Latin American applied research.docx

Latin American applied research.docx

《Latin American applied research.docx》由会员分享，可在线阅读，更多相关《Latin American applied research.docx（12页珍藏版）》请在冰豆网上搜索。

LatinAmericanappliedresearch

窗体顶端

窗体底端

窗体顶端

∙

CitadoporSciELO

∙

CitadoporGoogle

∙

SimilaresenSciELO

∙

SimilaresenGoogle

LatinAmericanappliedresearch

versión ISSN0327-0793

Lat.Am.appl.res. v.34 n.4 BahíaBlanca oct./dic. 2004

Controlofpusherfurnacesforsteelslabreheatingusinganumericalmodel

P.Marino1,A.Pignotti2andD.Solis3

CentrodeInvestigaciónIndustrial,FUDETEC,2804Campana,Argentina

1sidmrp@

2api@

3sidpsol@

Abstract−Steelslabsarereheatedinpusher-typefurnacesuptoatemperatureof1200oCinthesteelsheetmanufacturingprocess.Inthisarticlewedescribeacontrolsystemthatusesanon-linenumericalmodeltocalculatethefurnacesetpointsinordertoimprovetheheatingquality.Examplesofactualfurnaceoperationwithandwithoutthesystemarepresentedtoshowtheimprovementsthatareobtainedhandlingtypicalnon-stationarysituations.

Keywords−SteelIndustry.ReheatingFurnaces.FurnaceControl.

I.INTRODUCTION

Inthesteelstripmanufacturingprocess（Figure1）,steelslabsobtainedfromcontinuouscastingarereheateduptotemperaturesofapproximately1200oCpriortotherollingprocess.Therequiredtemperatureattheendofsuchprocesshastobecomprehendedwithinanarrowrangedeterminedbythesubsequenton-lineheattreatmentprocess.Slabreheatinginpusherfurnacesisoneofthesourcesofvariabilitythatproducedeparturesfromthatnarrowrange.

InthecaseofSIDERAR'shotrollingfacilityinSanNicolás,Argentina,fourpusher-typefurnacesareusedtoreheatslabsthatareapproximately6mlong,between0.65and1.53mwide,andfrom0.18to0.20mthick.Thesefurnacesarenamedafterthewaytheslabsarepushedforwardinsidethefurnace.Everytimeahotslabhastobedischargedtoberolledanewslabisintroducedintothefurnaceandtheintermediateslabsarepushedsidewaystowardsthefurnaceoutlet.Inthefirstpartofthefurnacetheslabsaresupportedbyfourrefrigeratedskids,whileneartheoutlettheylieonarefractoryhearththatisintendedtodiminishthetemperatureinhomogeneitygeneratedbytheskidsTheheatingpowerissuppliedbygasburnersthatuseeithernaturalgasoramixtureofnaturalandcokegasesandarearrangedinseveralzones.Typicallyonepreheatingzone,twoheatingzones（anupperandalowerone）andonesoakingzonearepresent（Figure2）.Theburnersofeachzonearecontrolledthroughthermocouplesetpoints:

acontrolloopregulatestheairandgasflowratestomatchthesetvaluewiththetemperaturemeasuredbyaproperlyplacedzonethermocouple.Therefore,theproblemoffurnacetemperaturecontrolisthatofspecifyingthesetpointsthatproduceanadequateslaboutlettemperaturedistribution.

Tomonitortheslaboutlettemperaturethereisaninfraredpyrometerattherougherexit（R4inFigure1）,whichmeasurestheslablongitudinaltemperatureprofileontheuppersideoftheslab.Themeantemperatureandthemaximumtemperaturedifferenceofthisprofilearethetargetvariablesofthefurnacecontrolanddefinetheheatingquality.Althoughitwouldbedesirabletohaveameasurementpointclosertothefurnaceoutlet,theoxidelayerthatisformedduringtheheatingprocessandthatisremovedbyadescalerattherougherinlet,preventsareliablemeasurementpriortotherougherexit.

Figure1.Schematicillustrationofthesteelstripmanufacturingprocess

Traditionallythefurnacesareoperatedmanually,basedonsetpointtablesthatcorrespondtosteadystateoperation.TherearealsosomeautomaticactionsthatareimplementedintheProgrammableLogicController（PLC）thatisusedtohandlethesignalsfromtheprocesssensorsandtoregulatetheprocessactuators.Althoughmanualoperationgivesareasonableheatingqualitywhenthefurnaceisinsteadystate,thereusuallyarechangesintheslabgeometry,inthecycletime,intheinlettemperature,andtherearedownstreameventsthatproducehaltsintheline.Allthesesituationsproducedeparturesfromstationaryoperationandgeneratevariationsintheslabmeantemperatureattherougheroutlet.Forinstance,toavoidslaboverheatinginmanualoperation,wheneverahaltoccurs,thegasandairflowratesareautomaticallydecreasedbythePLC.Whenoperationisresumed,compensatingforthiseffectisadelicatetaskthatonlyexperiencedoperatorsareabletocarryoutwithrelativesuccess.

Inthisarticlewedescribeacontrolsystembasedonanumericalmodeloftheprocessthatisusedtoautomaticallycalculatezonetemperaturesetpointsthatareintendedtominimizethedeparturesoftheslabmeantemperatureattherougherexitfromthecorrespondingprocessobjectives.

II.NUMERICALMODELS

Theuseofnumericalmodelstoimprovethedesignofthisprocesshasbeentheobjectiveofseveralanalyses.Averycompletesteadystatemodelwaspresentedin（Barr,1995）.Thatmodelwasusedtostudytheinfluenceoftheskidconfigurationontheslabhomogeneity.

Duetothefactthatthemodelsareusedinafactoryenvironment,theuseofstandardPChardwareismandatory,thusrestrictingthemodelcomplexity.However,thecontinuousincreaseinthecomputingpowerhasrecentlyallowedthedevelopmentofdetailednumericalmodelscapableofperformon-line.

（Correiaetal.,2002）investigateparametricallytheuseof2-Dzonemodelstopredictthethermalbehaviorofacontinuouslyoperatedmetalreheatingfurnace.In（Boineauetal.,2002）aCFDcodewithamoduletocalculatetheradiativeexchangesusingazoneformulationwasadaptedtosimulatetransientseliminatingthefluiddynamiccalculation.Althoughaseriesofoff-lineanalyseswaspresented,on-lineresultswerenotgiven.Alsoin（Honneretal.,2002）,aCFDmodelwasusedtocalibrateasimpleronewhichusesadjustablecoefficientstoevaluatetheradiativeandconvectiveheatfluxes.On-lineresultsshowedagoodagreementbetweencalculatedandmeasuredvalues.

Inapreviouspaper（Marinoetal.,2002）wepresentedadetailednumericalmodeloftheslabreheatinginpusher-typefurnacesandshowedthatthetemperaturescalculatedbythemodelareinagreementwithvalidationmeasurementsmadewithinstrumentedslabs.ModelresultswerealsosuccessfullycomparedwiththepyrometermeasurementsatR4.Themainmodelfeaturesare:

∙3-dimensionalandspectralcalculationoftheradiativeexchangesinthecombustionchamberusingthezonemethod（HottelandSarofim,1967）

∙DetaileddescriptionoftheradiativepropertiesofthecombustionproductsfromRADCAL（Grosshandler,1993）

∙Thecombustionproducttemperaturesarecalculatedfromthermocouplemeasurements

∙2-dimensionalcalculationoftheslabtemperaturedistribution（neglectinginhomogeneitiesintheslabwidth）

Similarfeaturesarepresentinmodelsdevelopedfordifferentkindoffurnaces（MarinoandPignotti,1997;Altschuleretal.,2000;Marino,2000）.

III.FURNACECONTROL

Inthissectionwedescribethealgorithmthatisusedalongwiththemodeltoautomaticallycalculatethezonetemperaturesetpointsthatareintendedtominimizethedeparturesofthemeanslabtemperaturefromthetargetvalue.Thisalgorithmincludestheevaluationoftheeffectofchangesinthezonetemperaturesontheslabfuturethermalevolution.Duetothefactthatboththemodelandthecontrolalgorithmhavetoperforminrealtime,andthatthiscalculationhastobeupdatedfrequentlyinordertotakeintoaccountpossiblechangesintheloadgeometry,cycletime,thermocouplemeasurements,etc.,thereareboundsonthecomplexityofthealgorithmusedtocalculatethetemperaturesetpoints.Thecurrentpracticeistoperformthiscalculationevery30seconds.

A.ControlAlgorithm

Inthecontrolalgorithmatargetaverageslabtemperatureattherougherexitisdefinedforeveryslabinthefurnace,accordingtotheproductandprocessrequirements.Itdependsontheslabgeometryandthestripfinalthickness.Intermediatetargettemperaturesattheendofthepreheatingandheatingzonesarealsodefined.

Figure2.LongitudinalsectionofSIDERAR#3pusherfurnace

Foreachzonetheupdatedsetpointtemperatureisdeterminedbycomparingthetargettemperatureattheendofthezone,andthemodelpredictedaverageslabtemperaturewhenitreachestheendofthezone.Inpractice,thefollowingequationissolved:

=0

where

Wi:

isaweighingfactorthatdependsontheslabdistancetotheendofthezoneandonthedepartureofthecurrentslabtemperaturefromitsdesiredstationaryvalue

Tci:

isthecalculatedmeantemperaturethattheslabiwillhaveattheendofthezoneundertheassumptionthattheslabvelocityremainsconstant.Thisvalueisafunctionoftheupdatedzonesetpointtemperature

Tobji:

isthetargettemperatureofslabiattheendofthezone

Thesummationcomprisesalltheslabsofthezoneandalsopartofthosethatareintheprecedingzone.Boththeobtainedzonetemperaturesandtheirratesofchangearelimitedinordertopreservetheintegrityoftherefractories.

B.TargetOffset

Asthemodelcalculatesthemeanslabtemperatureatthefurnaceoutlet,buttheheatingprocessreferencetemperatureismeasuredwithapyrometerontheslabuppersurfaceaftertherougherstand,thereisadropinthemeanslabtemperaturebetweenthefurnaceandtherougherexit.Thisdropmayvarydependingonthetransferencetime,theperformanceofthedescaler,andtherefrigerationoftherollingcylinders.Thus,acorrectionterm（offset）isintroducedforthetargettemperatureatthefurnaceoutlet.Thistermisintendednotonlytocompensatefordeparturesfromaconstanttemperaturedropfromthefurnaceout