大型燃煤电站锅炉低NOx燃烧优化运行的研究

上海交通大学硕士学位论文大型燃煤电站锅炉低NOx燃烧优化运行的研究姓名：蒋欣军申请学位级别：硕士专业：动力工程指导教师：罗永浩;李秋白20071101IIINOx300MW#7330MWBPNOxIVDCSNOxBPVCOMBUSTIONOPTIMIZATIONOPERATIONOFHIGHCAPACITYCOAL-FIREDPOWERSTATIONBOILERABSTRACTBoilersareoneofthethreemainframesofcoal-firedgenerations.Atpresent,coal-firedboilersondutywith300MWloadhaveproblemsofhighemissionsofpollutantandlowcombustionefficiencyinChinabecauseofself-defectsofequipmentsandlacksofexcellentoperatingtechniques.toimproveboilerefficiencyandsavelimitedcoalresources,reduceemissionsofpollutantproducedinprocessoffiringcoalareurgenttorealizecontinuabledevelopmentofenergyresourcesinChinaaswellasaproblemtobeovercomebyelectricalscientistsandtechnicians.Focusingonthe330MWcoal-firedboilerofNo.7unitofJianbipowerplant,thispaperdetailedlystudiedNOxformationmechanismsandsomemainfactorsinfluencingitsemissions,andanalizedrelationsofcontrollingrulesbetweenreducingNOxemissionsandimprovingboilerefficiencybycombustionmodification.Atthebasisoftheabovestudy,acombustionmodificationhotexperimentwascarriedout.Accordingtoeveryexperimentwithdifferentworkingconditions,characteristicrelationbetweenNOxemissionsandsomeoperatingvariableswasanalized.Someproblemswhichexistintheboiler’scurrentoperatingconditionsarelisted.Andsomesuggestionisgiven.Theutilityboilerinstallationishuge.Ithascharacteristicsofmultivariate,longVIlaggingandnonlinear.Itisalmostimpossibletopredictoutputsofcombustionprocessinrealtimebysometraditionalphysicalandchemicalmodelsbecauseofcomplicationofburningpulverizedcoalinboiler’sfurnace.Inordertosolvethisproblem,theartificialneuralnetworkcouldbeutilized.Inthispaper,arealtimepredictingmodelofNOxemissionsandboilerefficiencyisbuiltbyusingBPneuralnetwork.Samplesaredatasaquiredfromtheaboveexperiment.CombininggeneticalgorithmsusingnonlinearoptimizationsearchingtechnologywiththeBPneuralnetworkmodel,someoptimumoperatingplansunderdifferentcoalwascalculated.Resultsofhighefficiencyandlowpollutantcouldbeaquirediftheboileroperatesusingtheplan.TheoperationplancaneitherguideoperatorstooperatetheboilerorbecombinedwithDCSsupplingbasisofmodelforachievinglowNOxcombustionoptimizationclosed-loopoperation.KEYWORDS:Generationboiler,combustionoptimization,NOx,boilerefficiency,neuralnetwork,BPalgorithms,Bayesgeneralization,optimizationsearching,geneticalgorithms702008121871200812182008121811.11.1.120062450NOx67NOx2010NOx19952292020NOx3500/[1]1.1.2NOxNOxNOxNONO2N2ON2O3NOxNOx[2-5]1)NOxHNO3SO2,4000151.150002)NOxNO2CHOPANH2SO4PAN3)NONOHbCONO2SO2NO2NO4-5NO2NO+O2=2NO2NO28100ppmNO2NO24N2O5N2ON2OCO2CO21.1.3NOxNOxNOx19968GB13223-19961997111000t/hNOx650mg/m31000mg/m31997NOx650mg/m3400500mg/m320047120NOx450mg/m3NOx1kgNOx0.6320077NOx1.26/kgNOx1.2#7NOxBPNOxNOxNOx/2NOxNOxNOx3#73NOx4BP3NOx5BPNOxNOx6NOx4NOx2.1NOx2.1.1NOxNONO2N2ONH3HCNN2ONOx10%90%NOxNOx1NOx(ThermalNOx)NOx1500NOxNOx2NOx(PromptNOx)CHNOxNOx3NOx(FuelNOx)NOxNOx21NOxNOxNOx60~80%NOx21NOxFig21.AmountofNOxfordifferentformationpaths3NOx(mg/Nm)18001400100050010001500ThermalNOxPromptNOxFuelNOxFurnaceTemperatrue(C)NOx51500NOx20%NOx[7]NONO2NO90%NO2510%N2O1%[8]NONOxNOxNOxNOxNOxNOxNOx2.1.2NOxNN2NOxNOxNOxNOxNOxNOxN2OOHO2NOx(Zeldovich)[9]NNONO+⇔+2(21)ONOON+⇔+2(22)ZeldovichNOx23ZeldovichHNOOHN+⇔+(23)NOx[7](21)(21)NOxZeldovichZeldovichNOx24NOxN(24)NOxNOxCOCCO2OH(23)NOxNON2++−=−−−][][][1][][][][2][32212222121OHkOkNOkOkNOkkNkOdtNOdNOx694.5107J/mol1200NNONO+⇔+2NOx1500NOx2000NOxNOx25%[10]NOx(Zeldovich)NOxNOx1500NOx1500NOx[11]NOxNOxNOxNOxNOxNONO(21)NOO(21)NONONOxNOxNOxO2N2NOx2.1.31971FenimoreNOxNOxCHiN2NOxNOxNOxNOxNOxNOx7NOxNOxNOxNOxHCNHCNCNNHN[9]NHCNNCH+→+2(25)NHHCNNCH+→+22(26)NCNHNCH+→+222(27)NCNNC+→+2(28)CNNC222→+(29)CHCH2(25)(26)HCNNOxNON290%NOxHCNNOxNOxCHCH2NOxNOxNOxNOxNOxNOxNOxNOxNOxNOxNOx[12]NOxNOxCOH2NOxNOxNOx2.1.4NOx822NOxFig22.FormationoffuelNOxtopNvalleyNcenterN23NFig23.SchemeofNatomslocationincoal0.82%[13]C5H5NC5H5NH2CN[(25.363)107Jmol]NN(94.5107Jmol)NONOxNOx2040%NON222NOx75~90%NOx[14]NOxNOxNOxNOx[15]NPyrrolicN5PyridinicN6QuaternaryNQ[16]N5N6NQNQN6NQNtopNvalleyNcenterN23topNNOvalleyNcenterNtopNNOxNNNNOx[15]NHCNNH3NNNNOxN2NNOxN2NOx924WendtNOxFig24.FuelNOxformationmodelprovidedbyWendtetc.12-4WendtNOxNOxNOxNNNNNN/NNNN/NNOxNOxNOxNOx[8,13,16,17]HCNNH3HCNNH3HCNNH3[17]NNOx[18]][][3QNNH=(210)]/[][][][CNPNPHCNrd+∝(211)]/[][][][][CNPNPQNVolNrd++=(212)][QN][NPd][NPrQuaternaryNQPyrrolicN5PyridinicN610%N[19]HCNNH3NNQNH3N5N6NHCNHCNNH3HCNHCNNH3[20]HCNNH3O/NO/NHCNNH3HCNNH3[9]HCNCoalfuelNCharNVolatileNTarNNH3NON2NOx1021.HCNNH3Table2-1.ThemainreactionequationsandreactionvelocityofHCNandNH3AE(J/gmol)Λ+→+NOOHCN21.01010280300Λ+→+2NNOHCN3.001012251000Λ+→3NHHCN1.941015328500Λ+→+NOONH234.00106133900Λ+→+23NNONH1.801014113000HCNNH3NOxNOxHCNNH3NOxN2NOxNOxNOx1NOxCHiNOxNOxNOxNOx2NNNON2NOON→+2)((213)ΛΛ+→+2)(NNON(214)PershingWendtNNNOxN[21]()cccySRxβ+−=1132.10.2(215)cxNNOcyNcβSR(215)NNONOx1110ms300msNOx2030%NNOxNOxNOxNOxNOxCO[14]2.1.5NOxNOxNOxHCNN2NOxNOxNOx[22]2NHCNNOCHi→+→+Λ(216)CHiNOxNOxNOxNOxCH3CHNOHCNHHHCNNOCH++→+3HCNNOxNOxNOxNOxNOxNOx2.2NOxNOx[23]NOx[23,24]NOxN2NOxSCRSNCRNOx[25,26,27][28,29]2.2.1NOxN