulation 3-domensionally NOx formation during coalcombustion. Proceeding of the CSEE, 2002, 22: 156-160. (In Chinese)[7] Fenimore C P. Studies of fuel-nitrogen in rich flame gases. In: I7th Symposium(international)on Combustion. PittsburgThe Combustion Institute. 1979. 661-669[8] De Soete GG. Overall reaction rates of NO and N2 formation from fuel nitrogen. In: 15th Symposium(international)orCombustion. Pittsburgh, PA: The Combustion Institute, 1975, 1093-1102Xu H J, Smoot L D, Tree D R, et al. Prediction of nitric oxide destruction by advanced reburning. Energy Fuels, 2001, I541-551[10 Dimitriou D J, Kandamby N, Lockwood F C. A mathematical modeling technique for gaseous and solid fuel reburningpulverized coal combustors. Fuel, 2003, 82: 2107-2114[Il Glarborg P, Miller J A, Kee R J Kinetic modeling and sensitivity analysis on nitrogen oxide formation in well stirred reactorsCombus.Flme,1986,65:177-202[12 Bowman C T Chemistry of gaseous pollutant formation and destruction. In Fossil Fuel Combustion, Bartok W, Sarofim A. F.Ed. Canada: J. Wiley and Sons, 1991[13 Leung K M, LindstedR P Detailed kinetic modeling of C1-C3 alkane diffusion flames. CombuS. Flame, 1995, 102: 129-160[14] Chagger H K, Goddard P R, Murdoch P, et al. Effect of SO2 on the reduction of Nox by reburning with methane Fuel, 199170:1137-1142中国煤化工CNMHG