gasestokyohotn0744

ANOVERVIEWOFTHEIODINEBEHAVIOURINTHETWOFIRSTPHEBUSTESTSFPT-0ANDFPT-1D.
Jacquemain,N.
Hanniet,C.
PoletikoInstitutdeProtectionetdeSretéNucléaireIPSN/DRS-C.
E.
Cadarache13108StPaulLezDuranceCédex,FranceR.
CrippsAccidentProgressionSectionPaulScherrerInstituteCH-5232Villigen,SwitzerlandC.
WrenResearchChemistryBranchAECL/WhiteshellLaboratoriesPinawa,Manitoba,R0E1L0,CANADAD.
PowersSandiaNationalLaboratoriesNuclearFacilitiesSafetyO.
Box5800,MailStop0744,Albuquerque,NM87185,USAY.
DrossinosCECJointResearchCenter,21020,ISPRA,ITALYF.
FunkeNuclearPowerGenerationDivision,SIEMENS/KWU,D-91058Erlangen,GERMANYB.
HerreroCIEMATMadridAvda.
Complutense22-28040Madrid,SPAINABSTRACTThePHEBUSF.
P.
programisawideinternationalefforttoinvestigate,throughaseriesofin-pileintegralexperiments,LWRsevereaccidentphenomena,inparticularbundledegradationandthesubsequentreleaseandtransportofradioactivematerialsuptothecontainment.
Twotestssimulatingalowpressurecoldlegbreakunderasteamrichenvironmenthavealreadybeensuccessfullyperformed:FPT-0withtraceirradiatedfuelandFPT-1withre-irradiatedBR3fuel.
Bothtestshaveprovidedexperimentaldataofhighinterest,particularlyconcerningiodinereleasefromthefuelbundle,transportinthereactorcoolantsystemandbehaviourinthecontainment.
TheanalysisofFPT-1,currentlyinprogress,showsthattheresultsexhibitcommonfeaturesanddifferenceswiththeprevioustestFPT-0.
Inbothtests,highestgaseousiodinefractionsareobtainedafterZicaloy-oxidationphases;theamountmeasuredinFPT-1,referredtotheinitialbundleinventory,ishoweversignificantlylowerthaninFPT-0.
ThefastdepositionofiodineinthesumpasinsolubleAgIobservedinFPT-0appearstobeconsiderablysloweddowninFPT-1toapointthatiodinere-volatilisationfromthesump,whichwasinhibitedinFPT-0,becomespossibleinFPT-1.
Bothtestsresultsyieldindicationthatthegaseousiodineonthelongtermisproducedbyreleaseoforganiciodidesfromthepaintedcondensersurfaces.
ThedataobtainedfromthePHEBUStestsindicatethatthecurrentmodellinginiodinechemistrycodesisnotfullyadequatetomodeltheobservedphenomena.
Thecomparisonofmodellingandexperimentaldatahasidentifiedareasofnecessarydevelopments:modellingofthermodynamicsandkineticsofchemicalreactionsinthereactorcoolantsystem,ofkineticsoforganiciodidesformation/decompositionandofAg/Iinteractionsinthecontainment.
ThelaterrequiringamoreaccurateknowledgeoftheAgphysicalandchemicalstateinthecontainment(surfacecharacteristicsandoxidationstate).
ImplicationofPHEBUSresultsforLWRiodinesourcetermstudiesarebrieflydiscussed.
INTRODUCTIONThePHEBUSFPprogramoffersthefirstopportunity,throughaseriesofin-pileintegralexperiments,forevaluationoffissionproductsbehaviourinthepresenceofmixesofradioactiveandnon-radioactivematerials(structuralandcontrolrodmaterials),gaseffluents(H2O/H2mixtures),radiationfields,temperaturesandpressuresprototypicofLWRsevereaccidentconditions(Krischer,1992;VonderHardt,1994).
ThefirstPHEBUStest,FPT0,performedinDecember1993withtrace-irradiatedfuelinthepresenceofsilver-indium-cadmiumcontrolrodmaterialandessentiallyunderoxidisingconditionswasintendedprimarilyasameanstoevaluatetheadequacyofthetestfacilityandtestprocedurestobeusedthroughouttherestofthePHEBUStestprogramtostudyfissionproductbehaviour.
TheFPT0testhas,however,providedinferencesandsuggestionsaboutissuesthatshouldbepursuedinfuturetestsintheprogram.
Thiscertainlyistrueinthecaseofiodinechemistryinthecircuitandinthecontainment(Jacquemain,1997).
ThesecondtestFPT1performedinJuly1996,withatestbundledesignanddegradationscenariosimilartothatofFPT-0,shouldallowtocheckthattheun-predictedphenomenaputinlightinFPT0arereproducedwithBR3re-irradiatedfuel,i.
e.
:-theexistenceofsignificantamountofgaseousiodineinthecoldlegoftheprimarycircuitandinthecontainmentearlyduringthedegradationphase,-thecorrelationbetweenbundledegradationevents(fueloxidation,materialre-locationpoolformation)andthephysicalformandamountofiodinetransportedthroughtheprimarycircuit,-thelowiodineretentionintheprimarycircuitpipesbutthehighaffinityofiodineforthecontainmentwallsbothinthesumpandintheatmosphere,-theroleplayedbycontrolrodandstructuralmaterialiniodinetransporttothecontainment,-thepresenceofAginlargeexcessrelativetoiodineinthecontainmentandtheformationofinsolubleandnon-volatileiodineinthesump.
ThispaperpresentsanoverviewoftheiodinedatacollectedduringthePHEBUStests,insistingonthegaseousiodineresultsintheexperimentalcircuitsandinthecontainment.
ResultsconcerningotheraspectsofPHEBUSresultsaredescribedmoreextensivelyinpaperspublishedelsewhere(Clément,1998;Hanniet,1998).
ThefirstsectionofthepaperwillbedevotedtoashortdescriptionofthePHEBUStestfacilityandinstrumentationdedicatedtofissionproductsandgaseousiodinemeasurements.
ThesecondandthirdsectionwillgiveadescriptionofthemainresultsobtainedinthePHEBUScircuitandcontainmentconcerninggaseousiodine,insistingonthereleasedmassandkineticsofreleaseandtheavailableinformationexistingontheiodineformsinthecircuitandthecontainment.
ThefourthsectionwillpresentcomparisonofmodellingandexperimentaldataperformedwiththeiodinechemistrycodeIODE4.
2developedatIPSN.
ThelastsectionwilldiscussimplicationsofPHEBUStestsresultsforLWRiodinesourcetermstudies.
THEPHEBUSFACILITYANDINSTRUMENTATIONDescriptionofthefacilityWerecallheretheinformationnecessaryfortheunderstandingofthepaper.
ForamorecompletedescriptionconcerningthetestfacilityandthePHEBUS-FPproject,seeSchwarz(1998).
Thefacilitywasdesignedtosimulatethereactorcore,primarycircuitwithasteamgeneratorandthereactorcontainmentofaLWR,thescaling-downfactorbeing,forthethreeparts,about1/5000ascomparedtoa900MWeLWR(figure1).
Thetestbundlecomprised20lightwaterreactor(LWR)fuelrodsandonecentralsilver-indium-cadmiumcontrolrodwithstainlesssteelcladding.
TwoZircaloygridsarelocatedat220and740mmfromthebottomofthefuelcolumn.
Thetestpackage,surroundedbyaninsulatingshroudandcontainedinapressurisedin-piletube,isinsertedintoapressurizedwaterlooplocatedatthecentreofthe40MWPHEBUSreactorcore.
.
ThebundlewasirradiatedforafewdaysinthePHEBUSreactor.
Theobjectiveoftheirradiationistore-buildinthetestfueltheinventoryinshort-livedfissionproductssuchasI131,toinduceradiationdoseslargeenoughtostudyradiochemistryeffectsinthecontainment.
Theinitialbundleinventoryandtheobtainedinventoryofthemainfissionproductsaregivenintable1forthetwotests.
Themaindifferencebetweenthetwotestsarisefromthedifferenceinfuelburn-up:thefissionproductinventoriesaremuchlargerinFPT-1.
TheupperplenumabovethebundleisconnectedtoanhorizontallinemadeofInconelandtraceheatedto700°C.
ItconveysthegasesandaerosolgettingoutofthetestbundletoaUtubesimulatingaPWRsteamgenerator.
InFPT-0andFPT-1,thewallsofthesteamgeneratormodelweremaintainedat150°C,i.
e.
abovetheconditionsofsteamsaturationfortheconsideredpressure.
TheoutletoftheUtubeisthenconnectedtothecontainmentmodel,thussimulatingacoldlegbreak.
Afterthehorizontallineat150°C,a10m3cylindercollectstheaerosol,gasandsteam/hydrogeneffluentsconveyedbythecircuitduringthetest,simulatingareactorcontainmentbuildinginpresenceofabreakintheprimarycircuit.
Particulardesignfeaturesofthecontainmentvesselareasumpatthebottom(thesumpisinitiallycomposedof100lofwateratpH=5.
0(0.
3Mboricacid,1.
7510-4MNaOH))andagroupofthreecondensersintheupperpart(paintedwithRIPOLINepoxypaint,currentlyusedinEDFplants),whicharedesignedtocontrolsteamcondensationandwhichthussimulatethecoldstructuresofareactorbuilding.
Theoutervesselwallisheatedtoavoidsteamcondensationandsubsequentaerosoldepositiononthecontainmentwalls.
Figure1:SchematicrepresentationofthePHEBUStestfacilityforFPT-0andFPT-1.
The900MWeLWRtoPHEBUSscaling-downfactoriscloseto1/5000forthe3mainmodelledcomponentsofareactordepictedinthefigure(reactorcore,primarycircuitwithasteamgenerator,containment)Table1:FPT-0andFPT-1initialbundleinventoriesformainelementsofinterestElementFPT-0,mass(g)FPT-1,mass(g)ElementFPT-0,mass(g)FPT-1,mass(g)BundlematerialU9284.
09163.
0Fe179.
8173.
9Pu-47.
6Sn50.
649.
5Zr3417.
33475.
7Cr51.
649.
7Ni34.
534.
2Mn3.
64.
6ControlrodmaterialAg476.
8478.
3Cd29.
829.
9In89.
489.
7ThermocouplematerialRe706.
0482.
12FissionproductsBa0.
1510.
1Mo0.
1820.
2La0.
0957.
36Te0.
0472.
53Ru0.
1511.
2Cs0.
1415.
9I0.
0361.
12FissionproductsmeasurementsTheevaluationoftheamountoffissionproductsandactivationproductsinvariouspartsofthefacilityisobtainedbyapplyingseveraltechniques.
Duringthetest,on-line(spectrometersmeasuredtheactivitiesofthenumerousradionuclidesreleasedbythefuel.
Threearelocatedalongthecircuit(inletofsteamgenerator(SG),alongtheSGandoutletoftheSG);onemeasurestheactivityofthecontainmentatmosphereaswellasthatofthedepositsonthecontainmentwallsandcondensers,finallyonemeasurestheactivityofthewaterinthesump(figure1).
Afterthetest,inFPT-1,theevaluationoftheamountoffissionproductsandactivationproductsstillpresentinthefuelregionordepositedintheverticalchannelwasobtainedpost-testbyquantitativeg-spectrometryofthetestdevice(Cornu,1997).
Inbothtests,samplingsofreleasedmaterialswereperformedbyaspecificinstrumentationplacedinfurnacesatspecificpointsonthehorizontallineat700°C(pointC,figure1)andonthehorizontallineat150°C(pointG,figure1).
Aerosolswerecollectedoninertialimpactorsandfilters.
GascapsuleslocateddownstreamfiltersandimpactorswereusedtocollectgasatpointG.
ThermalgradienttubeswereusedatpointCtocondensevapours.
Thecontainmentwasstronglyinstrumented.
Aerosolswerecollectedoninertialimpactors,filtersandsequentialsedimentationcoupons.
GascapsuleslocateddownstreamfiltersorimpactorsandMay-packswereusedtosamplethecontainmentgas.
May-packsallowforthediscriminationofvolatileiodinespeciesonsuccessivefilteringstages(moleculariodineischemi-sorbedonthe"Knit-mesh"filterstageandorganiciodineontheZeolitefilterstage).
ThedifferentfilterstagesofoneMay-packwere(scannedon-lineduringbothtests.
DuringFPT-1,alargegascapsulepartiallyfilledwithZeolite3wasalso(-scannedon-line.
Allsamplinginstrumentsandsectionsofthecircuitmaincomponentarerecoveredbyremotehandlingassoonastheexperimentalinstallationisbacktoatmosphericpressureandroomtemperature.
Theyaretransferredtoahotcellunderthe"FPcaisson"wherefirstinspectionsandg-scansarecarriedout,beginningwiththosesamplerswhichhavetobescannedforI131analysis.
FPT-0andFPT-1testconditionsThefirsttestoftheprogramme,FPT-0,usingfreshfuel,wasperformedfromDecember2to6,1993.
Thesecond,FPT-1,withirradiatedfuel,fromJuly26to30,1996.
Bothexperimentswereoperatedundersimilarthermal-hydraulicconditions,themaindifferencebetweenthembeingtheburn-upofthetestfuel.
Thetestscenariosweredividedinexperimentalphasesaslistedintable2.
Theexperimentsconsistedfirstof5hourstransientsduringwhichtheinletcoolant(steam)flowvariedbetween0.
5g/sand3g/sinFPT-0(2g/sinFPT-1).
Themainobjectivewastoinvestigatebundledegradation,fissionproductreleaseandtransportunderlowpressure(0.
2MPa)andoxidising(steam-rich)conditions.
Thebundlefissionpowerwasgraduallyincreaseduntilcladdingrupture,oxidation,andfuelmeltingoccurred.
HightemperaturesreachedduringthetransientsresultedinZircaloyandcontrolrodabsorberalloymelting,fuelliquefaction,materialrelocation,fuelmeltingandthereleaseofhydrogen,aerosolsandfissionproducts.
Thetransientswereterminatedbyacoreshut-downandacoolingofthebundlewithsteam.
ExperimentalresultsconcerningthebundledegradationinPHEBUStestsaredescribedindetailinseparatepapers(VonderHardt,1995;Schwarz,1998).
Duringthedegradationphase,thecontainmentvesselboundaryconditionswereadjustedinordertolimittherelativehumidityratiotoabout70-80%,avoidinguncontrolledsteamcondensationontheouterwallswhichcouldforcefissionproductdepositiononundesirablenoninstrumentedspots.
Thestudyoftheaerosolsettlingandsurfacedepositioninthecontainmentlastedfor19hoursafterthebundledegradationinFPT-0andforabout60hoursinFPT-1(table2)duringtheso-calledaerosolphase.
Theaerosolphasewasfollowedbythewashingoftheellipticbottompartofthecontainment.
Thewashingwasperformedwiththewaterofthesumptobringthesettledaerosolsintothesump.
Theobjectiveofthisexperimentalphaseistobringsignificantradioactivematerialintothesumptoobserveradiolyticaleffectsduringthefollowingchemistryphase.
Itrequiredapreparatoryphaseduringwhichcontrolledtemperatureswerechangedtoavoidsignificantwatervaporisationfromthesumpduringtheoperation(table2).
Afterthewashing,anotherpreparatoryphasewasnecessarytore-adjusttemperaturestotheirdesiredlevelforthechemistryphase.
Thislastexperimentalphaselastedforabout4daysinFPT-0andabout1dayinFPT-1,duringwhichtheiodinechemistryinthecontainmentatmosphereandthesumpwasstudied.
Theobjectiveistomeasureiodinevolatilityonthelongtermandtoprovidesomeunderstandingontheprocessesattheoriginoftheformationofvolatileiodine.
Table2:ContainmentboundaryconditionsduringexperimentalphasesofthePHEBUStestsContainmenttemperatures(°C)PhasesDuration(hours)wallscondensersSumpFPT-0FPT-1FPT-0FPT-1FPT-0FPT-1FPT-0FPT-1Bundledegradation55110110741109090Coreshut-down-ContainmentisolationAerosolphase18.
56011011074then1101109090Preparatoryphase64.
5110to120110to120110to40110to4090to4090to40Washing0.
250.
2512012040404040Preparatoryphase67120to130120to13040to11040to11040to9040to90Chemistryphase96181301301101109090Containmentdepressurisation-endofexperimentalphasesCURRENTUNDERSTANDINGOFTHEGASEOUSIODINEBEHAVIOURINFPT-0ANDFPT-1ReleasefromthefuelbundleandtransportthroughthecircuithotlegThedegradationphaseresultedinalargereleaseoffissionproductsandactivationproducts(aerosolandstructuralmaterials)inthecircuitandinthecontainmentvesselasillustratedinfigure2forI131inFPT-1.
Inbothtests,thehighestreleaseratesweremeasuredduringafirsttemperatureescalationphase(firstescalationoftheZr-oxidationreaction),thenduringtheheat-upphasefollowingthefirstescalationphaseandfinallyduringasecondoxidationphase,occurringbeforethecoreshut-down(figure2).
Thelaterphaseprobablycorrespondstooxidationinlowerpartsofthetestbundle,duetomoving-downofhotmaterial.
Theintegralreleasefractionsfromthebundleforthemainelementslistedintable1provedtobeverysimilarinbothtests.
Fourkindsofnuclidescanbedistinguished:-highlyvolatilenuclides(releasefractionhigherthan50%):Te,I,Cs,SbandprobablyCd,-releasefractioninbetween10and50%:structuralmaterialandfissionproductssuchasMo,Ag,In,Sn,FeandCo,-releasefractionlyingbetween1and10%:materialsuchasW,Re,Mn,Ru,-releasefractionunder1%:materialsuchasBa,Sr,U,Nb,ZrandCe.
Inbothtests,iodinewasalmosttotallyreleasedfromthefuelbundleduringthedegradation(about86.
5%oftheinitialbundleinventoryinFPT-0),probablyasagas,consideringthehightemperaturesreachedinthefuel(Schwarz,1998).
Intheverticallineabovethebundle,gaseousiodineprobablyreactedpartiallywithAg,In,CsorRbtoformametaliodidevapour.
TheexistenceatthislocationofsignificantunreactedgaseousiodinefractionsisnotunexpectedasshownbythermodynamicevaluationsperformedforPHEBUSconditions(Gtzmann,1996).
Acomparisonoftheiodinemassflowingthroughthehotlegofthecircuit(pointC)withthatflowingthroughthecoolerzonesofthecircuit(steamgeneratordeposits+flowthroughthecoldleg(pointG))ispresentedinfigure2and3forFPT-1.
TheresultsindicateadeficitiniodineflowatpointCwithrespecttotheflowthroughthecoolerzonesofthecircuitamountingto28%oftheflow(measuredinthecoolerzonesofthecircuit)forthesamplingperformedduringtheearlydegradationphaseat11050-11349sand22%forthesamplingperformedduringthelatedegradationphaseat16766-16823s.
Thus,inthehotlegofthecircuit(pointC,700°C),atleast20to30%oftheFigure2:FPT-1-(top)Samplingtime-windows(thicklines)andcorrespondingnon-condensediodinefractionsinthehotlegofthecircuit(pointC,700°C);themeasuredevolutionofhydrogenvolumeconcentrationinthecircuitisshown,togetherwiththeevolutionoftheOLAMsignal(on-lineaerosolmonitor,signalattenuationscorrespondtoincreaseintheaerosolconcentrationinthecircuitcarriergas)andtheI131flowratesignal.
(bottom)Iodineflowratedatainthehotleg(pointC,700°C)andthecoldleg(pointG,150°C)ofthecircuit.
Notethattwoconsecutivedatapointswiththesamevaluecorrespondtothestartandtheend-timeofagivensampling.
Thicklinesonthetopofthefigureindicatesamplingtime-windowsyieldinginformationongaseousiodineatpointCandatpointG.
Steamandhydrogenflowratesinmol/saregiven,showingthatreducingconditionsarereachedonlyduringthefirstoxidationphase.
totaliodineflowwaspresentasaformwhichcouldnotbecondensedonthe150°CsamplingsystemsduringtheearlyandlateZr-oxidationphase.
Thisformmaybeattributedtogaseousiodinesincemetaliodidecompoundswouldhavebeencondensedinthesamplinglineandinthe150°Cfilterasevidencedbythesamplingperformedduringthemainreleasephase(samplingat13809-14101s,figure3).
NosimilardeterminationcouldbemadeforFPT-0sinceforthistesttheinstrumentationinthehotlegofthecircuitdidnotinclude150°Caerosolfiltrationsystems.
However,thepresenceofgaseousiodinefractionsinthecircuithotleginFPT-0cannotbeexcluded.
Apparently,theconditionsfavouringtheformationofgaseousiodinecorrespondtotheconditionsobtainedduringtheZr-oxidationphases.
DuringthefirstoxidationphaseinFPT-1,thecarriergascontainsH2inmolarexcesswithrespecttosteam.
Reducingconditionsinthecircuitwouldberesponsibleforthepresenceoflargefractionofgaseousiodine,aresultwhichcontradictsthermodynamicevaluationsmadetodateforPHEBUSconditions(see,Gtzmann,1996).
Figure3:FPT-1-Comparisonoftheiodinemassflowingthroughthehotlegofthecircuitwiththatflowingthroughtherestofthecircuit(steamgeneratorandcoldleg)forthethree700°C+150°Cfilterssamplings.
Characteristicsofeachsampledexperimentalphase(H2/H2O,Cs/I,Ag/IandAg/Cs)molarratiosaregiven.
DuringthesecondoxidationphaseinFPT-1,thecarriergasstillcontainssteaminmolarexcesswithrespecttohydrogen.
ResultshoweverindicatethattheCsconcentrationsarelowerduringthislatephaseofthetransient.
TheexistenceofsignificantgaseousiodinefractionatthistimemaybelinkedtotheselowCsconcentrations,anassertionwhichneedstobeconfirmedbythermodynamicevaluations.
Itisalsoworthnotingthattheamountofiodinecollectedonthe700°CfilteratpointCinFPT-1issignificant.
Thisfractionofcondensediodineat700°Ccannotbeattributedtoknownsimplemetalliciodidespeciesbuteithertoanunknowncomplexspeciesortoiodinetrappedincondensedmatter.
Thisfractionrepresentsmorethan10%ofthetotaliodineflowonthesamplingperformedduringthelateoxidationphase.
TheresultsobtainedinthehotlegofthecircuitforFPT-1confirmthecorrelationexistingbetweenbundledegradationevents(fueloxidation,materialre-location,poolformation)andthephysicalformandamountofiodinetransportedthroughthecircuit).
Moreelementstoestablishthiscorrelationareobtainedwiththecontainmentsampledata.
SteamgeneratordepositsIodinewasthentransporteduptothesteamgeneratortubeanddepositedtheremorethanotherfissionproductsandstructuralmaterials(inFPT-0,about27%ofthereleasediodinefractionagainstanaverage15%fortheotherelements).
TheiodinedepositionprofilemeasuredinFPT-1indicatethatmostofthedepositionoccurredattheentranceofthehotlegofthesteamgenerator(figure4).
Aresultwhichmaybeinterpretedeitherbywallandbulkcondensationofametaliodidevapour,orbychemisorptionofgaseousiodineonCddeposits(Cddepositswerefoundbypost-testanalysestobeverylargeatthislocationinFPT-0).
Onecannotexcludeheretootheexistenceofsignificantunreactedgaseousiodinefractions.
Inthesteamgeneratortemperaturerange(700°Cdownto150°C),reactionkineticsmaylimittheyieldsofreactioninvolvingiodineandvapourorcondensedmetals.
Theselimitationsaremoreimportantwhenthereactive(here,fissionproducts)concentrationsarelower.
SinceinFPT-0thefissionproductsconcentrationsaresignificantlylowerthaninFPT-1(cf.
table1),onecouldexpectlargerfractionsofunreactediodineintheFPT-0circuits.
Experimentalresultsinthecontainmentconfirmthistendancy(cf.
nextparagraphs).
Figure4:FPT-1-I131depositionprofileinthehotlegofthesteamgenerator.
Iodinedepositionprofileissingularwithrespecttothatmeasuredforothernuclidesinthatiodinedepositionislargerattheentranceofthehotleg.
TransportthroughthecoldlegofthecircuitInthecoldlegofthecircuit(pointG,150°C),mostoftheiodineisapparentlytransportedbyaerosols.
FPT-0gascapsuledataindicatedthatatleast2%ofthegaseousiodineflowwaspresentunderagaseousformafterthemainZircaloy-oxidationphaseandatthecoreshut-down.
Theselowfractionsmeasureddonotexcludethattemporarilygaseousiodinefractionsmayhavebeenmuchhigher.
Accordingtothefirstcontainmentsamplingperformedaftertheoxidationphase,morethanathirdofthetotaliodineflowwaspresentunderagaseousiodineformduringtheoxidationphase(cf.
nextparagraph).
FPT-1ZeolitegascapsuledatashownoevidenceofsignificantgaseousiodinefractionsontheZeolite,evenduringoxidationphases.
However,asamplingperformedduringthefirstoxidationphaseinthecontainmentshowsthattheinjectedgaseousiodineisnotefficientlytrappedonadoubleZeolitefilter,alargefractionofitisfoundinthedownstreamgasspace(atleast40%).
Basedonthisdata,theexistenceofaniodinespeciesnottractablebyZeolitefiltrationatpointGcanbesuspected.
ThegasspacevolumeoftheZeolitegascapsuleatpointGbeingsmallandthegaseousiodineproductionbeingprobablystronglytime-dependent(accordingtohotlegandcontainmentsamplings),itisnotsurprisingtofindnoorextremelysmallamountofiodineinthegasspacevolumeofthosecapsules.
ThosecoldlegsamplingsillustratethedifficultyencounteredduringPHEBUSteststomaptime-dependenteventswithalimitednumberofsamplingsandtodesignaninstrumentationdedicatedtogaseousiodinedetermination,notknowingthespeciesinvolved.
FPT-1resultsindicatethatthegaseousiodinespeciespresentinthecircuitmaynotbetractablebyZeolitefiltrationwhichexcludes,apriori,theexpectedgaseousI2andHIspecies.
TransporttothecontainmentandbehaviourduringthebundledegradationphaseMorethanhalfofthefuelinventoriesofiodinesuccessfullynegotiatedpassagethroughthePHEBUScircuittoreachthecontainmentmodel(63%inFPT-0).
Muchofthisiodinewasconveyedbyaerosolparticlesthateithersettledonthecontainmentbottomordepositedonthecontainmentsurfaces(bydiffusiophoresisonthepaintedcondenserandbyanunidentifiedmechanismonthecontainmentwalls)overaperiodofafewhours.
Inbothtests,themeasuredgaseousiodinefractionsareattheirhighestlevelat,orfollowing,Zircaloy-oxidationphases(figure5).
InFPT-0,therewasnosamplingperformedduringthefirstoxidationphase.
Thefirstsamplingperformedabout30minutesafterthisphaseyieldedagaseousiodinefractionrepresentingatleast33(12.
1%ofthecontainmentinventoryatthattime.
Thisresultdoesnotexcludethattemporarily,earlieraftertheoxidationphase,thegaseousiodinefractioncouldhavebeenevenhigher.
InFPT-1,thereweresamplingsperformedduringthetwooxidationphases.
Thesamplingperformedduringthefirstoxidationphaseindicatedthatatleast64.
05(0.
9%ofthecontainmentinventoryisgaseousatthattime.
Atthistime,about40%ofthecollectedgaseousiodineisnotretainedonthedoubleZeolitefilteroftheapparatus,showingthatasignificantfractionofthegaseousiodineisnottractablebyZeolitefiltration(whichaprioriFigure5:gaseousiodinefractionin%ofthecontainmentinventoryatsamplingtimesduringthebundledegradationphaseandtheearlyaerosolphase.
Evolutionofthecontainmentinventoryisshownasanindicationasfulllines.
DatauncertaintiesarediscussedinPantera(1998)andJacquemain(1999).
excludesforthisfractionanidentificationwiththeexpectedHI,I2ororganiciodides).
Latersamplingsperformedwiththesameapparatusdonotexhibitanymorethisspecificiodinebehaviour.
The"unknown"speciesdetectedearlyduringthetestprobablytransformsfastintoanotherchemicalspecieswhichisthenefficientlyretainedonZeolite.
Thesamplingperformedduringthesecondoxidationphaseyieldsagaseousiodineamountcomparabletothatobtainedduringthefirstoxidationphase.
AfterbothoxidationphasesinFPT-1,thegaseousiodinefractiondecreasesfast.
Bothdecreasesmaybefittedmathematicallybythesameexponentiallydecayingfunction.
Afastdecreasewhichmaybelinkedeithertoadiffusiophoresisofgaseousiodine(Weber,1992)atthepaintedcondensingsurfacesand/ortoafastchemicaltransformationasmentionedearlier.
ThevolatilefractionofiodineoriginatingfromtheprimarycircuitalsodisappearedinafewhoursinFPT-0,probablyduetodepositiononthecontainmentsurfaces(cf.
nextsection).
ItishowevernotpossibletocharacterisepreciselythegaseousiodineevolutionduringthebundledegradationphaseinFPT-0sinceonlyonemeasurementwasperformedduringthisphase.
Behaviouraftercontainmentisolation(aerosolandchemistryphase)Theevolutionofthegaseousiodinefractionduringtheaerosolphaseisdifferentinthetwotests.
InFPT-1,thegaseousiodinefractionincreasessignificantlyrightaftercontainmentisolation(byafactorof2,figure5and6).
Itremainsthenmoreorlessconstantduringabout5hours.
Theincreasemaybeattributedtoameasuredreleaseofgaseousiodinefromthepaintedcondensingsurfaces(figure7).
Thereleaserepresentsaboutathirdoftheiodinewhichwasdepositedonthecondensingpaintedsurfacesattheendofthebundledegradationphase,i.
e.
about0.
6%ofthecontainmentinventory.
May-packdata(figure9)indicatethat,atthistime,mostoftheiodine(75%ofit)isfoundontheZeolitestageofthedevice,consistentwithareleaseoforganiciodidescompounds.
Thereleasediodinemayhavetwoorigins:eithergaseousiodinewhichreacteddirectlywiththepaintduringthetesttransientorformationofiodideions,I-,bysolubilisationinthecondensedsteamofthesolubleiodinespeciesinitiallytransportedbytheaerosolandsubsequentreactionwithpaint.
Noconclusionispossiblefromtheexperimentaldataonthisquestion.
InFPT-0,thegaseousiodinefractiondecreasedexponentiallyduringtheaerosolphase(figure6).
Noobservationofasignificantdecreaseofiodineactivityonthepaintedcondensersurfaceswasmade.
Thedatadoesnotexcludehoweverthataminorrelease,insignificantwithrespecttothetotalamountofiodinedeposited,occurredduringthelongtermphase.
Forthelongterm,thegaseousiodineevolutioninFPT-1isprobablydeterminedbythemore"soluble"behaviourofiodine,comparedwithFPT-0,notedduringthebundledegradationphaseandtheaerosolphase.
Indeed,thesumpiodinevolumicactivityevolutionmeasuredduringtheaerosolphase(figure8)indicatesthattheremovalofiodinefromthebulkvolumeoftheaqueousphaseisslow.
Itrequiresabout50h.
ThisremovalprocessmaybeidentifiedeitherasaslowreactionofiodideionswithAgmaterialanddepositionofthismaterialonthesumpbottom,orbyaslowdepositionofaniodinesuspensionofthecolloidaltypeonthesumpbottomorbyaslowdepositionontothesumpsurfaces.
Thelatermechanism(ofdepositionontothesumpsurfaces)wasshowntobeunsignificantincontainmentmassbalanceevaluationpresentedinGarnier(1999).
Itisnotpossibletodecideonthefirsttwopossibilities,actuallyacombinationofbothisprobable.
Figure6:evolutionofthegaseousiodinefractioninthecontainmentofthetwofirstPHEBUStests.
ConfidenceintervalsonthevaluesarediscussedinPantera(1998)andJacquemain(1999).
Figure7:FPT-1comparisonoftheevolutionofthegaseousiodinefractioninthecontainmentatmosphereandoftheevolutionoftheiodinefractiondepositedonthepaintedcondensingsurfaces.
AsimilardecreaseoftheactivitywasmeasuredfortwoiodineisotopesI131andI133onthecondensers.
Otherisotopes,suchasTe132givenhereasanexample,donotexhibitthebehaviourobservedforiodine.
Figure8:FPT-1comparisonoftheevolutionofthegaseousiodinefractioninthecontainmentatmosphere(confidenceintervalsdiscussedinJacquemain(1999))andoftheevolutionoftheiodinefractioninthecontainmentsump(measuredactivityinterpretedasvolumicactivity).
Thedecreaseiniodinevolumicactivitycorrespondstoaremovalofiodinefromthebulkvolumeoftheaqueousphase.
Figure9:FPT-1-Distributionofgaseousiodineinthesequentialandon-lineMay-packsasafunctionofexperimentaltime.
Thegaseousiodineamounttrappedineachtypeofiodinefilter(successivelyKnit-mesh,silverZeoliteandKIimpregnatedcharcoal)wasnormalisedtothetotalgaseousiodineamounttrappedintheMay-packdevice.
Theevolutionofthegaseousiodinefractionafterthewashingphaseisdifferentinthetwotests.
InFPT-1,asignificantincreaseofthegaseousiodinefraction(byafactorcloseto2)wasmeasured(figure6).
ThisincreasemaybeattributedtotheformationandtransfertothecontainmentatmosphereofvolatileI2formedbyradiolyticoxidationofsolubleI-ions.
May-packdataindicatethatatthewashing75%oftheiodineistrappedontheKnit-meshstageoftheapparatus(figure9),consistentwithI2beingthemajorgaseousiodinespeciesatthattime.
Afterthewashing,theI2contributiontothegaseousiodinefractiondecreasesregularlyfrom75%downtoabout25%,probablyduetoaback-transferofI2tothesumpandsubsequentslowreactionwithAg.
Thetotalgaseousiodinefractionremainshoweverconstant,indicatingthatcontributionoftheiodinespeciesdepositedontheZeoliteincreases.
Thisobservationwouldindicatethatasignificantproductionoforganiciodidescompoundsstilloccursafterthewashingphase.
However,noevolutionoftheiodinefractiondepositedonthecondensingsurfaceofthecondensercanbenotedfromfigure7.
Thereleaseatthattimemaybetoosmalltobeobservable.
InFPT-0,noincreaseinthegaseousiodineconcentrationwasobservedatthewashing;thegaseousiodinefractionremainedstable(figure6)andrepresentedsolely0.
063(0.
032%ofthebundleinventory.
TheseresultswereconsistentwithanefficienttrappingoftheiodinebyAgandthesubsequentinhibitionofvolatileiodineformationbysumpradiolysis.
May-packdataindicatedthatmostoftheiodinewastrappedontheZeolitestageoftheapparatus,consistentwithorganiciodidesbeingthemajorgaseousiodinespeciesatthattime.
Aresultwhichwasattributedtothereleaseoforganiciodidesfromthepaintedcondensersurfaces.
Forthelongtermevolution,themaindifferenceinbetweenbothtestsisprobablydetermined,asmentionedearlier,bythemore"soluble"behaviourofiodinenotedinFPT-1.
ThisdifferenceinbehaviourmaybeattributedtothelargedifferenceinbetweentheAg/ImolarratiosdeterminedinFPT-0(>1000)andinFPT-1(~50),thisratiobeingdeterminantforthekineticsofthereactiontoforminsolubleAgI.
Inshort,AgIbeingformedfastinFPT-0,theprocessesinvolvingsolubleiodideasareactivespeciesarestronglyinhibitedinthistest,AgIformationbeingslowerinFPT-1,thoseprocessesmaybecomesignificantintheproductionofvolatileiodine.
Thoseprocessesare:-reactionofI-ions,formedbysolubilisationincondensedsteamofthesolubleiodinespeciesinitiallytransportedbytheaerosol,withthepaintofthecondenser,leadingtothereleaseoforganiciodidecompounds,-radiolyticoxidationofI-ionspresentinthesump,leadingtotheproductioninthesumpandtransfertothecontainmentatmosphereofmoleculariodineI2.
TheFPT-1data,takenasawhole,appeartobeconsistentwithaslowAg/Ireactionprocess,allowingfortheprocesseslistedabovetooccuranddeterminethelongtermvolatileiodinebehaviour.
FPT-0datawereconsistentwithafastiodinetrappingbysilver,inhibitingthoseprocessesandthelongtermvolatileiodinebehaviourwasessentiallydeterminedbytheevolutionofthegaseousiodinewhichwasinjectedfromtheprimarycircuitduringthebundledegradationphase.
Theresultsalsoshowthatorganiciodidescouldverywellbethemajorcontributortothegaseousiodinefractionattheendofbothtests.
Thoseresultshaveimportantapplicationsforiodinechemistrycodedevelopmentandiodinesourcetermevaluationasdiscussedinnextsections.
ImplicationsofPHEBUSresultsforiodinechemistrycodedevelopmentprocesseswithprototypicmixesofradioactiveandnon-radioactivematerials,radiationfields,temperaturesandpressures.
WhatissoughtfromthePHEBUStestresultsisacheckthatthecurrentunderstandingofthechemistryofiodineasembodiedinthecomputercodesissufficienttopredictadequatelyiodineconcentrationsinreactorcontainmentsandtheratesatwhichtheseconcentrationsarereached.
Iniodinechemistrycodes,validatedonthecurrentknowledgeofiodinechemistry,iodineconcentrationsandpartitioningratesaredeterminedfrom:-theformationofvolatileformsofiodine(I2ororganiciodides)inthesump,-transportofvolatileiodinespeciesfromthesumpintothebulkcontainmentatmosphere,-lossofiodinefromtheatmospherebyhomogenousreactionordepositiononsurfaces.
ThePHEBUStestresultsindicatethattheiodinebehaviourinthecontainmentcannotbedescribedsimplybythoseprocesses,mainlysincethesumpisnottheonlyvolatileiodinesource.
Ontheshortterm(duringthefissionproductsreleaseandaerosolphase),volatileiodineconcentrationsinthecontainmentwereessentiallyduetogaseousiodineinjectedbytheprimarycircuit.
Consequently,itisessentialtodevelopkineticsmodelsoftheprimarycircuitchemistry,abletocalculatetheiodinesourcetothecontainmentforthefollowingtworeasons:-theknowledgeofthegaseousiodineinventoryontheshorttermisimportantforsafetyanalysisofsituationsinwhichthecontainmentintegrityisnotpreserved,-volatileiodinefromtheprimarycircuitmayeitherdepositonsurfacesorreactwithaerosolmaterialinthecontainment.
Surfacedepositediodinemaydeterminethelongtermvolatileiodineconcentrationsinthecontainment(e.
g.
,releaseoforganiciodidesfromsurfaces).
Onthelongterm,productionofvolatileformsofiodinecanoccureitherintheaqueousphaseoronpaintedsurfacesinthecontainmentmodel.
Modelsoforganiciodideformationandbehaviourunderreactoraccidentconditionsarenotnearlyaswelldevelopedasmodelsofelementaliodineformationandbehaviour.
Aneffortofmodeldevelopmentforthesereactionsisalsorequired.
Volatileiodineformationintheaqueousphaseisknownfromlaboratorystudiestodependontheiodideconcentrationintheaqueousphase,theradiationdosetotheaqueousphase,thetemperature,thepHoftheaqueousphase,andthepresenceofsilver.
ThePHEBUStestsyieldedresultsconcerningtheaqueousconcentrationofiodine,theeffectofAg,thesolutionpHthathaveimplicationsonthemodellingofseverereactoraccidentsequencesinwhichthedegradationresultsinasignificantcontrolrodmaterialreleasetothecontainment.
SubstantialamountsofthereleasedsilverreachedthePHEBUScontainmentmodel.
Evidencetodatefromthereactorcircuitisthatsilverwasintheelementalformwithinthereactorcircuit.
But,onceitreachedthereactorcontainment,itwasoxidisedbyfairlypowerfuloxidisingagents(ozoneandnitrousoxidesproducedbyairradiolysis).
Whendepositedinwater,thepartiallyoxidisedsilverformedcolloidalsuspensionsthatcouldreactwithdissolvediodinetoformAgI(s),reducingsubstantiallytheaqueousiodineconcentrationandgreatlyreducingtherateofvolatileiodineformationfromthesump.
ThesituationintheFPT0caseis,however,equivocalsincethemechanismswhichyieldedtheformationofcolloidalAgIarenotcompletelyclarified.
Extrapolationandmodellingforthereactorcasewillalsorequireasubstantialanalyticaleffort:-developmentofmodelstodescribetheAgcontainmentchemistryintheatmosphereandinthesumpunderrepresentativeconditions,goingthroughaclarificationofthemechanismsresponsibleforAgprecipitateorcolloidformation,-developmentofmodelstodescribetheAgIformationandstabilityinrepresentativesumpconditions.
Establishmentofasteadystateatmosphericconcentrationofiodineafterthewashingisofimportanceforthemodelling.
ItmeansthatmodellingofthebehaviourofiodineinthePHEBUSatmosphereislikelytobeatractablechore.
Indeed,analysesofthedatadonetodateshowthatbysuitablemanipulationofrateconstants,atmosphericconcentrationsobservedintheFPT0testcanbepredictedtoanaccuracyadequateforreactoraccidentanalysis.
Unfortunately,theFPT0databaseisnotsufficienttoshowthatthenecessarymanipulationsofthemodelscanbeextrapolatedtoreactoraccidentconditions.
Resultsfromseveraltestswillhavetobeexaminedtohaveconfidenceinsuchextrapolation.
Itisnecessarytohaveafirmunderstandingoftheratesofvolatileiodinetransporttotheatmosphereandtheratesofvolatileiodineremovalbyprocessessuchas:-homogeneousdecompositionbyreactionwithozone,O3,nitrogenoxides,andwatervapour,-depositionontopaintedsurfacesinthecontainment-depositionontometalsurfacesinthecontainment.
DefinitivemeasuresofeachoftheseratesofiodineremovalwillbeneededtofullyutilisethecontainmentchemistryresultsofthePHEBUStests.
CharacterisationofthedetailsofsuchgasphasemasstransportofvolatileiodinespecieswouldgreatlyassistmodellingofthePHEBUStestresultsandtheextrapolationoftheresultstoreactoraccidentanalyses.
Inconclusion,asubstantialworkofdevelopmentandvalidationofcontainmentiodinechemistrycodesisrequiredtomodelthePHEBUScontainmentchemistry.
Onspecificquestions,separatetesteffectexperimentsareneededsincePHEBUSdoesnotnecessarilyprovideinformationaccurateenoughformodeldevelopmentandvalidation.
Besides,theFPT0databaseisnotsufficient,todate,toshowthattheiodinechemistrycodemodelscanbeextrapolatedtoreactoraccidentconditions.
Resultsfromseveraltestswillhavetobeexaminedtohaveconfidenceintheextrapolation.
ImplicationsofPHEBUSresultsforLWRiodinesourcetermConsequently,itisessentialtodevelopkineticsmodelsoftheprimarycircuitchemistry,abletocalculatetheiodinesourcetothecontainmentforthefollowingtworeasons:-theknowledgeofthegaseousiodineinventoryontheshorttermisimportantforsafetyanalysisofsituationsinwhichthecontainmentintegrityisnotpreserved,-volatileiodinefromtheprimarycircuitmayeitherdepositonsurfacesorreactwithaerosolmaterialinthecontainment.
Surfacedepositediodinemaydeterminethelongtermvolatileiodineconcentrationsinthecontainment(e.
g.
,releaseoforganiciodidesfromsurfaces).
Themostimportantresultsarethefollowing:-theproductionofgaseousiodinebyradiolysisofiodinewasmuchlowerthanexpected,duetothepresenceofsilvercomingfromthecontrolrod-anon-anticipatedresultisthatsomepercentofiodineenteringthecontainmentwasinvolatileform.
Thoseresultswhichconcernsthegaseousiodineamountinthecontainmentareveryimportantbecausetheycanhaveanimpactonthelatereleaseofiodineincaseofventingorincaseofcontainmentfailure.
CONCLUSION:futureexperimentsDatabasewillbeextendedinFPT-2withtheperformanceofatestwithamoresignificantreducingwindowandanewinstrumentationimprovedinthehotlegofthecircuit.
FPT-3withBoroncarbidetostudytheimpactonprimarycircuitchemistryandcontainmentchemistry.
Concerningprimarycircuitchemistryevenifintegralexperimentsprovemoreadequatetomeasurefractionsreleasedandchemicalforms.
However,thefactthatchemicaltransformationsintheRCSplayanimportantrlewhichcanbedifferentfordifferentaccidentalsequencesleadtothenecessitytoperformanalyticalstudiestovalidateachemicalmodel.
Studycoldlegandhotlegbreak,reducingandoxidisingconditions,FPconcentrations,presenceofCRDmaterialornot.
ExperimentstoinvestigatethoseaspectsareunderdiscussionExperimentstobetterdelineatemechanismsofformationofgaseousiodinefromwatersolutionsinarangeofconditions(effectsofAg,andBspecies),pH,effectofirradiation,hightemperatureneedstobepursuedinseparateeffectsexperiments.
Experimentstostudysystematicallythereleaseofvolatileorganiciodidesfromreactorpaintedsurfacesandothermaterialsuchasliners,cableshightemperatureradiolysis.
SuchexperimentsareplannedoralreadyrunningintheframeworkofthePHEBUSprojectandE.
U.
programs.
REFERENCESClément,B.
,Ktorza,C.
,Jones,A.
V.
,Zeyen,R.
,1998,"StatusandMainFindingsofthePHEBUSFPProgramme"ProceedingsoftheSARJ'98Meeting,Tokyo,November4-6Cornu,B.
,Giacalone,J.
-C.
,"QuantitativeGammaMeasurementsoftheTestDevice,AssessmentoftheReleases",PHEBUSProjectReport,InstitutdeProtectionetdeSretéNucléaire,Cadarache,July1997Gtzmann,O.
,1996,"FissionProductSpeciationinPHEBUSTestsFPT-0andFPT-1andthePossibleInfluenceofBoron",ReportFZKA5682,ForschungszentrumKarlsruheGmbH,Karlsruhe,January1996Hanniet,N.
,Garnier,Y.
,Jacquemain,D.
,1998,"FissionProductsandStructuralMaterialsRelease,TransportandContainmentBehaviourinPHEBUSFPT-0andFPT-1"ProceedingsoftheSARJ'98Meeting,Tokyo,November4-6Jacquemain,D.
,Hanniet,N.
,Hueber,C.
,Poletiko,C.
,Chuaqui,C.
A.
,Powers,D.
,Dickinson,S.
,Drossinos,Y.
,Funke,F.
,Herrero,B.
,1997,"AnOverviewoftheIodineBehaviourinPHEBUSFPT-0",Proceedingsofthe5thInternationalConferenceonNuclearEngineering-ICONE5,2181/1-2181/11,AmericanSocietyofMechanicalEngineer,New-YorkJacquemain,D.
,Hanniet,N.
,1999,"AnOverviewoftheGaseousIodineDataintheExperimentalCircuitsandtheContainmentintheTwoFirstPHEBUSTestsFPT-0andFPT-1",PHEBUSProjectReport,InstitutdeProtectionetdeSretéNucléaire,Cadarache,March1999Krischer,W.
,Rubinstein,M.
-C.
(Eds.
),1992,"ThePHEBUSFissionProductProject",ElsevierAppliedSciencePublications,LondonandNew-YorkPantera,L.
,Cornu,B.
,Garnier,Y.
,Siri,B.
,1998,"FPT-0:UncertaintiesandCoherenceofExperimentalData",PHEBUSProjectReport,InstitutdeProtectionetdeSretéNucléaire,Cadarache,November1998Schwarz,M.
,Clément,B.
,Ktorza,C.
,Jones,A.
V.
,Zeyen,R.
,1998,"ThePHEBUSF.
P.
InternationalResearchProgramonSevereAccidents:StatusandMainFindings",Proceedingsofthe26thWaterReactorSafetyInformationMeeting,Bethesda,Maryland,October26-28VonderHardt,P.
,Jones,A.
V.
,Lecomte,C.
,Tattegrain,A.
,1994,NuclearSafetyResearch,"ThePHEBUSFPSevereAccidentExperimentalProgramme",NuclearSafety35,Vol.
2VonderHardt,P.
,Tattegrain,A.
,1995,"PHEBUSProbesCoreMelt-downBehaviour:FirstResults",NuclearEuropeanWorldscanVol.
9/10Weber,C.
F.
,Beahm,E.
C.
,Kress,T.
S.
,1992,"ModelsofIodineBehaviourinReactorContainments",OakRidgeNationalLaboratoryReport,ORNL/TM-12202