下载

1下载券

加入VIP
  • 专属下载特权
  • 现金文档折扣购买
  • VIP免费专区
  • 千万文档免费下载

上传资料

关闭

关闭

关闭

封号提示

内容

首页 2

2

2

yinfulin
2014-04-11 0人阅读 举报 0 0 暂无简介

简介:本文档为《2pdf》,可适用于高等教育领域

EnergyandBuildings()–ContentslistsavailableatSciVerseScienceDirectEnergyandBuildingsjournalhomepage:wwwelseviercomMonito(IusingsMinJeonb,CaDepartmentoeUnivbDepartmentorsity,articlKeywords:IndoorairquaMultivariatea(MANOVA)PCAPLSregressionSeasondependSubwaystatioMetrosystemollutaualityetheependThraturfromdifferegrevelopodeltheglobalmodel©ElsevierBVAllrightsreservedIntroduMillionswayormeandconvenmostconvetanareasInlargeportio,andfoeverydaysourceofouSincepestationsdaiofindooraofhazardouorganiccomlatedinthetoheavyusInadditionsystem)alsspaceduetardousindo∗CorresponEmailadd$–doi:jctionofpeopleinmetropolitanareasaredependingonsubtrosystemsfortransportationduetotheirsimplicityienceSubwaysystemshavebeenconsideredasthenientmodeofpublictransportation,inmetropoliKorea,currently,linesofmetrosystem(ofwhichanareconstructedunderground)havebeencompletedurmillionsofpeopleareusingthemetrosystemsSeoulsubwaysystemisoneofthemanditistherpresentstudyoplespendaconsiderableamountoftimeinthemetroly,therehasbeenagrowingconcernovertheregulationirquality(IAQ)insubwaystations,Varioustypesspollutants,suchasparticulatematters(PMs),volatilepounds(VOCs),andformaldehydesremainaccumusubway(especiallylocatedinunderground)spaceduee,overcrowding,andinadequateventilationsystem,theoutdoorair(ortheairenteringfromventilationoaffects,theconcentrationsofPMsinsideindoorothepresenceofPMsintheoutdoorairThesehazorairpollutantsdirectlyaffectthepassengers’comfortdingauthorress:ckyookhuackr(CYoo)aswellashealth,Studies,oneffectsofdifferentindoorairpollutantsonhumanhealthhavebeenwidelycarriedoutTherefore,toensurepassengers’goodhealth,accuratemonitoringandpredictiontechniquesarenecessaryforregulationofIAQinthemetrosystemsRecently,severalstudies–onmonitoringanddiagnosisofIAQofdeferentmetrosystemshavebeenreportedBranishasmeasuredtheconcentrationsofPMsoftheambientairandtheairinsideundergroundtrainsHehasdevelopedacorrelationbetweenthemLauetalhavemonitoreddailyvariationsofIAQinHongKongmetrostationusingstatisticalmethods,suchasclusteringanalysis(CA)Kimetal,haveusedmultivariatemonitoringmethods,suchasprincipalcomponentanalysis(PCA)andpartialleastsquare(PLS)formonitoringIAQKametalhavemonitoredconcentrationsofPMsinsideLosAngelesstationanddevelopedacorrelationforvariationofconcentrationofPMswithnumberofpassengersutilizedthefacilityTheseresearchershavemainlyfocusedonthemonitoring,butnotonthepredictionofIAQinsidemetrostationTherefore,predictionmodelsneedtobedevelopedforproperregulationcontrolofIAQinsidemetrostationHence,thedevelopmentofpredictionmodelsforPMs(PMandPM)isthecentralthemeofthepresentstudyItisalsowellknownthattheIAQofmetrosystemsheavilydependentonlocalconditions,suchasweatherchangesandseasonalvariations,Kimetal,haveconcludedthattheseasonalvariationsaffectIAQbasedonthedifferentmodelsseefrontmatter©ElsevierBVAllrightsreservedenbuildringandpredictionofindoorairqualityeasondependentmodelsgKima,BSankaraRaoa,OnYuKanga,JeongTaiKimfEnvironmentalScienceandEngineering,CenterforEnvironmentalStudies,KyungHefArchitecturalEngineering,CenterforSustainableHealthyBuildings,KyungHeeUniveeinfolity(IAQ)nalysisofvarianceentmodelsnabstractUsually,varioustypesofhazardouspmetrosystemTocontrolindoorairqpredictivemodelwhichdoesnothavcurrentlyusedInthispaper,seasondcareofseasonalchanges,areproposeofPMandPMonplatform,tempetoFebruaryareobtainedknowthequantitativemeasureofthefall,summer,andwinter)PCAandPLSglobalmodel)andfourseasons(todeofthisstudyshowthattheseasonalmlocateenbuildAQ)insubwayormetrosystemshangKyooYooa,∗ersity,Yongin,RepublicofKoreaYongin,RepublicofKoreantsremainaccumulatedinundergroundsubwaystationsof(IAQ)insubwaystations,thecontrolstrategiesbasedontheeffectoftemperatureduetoseasonalvariations,havebeendentmodelsformonitoringandpredictionofIAQ,whichtakeerealtimedataofvariouspollutants(namely,concentratione,humidityandtheconcentrationofnitrogen)duringMarchSeoulsubwaystationMANOVAtesthasbeencarriedouttorencesamongdifferentdatasetsofthreeseasons(springandssionmethodsareappliedondatasetsofoneyear(todevelopseasonalmodels)tomonitorandpredicttheIAQTheresultsscanpredictthefuturedataofPMandPMpreciselythanMKimetalEnergyandBuildings()–obtainedforoneyearandoneseasonBut,theydidnotproposeaquantitativemethodtoknowwhetherIAQdependsonseasonalvariations,fromthegivenoneyeardatasetInthefirstpartofthepresentstudy,multivariateanalysisofvariance(Mvariationsaationsonthcompared:()thedataeliminatethamethodoLater,PCofoneyeartivelyThestelemonitoofIAQwhicitoringIAQthedependmethodisuofapreviouyear(nameseasonalprallyusedterentirephasconcentratiPLSglobalathesuperioOutlinesthetheorietisticalmetintroducedtionofIAQSectionpmetrostatioTheoryMultivaMANOVvectorsaresignificantlytionhas‘n’thenavectfollowingmxlj=x¯(x¯lwherel=,(x¯l−x¯)isanofthelthpoIfthenutoone(ie,g=),thenexpressionF=(∑gl=wherenlis(Wilk’slam�∗=∣∣W∣∣BWwhere|W|aexpressionsTableFtestrejectshypothesisH:�=�=···=�g=(where�ismeanvectoroflthpopulation,calculatedusingtheexpression�l=(xl,xl,xl,···xl,n)n)atlevel˛,if(∑gi=)(√)Fp(ibutiternaally,alvasofm,thecanalvagersxterncurariabfextxtercanbHGistnvaGCanstheewGCCisdvanalvnwhbnorhissingtnthnthiultivcethtedofinmmompprojeongemePCAwithethhileationPrincisaesthintoANOVA)approachisusedtoknowwhetherseasonalffecttheIAQToknowtheinfluenceofseasonalvarieIAQ,thedatawiththefollowingtwoscenariosare()therealdatawhichhaseffectoftemperatureandobtainedaftertakingouttheeffectoftemperatureToeinfluenceofseasonaltemperaturefromtherealdata,fexternalanalysisisusedAandPLSregressionmethodsareappliedondatasetsandfourseasonstomonitorandpredicttheIAQ,respecemethodsareappliedonrealtimedataobtainedfromringsystem(TMS)inSstation,KoreaThevariablesetshhavemutualdependencyareobtainedwhilemonusingPCAmethodThesevariablesetshelptoknowencybetweendifferentIAQvariablesPLSregressionsedforpredictingIAQofaparticulardayusingthedatasdayThisprocedureisfollowedtopredictIAQforonely,globalpredictioncurve)andfourseasons(namely,edictioncurves)Theglobalpredictionmodelisgenerminology,whichmeansthesinglemodelvalidovertheespace(orentireperiod)ofthedatasetRMSEvaluesofonpredictioncurves(ofPMandPM)obtainedusingndseasonalregressionmodelsarecomparedtoknowrityoftheseasonalregressionmodeloverglobalofthispaperareasfollowsThefirstsectionintroducessofMANOVA,externalanalysisandmultivariablestahodsInSection,motivationofthepresentstudyisandtheproposedmethodsformonitoringandpredicforfourdifferentseasonsandfullyearareexplainedresentstheresultsforthedataobtainedfromSeoulnFinallytheconclusionsofthisarticleareaddressedriateanalysisofvariance(MANOVA)Aisusedtoinvestigatewhetherthepopulationmeanthesameand,ifnot,whichmeancomponentsdifferIfadatasethas‘g’numberofpopulationsapopulanumberofvectorsandavectorhaspelementsinit,or‘j’ofapopulation‘l’canbedecomposedaspertheodel:−x¯)(xlj−x¯l),,gandj=,,,nx¯isanoverallvectormean,estimatedtreatmenteffect(wherex¯lisavectormeanpulation),and(xlj−x¯l)isresidualmberofelementsinavectoraregreaterthanorequalp≥)andthenumberofpopulationsarethree(ie,valueofFteststatisticiscalculatedbythefollowing:nl−p−p)(−√�∗√�∗)()thenumberofvectorsinthelthpopulationand�*bda)iscalculatedbythefollowingexpression:∣∣∣∣()nd|BW|aretheabsolutevaluesofWandBWTheforcalculatingmatricesWandBWareprovidedinF=whereFdistrExUsuexterntrationHencePM)externpassenthaneandairmainvceptsoIneables)X=Gwhereofmaiparts,andEiHcanbE=H−whereAabnormablesifilteraIntremovatureiPMiMSincorrelausedtThecocipalcTheseshipamonlyonstudymatrixthesemdatawinformPCAcapturposednl−p−p−�∗√�∗>Fp,(∑gi−ni−p−)(˛)()∑gi=nl−p−)(˛)istheupper(˛)thpercentileoftheonwithpand(∑gi=nl−p−)degreesoffreedomlanalysisprocessvariablescanbeclassifiedintothreegroups,riables,mainvariablesandothervariablesTheconcenainvariablesdependonexternalandothervariablesexpressionsforthemainvariables(suchasPMandbesubdividedintotwoparts:(a)thepartaffectedbyriablessuchastemperature,humidityandnumberofand(b)thepartaffectedbytheothervariables(otheralvariables)suchasventilationsystem,airfiltrationtains,Now,externalanalysisiscarriedoutonlestoremovetheeffectofexternalvariablesTheconernalanalysisaregivenbelow:nalanalysis,datamatrixX(ofmainandexternalvariewrittenascombinationofexternalandmainvariableshematrixofexternalvariablesdataandHisthematrixriablesdataThemainmatrixHcanbesplitintotwodE,whereGCisapartaffectedbyexternalvariablespartaffectedbyothervariablesThus,themainmatrixrittenasthecoefficientmatrixcalculatedusingC=(GTG)−GTHtagesofcarryingoutexternalanalysisare:detectingariationsinmainvariables,locatingthegroupofvariichcausetheabnormalvariationinmainvariables,tomalvariationsinthemainvariablestudy,wefocusedonEmatrix,whichisobtainedbyheeffectofchangesinexternalvariables(ie,temperisstudy)onchangesinmainvariables(ie,PMandsstudy)ariatestatisticalanalysiseconcentrationsofindoorairpollutantsarehighlywitheachother,multivariatestatisticalmethodsaredthedependencyamongdifferentvariablesofIAQonlyusedmultivariatestatisticalmethodsareprinonentanalysis(PCA)andpartialleastsquares(PLS)ctionmethodsgenerallyusedtoexplaintherelationdifferentvariablesusingcomplexdatasetsHowever,thod,namely,PCAisusedformonitoringinthepresentandPLSareusedtoobtainthevariance–covariancesmallnumberofprincipalcomponents(PCs)Thus,odshelptoreducethedimensionalityoftheoriginalretainingimportantinformationtodisplaythedatainaformatthatcanbeeasilyinterpretedipalcomponentanalysis(PCA)noptimaldimensionalityreductiontechniquewhichevarianceofadatasetTheoriginalmatrixXisdecomaprocessandnoisesubspaces:thematrixXcanbeMKimetalEnergyandBuildings()–TableMatricesBandWrequiredforcalculationof�*(seeEq())SourceofvariationMatrixofsumofsquareandcrossproducts(SSP)Degreesoffreedom(df)Treatmentg∑Residual(erTotal(correcwrittenastheresiduaX=TPTEwheretiisashipbetweeinformationthenumberInpractitoexplainmcanfacilitattaneouslycminimizatioPCAtoestimanalysis,anPartiaPLSmulfromthedaforaspartiinput)variaPLSregresstheinformacovariancePLSregrXandYThusingfollowX=TPTEY=UQTFwherepantherelationrespectivelyTandUaredifferentsamodelwhicY=X·LVwhereBisgMateriaMotivaConstrucisadifficultfocresedealicheslarpatedingmlemsToata(orcfindaation)neesbeevariaaeproropoinFingmsaretoSctedataisfall,aininamerecemedatasetionl,sumorecarrietemrpolltemperaB=l=nl(x¯l−x¯)(x¯l−x¯)Tror)W=g∑l=n∑j=(xlj−x¯l)(xlj−x¯l)Ttedforthemean)BW−g∑l=n∑j=(xlj−x¯)(xlj−x¯)TthesumoftheouterproductofvectorstiandpiandlmatrixE:=n∑i=ltipTiE()scorevectorthatcontainsinformationontherelationndifferentsamples,piisaloadingvectorthatcontainsontherelationbetweendifferentvariablesand‘n’isofindependentvariablesce,onlyafewprincipalcomponents(PCs)aresufficientostofthedataApplicationofPCAmodelonthedataetideterminetheoptimumnumberofPCsbysimulonsideringthereductioninthedimensionalityandthenoflossofdatainformationSeveraltechniquesexistinatethenumberofPCs,includingscreeplotting,paralleldcrossvalidation,lleastsquare(PLS)regressionmodeltivariateregressionmethodisusedtoframeamodeltaset,whichinturnhelpsforpredictionofthedataculardayThismethodisusedtocorrelateprocess(orbles(X)andasetofoutput(orresponse)variables(Y)ionreducesthedimensionsofXandYbyprojectingtionintoalowdimensionalspacethatmaximizesthebetweenthem,essionbuildsalinearmodelwhichrelatesthematricesesematricescanbedecomposedintobilineartermsingequations:=n∑i=ltipTiE()=n∑i=luiqTiF()dqaretheloadingvectorsthatcontaininformationonshipbetweendifferentprocessandoutputvariables,,nisthevectorofthenumberoflatentvariables(LVs),scorematrices(whichhastheinformationbetweenmples)andEandFaretheresidualsThePLSregressionhrelatesXandYcanbeexpressedinasusuallytorepwhenapproaparticuformuloperatmultipregionoutputmodelTowaystmodeltemhastudy,andPMThApshownoutusmodelappliediscolleyeardspringforobtMayforsumfromDthatthThedapopulaandfalBefsisisc(duetothataiduetosis,temF()ivenbyB=W(PTW)−QT,inwhichW=(XTYq)T–lsandmethodstionofthisstudytingmodelsfromtimeserieswithnontrivialdynamicsproblemConventionaldatabasedmodelingmethodsofPManrespectivelMANOVationsaffecpopulationforwinter)fallistakenpopulatioobservedTthedatawg−g∑l=n−gg∑l=nl−usonglobalapproaches,whichadoptasinglemodelnttheinput–outputbehaviorofsystemHowever,ngwithlargedatasetsfromenvironmental,theglobalareinefficienttocharacterizethedatabelongstoaeriod(orduration)andaredifficulttooptimizetheproblemandupdatethedataonlineoncetheprocessodechangesOntheotherhand,localapproachesuseodels,whicharevalidincertainlocalizedoperatingestimateandproduceanaccurateforecastofsystemgiventime,localmodelsarechosenaccordingtotheriteria)definedontheinputdata–naccurateIAQpredictionmodelinundergroundsubs,aseasondependentlocalmodel(namely,seasonaldstobedeveloped,sincetheIAQofthemetrosysninfluencedbyseasonaltemperaturevariationsInthistionsofseasonaltemperatureonconcentrationofPMrestudiedandvalidseasonalmodelsaredevelopedposedmethodsedframeworkformonitoringandpredictionofIAQisigMonitoringofseasonalchangesinIAQiscarriedultivariateanalysisofvariance(MANOVA)PredictionusedtocarryoutpredictionofIAQThesemethodsaresubwaystation,inKoreaThedatausedinthisstudyfromMarch,toFebruary,Thewholeonedividedintothreeseasonaldatasets,namely,oneforoneforsummer,andtheotherforwinterThedurationsgdatasetsforspringandfallarefromMarchtondfromOctobertoNovember,respectively,andwinterarefromJunetoSeptemberandbertoFebruary,respectivelyItistobenotetasetsofspringandfallaretreatedasonepopulationtsofsummerandwinteraretreatedasndandrds,respectivelyNumberofsamplesobtainedforspringmerandsinterare,and,respectivelyarryingouttheMANOVAtestonIAQ,externalanalydoutforremovingtheinfluenceofseasonalchangesperaturevariations)fromtheIAQdataItiswellknownutants,especiallyconcentrationofPMs,willgetaffectedperaturechangesTherefore,intheexternalanalytureisreferredasexternalvariableandconcentrationsdPM(whichhasdiameterswith�mand�m,y)arereferredasmainvariablesAtestsarecarriedouttoknowwhetherseasonalvaritIAQForMANOVA,oneyeardataisdividedintothrees,oneforspringandfall,oneforsummer,andtheotherItistobenotedthatthedataoftwoseasonsspringandasonepopulationSignificantdifferencesamongthesensaccordingtochangesinseasonaltemperatureareoknowtheinfluenceofseasonalvariationsontheIAQ,iththefollowingtwoscenariosarecompared:()theMKimetalEnergyandBuildings()–redicrealdatawhaftertakingnalanalysisairpollutanIAQmodelFinally,monitoringeffectofvarmodelsofIoneforsumthemonitopredictionmtionmodelsstudyarethvariablesarcriteriaforstudyarethofPMs(t−thetempermodels(thrIAQmodel(ofseasonalSeoulsTheobjestationonlairpollutanSeoulsubwplatformanlutants(NOhumiditywofNO,NOofnitrooxsuredby�utionsobtecifiremeinTadailluesFigFrameworkformonitoringandpichhaseffectoftemperatureand()thedataobtainedouttheeffectoftemperatureintherealdatabyexterEvaluationofinfluenceoftemperatureontheindoortsofthemetrosystemisrequiredtoproposeseasonalmultivariatestatisticalmethodsareusedtoproposeandpredictionmodelsforIAQ,whichtakescareofiationsindifferentseasonsMonitoringandpredictiondistriblengthThespmeasusentedTheMarchagevaAQaredevelopedforthreeseasons,oneforspringfall,mer,andtheotherforwinterPCAisusedtodevelopringmodels,whilePLSregressionisusedtodevelopodelsforthreeseasonsInthedevelopmentofpredicofIAQ,output(orresponse)variable(Y)takenforthisecurrentconcentrationsofPM(t)andPM(t)TheseeconsideredsinceconcentrationsofPMsaretakenasmonitoringIAQInputvariables(X)takenforthisecurrentconcentrationofnitrate,previousdaydata),temperatureandhumidityToverifytheaccuracyofaturedependentmodels,theperformanceofseasonaleemodelsforyearperiod)arecomparedwithaglobalonemodelforyearperio

用户评价(0)

关闭

新课改视野下建构高中语文教学实验成果报告(32KB)

抱歉,积分不足下载失败,请稍后再试!

提示

试读已结束,如需要继续阅读或者下载,敬请购买!

文档小程序码

使用微信“扫一扫”扫码寻找文档

1

打开微信

2

扫描小程序码

3

发布寻找信息

4

等待寻找结果

我知道了
评分:

/8

2

VIP

在线
客服

免费
邮箱

爱问共享资料服务号

扫描关注领取更多福利