accident source terms for light-water nuclear power plants

NUREG-1465

Accident Source Terms forLight-Water Nuclear Power Plants

Final Report

U.S. Nuclear Regulatory Commission

Offlce of Nuclear Regulatory Research

L Soffer, S. B. Burson, C. M. Ferrell,R. Y. Lee, J. N. Ridgely

@* '' .'/

.. I *0

NUREG-1465

Accident Source Terms for

Accident Source Terms forLight-Water Nuclear Power Plants

Final Report

Manuscript Completed: February 1995Date Published: February 1995

L Soffer, S. B. Burson, C. M. Ferrell,R. Y. Lee, J. N. Ridgely

Division of Systems TechnologyOffice of Nuclear Regulatory ResearchU.S. Nuclear Regulatory CommissionWashington, DC 20555-0001

9 A I8,4

. ... ...

Abstract

In 1962 The U.S. Atomic Energy Commission publishedTLD-14844, "Calculation of Distance Factors for Powerand Test Reactors" which specified a release of fissionproducts from the core to the reactor containment inthe event of a postulated accident involving a"substantial meltdown of the core." This "source term,"the basis for the NRC's Regulatory guides 1.3 and 1.4,has been used to determine compliance with the NRC'sreactor site criteria, 10 CFR Part 100, and to evaluateother important plant performance requirements.During the past 30 years substantial additional

information on fission product releases has beendeveloped based on significant severe accidentresearch. This document utilizes this research byproviding more realistic estimates of the "source term"release into containment, in terms of timing, nuclidetypes, quantities, and chemical form, given a severecore-melt accident. This revised "source term" is to beapplied to the design of future Light Water Reactors(LWRs). Current LWR licensees may voluntarilypropose applications based upon it. These will bereviewed by the NRC staff.

iii NUREG-1465

CONTENTS

Page

Abstract .............................................................................. iii

Preface......................................................................................... vii

1 Introduction And Background ...................................... . I

1.1 Regulatory Use of Source Terms ..I........................................................ 11.2 Research Insights Since TID-14844 ....................................................... 2

2 Objectives And Scope ...................................... . 32.1 General .............................................................................. 32.2 Accidents Considered . ........................................................... 32.3 Limitations ............................................................................ 4

3 Accident Source Terms ...................................... .5

3.1 Accident Sequences Reviewed ........................................................... 53.2 Onset of Fission Product Release ......................................................... 53.3 Duration of Release Phases ........................................................... 73.4 Fission Product Composition and Magnitude .......................... . 93.5 Chemical Form ........................................................................ 103.6 Proposed Accident Source Terms ...................... 123.7 Nonradioactive Aerosols ...................... 14

4 Margins And Uncertainties ...................... 154.1 Accident Severity and Type ......... .. 154.2 Onset of Fission Product Release ......... . . 154.3 Release Phase Durations ......... .. 154.4 Composition and Magnitude of Releases .................................................. 164.5 Iodine Chemical Form .......................................................... 17

5 In-Containment Removal Mechanisms .......................................................... 17

5.1 Containment Sprays .......................................................... 185.2 BWR Suppression Pools .......................................................... 185.3 Filtration Systems ............................ .............................. 195.4 Water Overlying Core Debris ........................................................... 205.5 Aerosol Deposition ............................ .............................. 20

6 References ........................................................... 21

TABLES

1.1 Release Phases of a Severe Accident ............... ....................................... 23.1 BWR Source Term Contributing Sequences ........................ 5.......................3.2 PWR Source Tberm Contributing Sequences . ............................................... 63.3 Contribution of LOCAs to Core Damage Frequency (CDF)-Internal Events ....... .......... 73.4 In-Vessel Release Duration for PWR Sequences ................. .......................... 83.5 In-Vessel Release Duration for BWR Sequences .................. ......................... 93.6 Release Phase Durations for PWRs and BWRs ................... ......................... 93.7 STCP Radionuclide Groups ............ ............................................ 103.8 Revised Radionuclide Groups ........................... 103.9 Fraction of mean core damage frequency with high, intermediate, and low pressure sequences

(internal events only unless otherwise noted) ....................... 11

v NUREG-1465

CONTENTS (Cont'd)

Page

3.10 Mean Values of Radionuclides Into Containment for BWRs,Low RCS Pressure, High Zirconium Oxidation ............ ................................. 11

3.11 Mean Values of Radionuclide Releases Into Containment for PWRs, Low RCS Pressure,High Zirconium Oxidation ....................... ............................. 11

3.12 BWR Releases Into Containment .................................................... 133.13 PWR Releases Into Containment .................................................... 134.1 Measures of Low Volatile In-Vessel Release Fractions ............... ....................... 16

APPENDICES

A Uncertainty Distributions ...................................................... 24

B STCP Bounding Value Releases .......................... ............................ 28

NUREG-1465 vi

Preface

In 1962, the Atomic Energy Commission issuedTechnical Information Document (IMD) 14844,"Calculation of Distance Factors for Power and TestReactors." In this document, a release of fissionproducts from the core of a light-water reactor (LWR)into the containment atmosphere ("source term") waspostulated for the purpose of calculating off-site dosesin accordance with 10 CFR Part 100, "Reactor SiteCriteria." The source term postulated an accident thatresulted in substantial meltdown of the core, and thefission products assumed released into the containmentwere based on an understanding at that time of fissionproduct behavior. In addition to site suitability, theregulatory applications of this source term (inconjunction with the dose calculation methodology)affect the design of a wide range of plant systems.

In the past 30 years, substantial information has beendeveloped updating our knowledge about severe LWRaccidents and the resulting behavior of the releasedfission products. The purpose of this document is to

provide a postulated fission product source termreleased into containment that is based on currentunderstanding of LWR accidents and fission productbehavior. The information contained in this documentis applicable to LWR designs and is intended to formthe basis for the development of regulatory guidance,primarily for future LWRs. This report will serve as abasis for possible changes to regulatory requirements.However, acceptance of any proposed changes will beon a case-by-case basis.

Source terms for future reactors may differ from thosepresented in this report which are based upon insightsderived from current generation light-water reactors.An applicant may propose changes in source termparameters (timing, release magnitude, and chemicalform) from those contained in this report, based uponand justified by design specific features.

NUREG-1465

1 INTRODUCTION AND BACKGROUND

1.1 Regulatory Use of Source TermsThe use of postulated accidental releases of radioactivematerials is deeply embedded in the regulatory policyand practices of the U.S. Nuclear RegulatoryCommission (NRC). For over 30 years, the NRC'sreactor site criteria in 10 CFR Part 100 (Ref. 1) haverequired, for licensing purposes, that an accidentalfission product release resulting from "substantialmeltdown" of the core into the containment bepostulated to occur and that its potential radiologicalconsequences be evaluated assuming that the contain-ment remains intact but leaks at its maximum allowableleak rate. Radioactive material escaping from thecontainment is often referred to as the "radiologicalrelease to the environment." The radiological releaseis obtained from the containment leak rate and aknowledge of the airborne radioactive inventory in thecontainment atmosphere. The radioactive inventorywithin containment is referred to as the"in-containment accident source term."

The expression "in-containment accident source term,"as used in this document, denotes the radioactivematerial composition and magnitude, as well as thechemical and physical properties of the material withinthe containment that are available for leakage from thereactor to the environment. The "in-containmentaccident source term" will normally be a function oftime and will involve consideration of fission productsbeing released from the core into the containment aswell as removal of fission products by plant featuresintended to do so (e.g., spray systems) or by naturalremoval processes.

For currently licensed plants, the characteristics of thefission product release from the core into thecontainment are set forth in Regulatory Guides 1.3 and1.4 (Refs. 2,3) and have been derived from the 1962report, TID-14844 (Ref. 4). This release consists of100% of the core inventory of noble gases and 50% ofthe iodines (half of which are assumed to deposit oninterior surfaces very rapidly). These values were basedlargely on experiments performed in the late 1950sinvolving heated irradiated U0 2 pellets. TID-14844also included 1% of the remaining solid fissionproducts, but these were dropped from consi-deration in Regulatory Guides 1.3 and 1.4. The 1% ofthe solid fission products are considered in certainareas such as equipment qualification.

Regulatory Guides 1.3 and 1.4 (Refs. 2 and 3) specifythat the source term within containment is assumed tobe instantaneously available for release and that theiodine chemical form is assumed to be predominantly

(91%) in elemental (02) form, with 5% assumed to beparticulate iodine and 4% assumed to be in organicform. These assumptions have significantly affected thedesign of engineered safety features. Containmentisolation valve closure times have also been affected bythese assumptions.

Use of the TID-14844 release has not been confined toan evaluation of site suitability and plant mitigationfeatures such as sprays and filtration systems. Theregulatory applications of this release are wide,including the basis for (1) the post-accident radiationenvironment for which safety-related equipment shouldbe qualified, (2) post-accident habitability requirementsfor the control room, and (3) post-accident samplingsystems and accessibility.

In contrast to the TID-14844 sourge term andcontainment leakage release used for design basisaccidents, severe accident releases to the environmentfirst arose in probabilistic risk assessments (e.g.,Reactor Safety Study, WASH-1400 (Ref. 5)) inexamining accident sequences that involved core meltand containments that could fail. Severe accidentreleases represent mechanistically determined bestestimate releases to the environment, includingestimates of failures of containment integrity. This isvery different from the combination of the non-mechanistic release to containment postulated byTID-14844 coupled with the assumption of very limitedcontainment leakage used for Part 100 siting calcula-tions for design basis accidents. The worst severeaccident releases resulting from containment failure orcontainment bypass can lead to consequences that aremuch greater than those associated with a TID-14844source term released into containment where thecontainment is assumed to be leaking at its maximumleak rate for its design conditions. Indeed, some of themost severe releases arise from some containmentbypass events, such as rupture of multiple steamgenerator tubes.

Although severe accident source terms have not beenused in individual plant licensing safety evaluations,they have had significant regulatory applications.Source terms from severe accidents (beyond-design-basis accidents) came into regulatory consideration andusage shortly after the issuance of WASH-1400 in 1975,and their application was accelerated after the ThreeMile Island accident in March 1979. Currentapplications rely to a large extent on the results ofWASH-1400 and include (1) part of the basis for thesizes of emergency planning zones for all plants, (2) thebasis for staff assessments of severe accident risk in

I NUREG-1465

plant environmental impact statements, and (3) part ofthe basis for staff prioritization and resolution ofgeneric safety issues, unresolved safety issues, andother regulatory analyses. Source term assessmentsbased on WASH-1400 methodology appear in manyprobabilistic risk assessment studies performed to date.

1.2 Research Insights SinceTID-14844

Source term estimates under severe accident conditionsbecame of great interest shortly after the Three-MileIsland (rMI) accident when it was observed that onlyrelatively small amounts of iodine were released to theenvironment compared with the amount predicted tobe released in licensing calculations. This led a numberof observers to claim that severe accident releases weremuch lower than previously estimated.

The NRC began a major research effort about 1981 toobtain a better understanding of fission-producttransport and release mechanisms in LWRs undersevere accident conditions. This research effort hasincluded extensive NRC staff and contractor effortsinvolving a number of national laboratories as well asnuclear industry groups. These cooperative researchactivities resulted in the development and applicationof a group of computer codes known as the SourceTerm Code Package (STCP) (Ref. 6) to examinecore-melt progression and fission product release andtransport in LWRs. The NRC staff has also sponsoredsignificant review efforts by peer reviewers, foreignpartners in NRC research programs, industry groups,and the general public. The STCP methodology forsevere accident source terms has also been reflected inNUREG-1150 (Ref. 7), which provides an updated riskassessment for five U.S. nuclear power plants.

As a result of the NRC's research effort to obtain abetter understanding of fission product transport andrelease mechanisms in LWRs under severe accidentconditions, the STCP emerged as an integral tool foranalysis of fission product transport in the reactorcoolant system (RCS) and containment. The STCPmodels release from the fuel with CORSOR (Ref. 8)and fission product retention and transport in the RCSwith TRAPMELT (Ref. 9). Releases from core-concreteinteractions are modeled using the VANESA andCORCON (Ref. 10) codes. Depending upon thecontainment type, SPARC or ICEDF (Refs. 11,12) areused in conjunction with NAUA (Ref.13) to model thetransport and retention of fission product releases fromthe RCS and from core-concrete interactions into thecontainment, with subsequent release of fissionproducts to the environment consistent with the stateof the containment.

Improved modeling of severe accident phenomena,including fission product transport, has been providedby the recently developed MELCOR (Ref. 14) code. Atthis time, however, an insufficient body of calculationsis available to provide detailed insights from this model.

Using analyses based on the STCP and MELCORcodes and NUREG-1150, the NRC has sponsoredstudies (Refs. 15-17) that analyzed the timing,magnitude, and duration of fission product releases. Inaddition, an examination and assessment of thechemical form of iodine likely to be found withincontainment as a result of a severe accident has alsobeen carried out (Ref. 18).

In contrast to the instantaneous releases that werepostulated in Regulatory Guides 1.3 and 1.4, analyses ofsevere accident sequences have shown that, despitedifferences in plant design and accident sequence, suchreleases can be generally categorized in terms ofphenomenological phases associated with the degree offuel melting and relocation, reactor pressure vesselintegrity, and, as applicable, attack upon concrete belowthe reactor cavity by molten core materials. Thegeneral phases, or progression, of a severe LWRaccident are shown in Table 1.1.

Table 1.1 Release Phases of a Severe Accident

Release Phases

Coolant Activity ReleaseGap Activity ReleaseEarly In-Vessel ReleaseEx-Vessel ReleaseLate In-Vessel Release

Initially there is a release of coolant activity associatedwith a break or leak in the reactor coolant system.Assuming that the coolant loss cannot be accommo-dated by the reactor coolant makeup systems or theemergency core cooling systems, fuel cladding failurewould occur with a release of the activity located in thegap between the fuel pellet and the fuel cladding.

As the accident progresses, fuel degradation begins,resulting in a loss of fuel geometry accompanied bygradual melting and slumping of core materials to thebottom of the reactor pressure vessel. During thisperiod, the early in-vessel release phase, virtually allthe noble gases and significant fractions of the volatilenuclides such as iodine and cesium are released intocontainment. The amounts of volatile nuclides releasedinto containment during the early in-vessel phase arestrongly influenced by the residence time of theradioactive material within the RCS during coredegradation. High pressure sequences result in long

NUREG-1465 2

residence times and significant retention and plateoutof volatile nuclides within the RCS, while low pressuresequences result in relatively short residence times andlittle retention within the RCS and consequently higherreleases into containment.

If failure of the bottom head of the reactor pressurevessel occurs, two additional release phases may occur.Molten core debris released from the reactor pressurevessel into the containment will interact with theconcrete structural materials of the cavity below thereactor (ex-vessel release phase). As a result of theseinteractions, quantities of the less volatile nuclides maybe released into containment. Ex-vessel releases areinfluenced somewhat by the type of concrete in thereactor cavity. Limestone concrete decomposes toproduce greater quantities of CO and CO2 gases thanbasaltic concrete. These gases may, in turn, spargesome of the less volatile nuclides, such as barium andstrontium, and small fractions of the lanthanides intothe containment atmosphere. Large quantities ofnon-radioactive aerosols may also be released as aresult of core-concrete interactions. The presence ofwater in the reactor cavity overlying any core debris cansignificantly reduce the ex-vessel releases (bothradioactive and non-radioactive) into the containment,either by cooling the core debris, or at least byscrubbing the releases and retaining a large fraction inthe water. The degree of scrubbing will depend, ofcourse, upon the depth and temperature of any wateroverlying the core debris. Simultaneously, andgenerally with a longer duration, late in-vessel releasesof some of the volatile nuclides, which had deposited inthe reactor coolant system during the in-vessel phase,will also occur and be released into containment.

Two other phenomena that affect the release of fissionproducts into containment could also occur, asdiscussed in Reference 7. The first of these is referredto as "high pressure melt ejection" (HPME). If theRCS is at high pressure at the time of failure of thebottom head of the reactor pressure vessel, quantitiesof molten core materials could be injected into thecontainment at high velocities. In addition to apotentially rapid rise in containment temperature, asignificant amount of radioactive material could also beadded to the containment atmosphere, primarily in theform of aerosols. The occurrence of HPME isprecluded at low RCS pressures. A secondphenomenon that could affect the release of fissionproducts into containment is a possible steam explosionas a result of interactions between molten core debrisand water. This could lead to fine fragmentation ofsome portion of the molten core debris with an increasein the amount of airborne fission products. While smallscale steam explosions are considered quite likely tooccur, they will not result in significant increases in the

airborne activity already within containment. Largescale steam explosions, on the other hand, could resultin significant increases in airborne activity, but aremuch less likely to occur. In any event, releases ofparticulates or vapors during steam explosions will alsobe accompanied by large amounts of water droplets,which would tend to quickly sweep released materialfrom the atmosphere.

2 OBJECTIVES AND SCOPE

2.1 GeneralThe primary objective of this report is to define arevised accident source term for regulatory applicationfor future LWRs. The intent is to capture the majorrelevant insights available from recent severe accidentresearch on the phenomenology of fission productrelease and transport behavior. The revised sourceterm is expressed in terms of times and rates ofappearance of radioactive fission products into thecontainment, the types and quantities of the speciesreleased, and other important attributes such as thechemical forms of iodine. This mechanistic approachwill therefore present, for regulatory purposes, a morerealistic portrayal of the amount of fission productspresent in the containment from a postulated severeaccident.

2.2 Accidents ConsideredIn order to determine accident source terms forregulatory purposes, a range of severe accidents thathave been analyzed for LWR plants was examined.Evaluation of a range of severe accident sequences wasbased upon work done in support of NUREG-1150(Ref. 7). This work is documented in NUREG/CR-5747(Ref. 17) and employed the integrated Source TermCode Package (STCP) computer codes, together withinsights from the MELCOR code, which were used toanalyze specific accident sequences of interest toprovide the accident chronology as well as detailedestimates of fission product behavior within the reactorcoolant system and the other pertinent parts of theplant. The sequences studied progressed to a completecore melt, involving failure of the reactor pressurevessel and including core-concrete interactions, as well.

A key decision to be made in defining an accidentsource term is the severity of the accident or group ofaccidents to be considered. Footnote 1 to 10 CFR Part100 (Ref. 1). in referring to the postulated fissionproduct release to be used for evaluating sites, notesthat "Such accidents have generally been assumed toresult in substantial meltdown of the core withsubsequent release of appreciable quantities of fissionproducts." Possible choices range from (1) slight fueldamage accidents involving releases into containment

3 NUREG-1465

of a small fraction of the volatile nuclides such as thenoble gases, (2) severe core damage accidents involvingmajor fuel damage but without reactor vessel failure orcore-concrete interactions (similar in severity to theTMI accident), or (3) complete core-melt events withcore-concrete interactions. These outcomes are notequally probable. Since many reactor systems must failfor core degradation with reactor vessel failure to occurand core-concrete interactions to occur, one or moresystems may be returned to an operable status beforecore melt commences. Hence, past operational andaccident experience together with information onmodern plant designs, together with a vigorous programaimed at developing accident management procedures,indicate that complete core-melt events resulting inreactor pressure vessel failure are considerably lesslikely to occur than those involving major fuel damagewithout reactor pressure vessel failure, and that these,in turn, are less likely to occur than those involvingslight fuel damage.

For completeness, this report displays the mean oraverage release fractions for all the release phasesassociated with a complete core melt. However, it isconcluded that any source term selected for a particularregulatory application should appropriately reflect thelikelihood associated with its occurrence.

It is important to emphasize that the release fractionsfor the source terms presented in this report areintended to be representative or typical, rather thanconservative or bounding values, of those associatedwith a low pressure core-melt accident, except for theinitial appearance of fission products from failed fuel,which was chosen conservatively. The release fractionsare not intended to envelope all potential severeaccident sequences, nor to represent any singlesequence, since accident sequences yielding both higheras well as lower release fractions were examined andfactored into the final report presented here.


The NRC staff also intends to allow credit for removalor reduction of fission products within containment viaengineered features provided for fission productreduction such as sprays or filters, as well as by naturalprocesses such as aerosol deposition. These arediscussed in Section 5.

2.3 LimitationsThe accident source terms defined in this report havebeen derived from examination of a set of severeaccident sequences for LWRs of current design.Because of general similarities in plant and core designparameters, these results are also considered to beapplicable to evolutionary LWR designs such asGeneral Electric's Advanced Boiling Water Reactor(ABWR) and Combustion Engineering's (CE) System80+.

Currently, the NRC staff is reviewing reactor designsfor several smaller LWRs employing some passivefeatures for core cooling and containment heatremoval. While the "passive" plants are generallysimilar to present LWRs, they are expected to havesomewhat lower core power densities than those ofcurrent LWRs. Hence, an accident for the passiveplants similar to those used in this study would likelyextend over a longer time span. For this reason, thetiming and duration values provided in the releasetables given in Section 3.3 are probably shorter thanthose applicable to the passive plants. The releasefractions shown may also be overestimated somewhatfor high pressure sequences associated with the passiveplants, since longer times for accident progressionwould also allow for enhanced retention of fissionproducts in the primary coolant system during coreheatup and degradation. Despite the lack of specificaccident sequence information for these designs, thein-containment accident source terms provided belowmay be considered generally applicable to the "passive"designs.

The accident source terms provided in this report arenot considered applicable to reactor designs that arevery different from LWRs, such as high-temperaturegas-cooled reactors or liquid-metal reactors.

Recent information has indicated that high burnup fuel,that is, fuel irradiated at levels in excess of about 40GWD/MTU, may be more prone to failure duringdesign basis reactivity insertion accidents (RIA) thanpreviously thought. Preliminary indications are thathigh burnup fuel also may be in a highly fragmented orpowdered form, so that failure of the cladding couldresult in a significant fraction of the fuel itself beingreleased. In contrast, the source term contained in thisreport is based upon fuel behavior results obtained atlower burnup levels where the fuel pellet remainsintact upon cladding failure, resulting in a release onlyof those fission product gases residing in the gapbetween the fuel pellet and the cladding. Because ofthis recent information regarding high burnup fuels, theNRC staff cautions that, until further informationindicates otherwise, the source term in this report(particularly gap activity) may not be applicable for fuel

NUREG-1465 4

--- - --

irradiated to high burnup levels (in excess of about 40GWDIMTU).

3 ACCIDENT SOURCE TERMS

The expression "in-containment source terms," as usedin this report, denotes the fission product inventorypresent in the containment at any given time during anaccident. lb evaluate the in-containment source termduring the course of an accident, the time-history of thefission product release from the core into thecontainment must be known, as well as the effect offission product removal mechanisms, both natural andengineered, to remove radioactive materials from thecontainment atmosphere. This section discusses thetime-history of the fission product releases into thecontainment. Removal mechanisms are discussed inSection 5.

3.1 Accident Sequences ReviewedAll the accident sequences identified in NUREG-1150were reviewed and some additional Source lbrm CodePackage (STCP) and MELCOR calculations werep&formed. The dominant sequences which are

considered to significantly impact the source term aresummarized in Thble 3.1 for BWRs and Table 3.2 forPWRs.

3.2 Onset of Fission Product ReleaseThis section discusses the assumptions used in selectingthe scenario appropriate for defining the early phasesof the source term (coolant activity and gap releasephases). It was considered appropriate to base theseearly release phases on the design basis initiation thatcould lead to earliest fuel failures.

A review of current plant final safety analysis reports(FSARs) was made to identify all design basis accidentsin which the licensee had identified fuel failure. For allaccidents with the potential for release of radioactivityinto the environment, the class of accident that had theshortest time until the first fuel rod failed was thedesign basis LOCA. As might be expected, the timeuntil cladding failure is very sensitive to the design ofthe reactor, the type of accident assumed, and the fuelrod design. In particular, the maximum linear heatgeneration rate, the internal fuel rod pressure, and thestored energy in the fuel rod are significantconsiderations.

Table 3.1 BWR Source Term Contributing Sequences

Plant Sequence Description

Peach Bottom TC1 ATWS with reactor depressurizedTC2 ATWS with reactor pressurizedTC3 TC2 with wetwell ventingTB1 SBO with battery depletionTB2 TB1 with containment failure at vessel failureS2E1 LOCA (2"), no ECCS and no ADSS2E2 S2E1 with basaltic concrete

V RHR pipe failure outside containmentTBUX SBO with loss of all DC power

LaSalle TB SBO with late containment failure

Grand Gulf TC ATWS early containment failure fails ECCSTBI SBO with battery depletionTB2 TB1 with H2 burn fails containmentTBS SBO, no ECCS but reactor depressurizedTBR TBS with AC recovery after vessel failure

SBO Station Blackout LOCA Loss of Coolant AccidentRCP Reactor Coolant Pump RHR Residual Heat RemovalADS Automatic Depressurization System ATWS Anticipated Transient Without Scram

5 NUREG-1465

Table 3.2 PIVR Source Term Contributing Sequences

Plant Sequence Description

Surry AG LOCA (hot leg), no containment heat removal systemsTMLB' LOOP, no PCS and no AFWS

V Interfacing system LOCAM3B SBO with RCP seal LOCA

S2D-8 SBLOCA, no ECCS and H2 combustionS2D-p SBLOCA with 6" hole in containment

Zion S2DCR LOCA (2"), no ECCS no CSRSS2DCF1 LOCA RCP seal, no ECCS, no containment sprays,

no coolers-H 2 burn or DCH fails containmentS2DCF2 S2DCF1 except late H2 or overpressure failure of

containmentTMLU Transient, no PCS, no ECCS, no AFWS-DCH fails

containment

Oconee 3 TMLB' SBO, no active ESF systemsS1DCF LOCA (3"), no ESF systems

Sequoyah S3HF1 LOCA RCP, no ECCS, no CSRS with reactor cavityflooded

S3HF2 S3HF1 with hot leg induced LOCA3HF3 S3HF1 with dry reactor cavity

M3B LOCA (3") with SBOTBA SBO induces hot leg LOCA-hydrogen burn fails

containmentACD LOCA (hot leg), no ECCS no CS

M3B1 SBO delayed 4 RCP seal failures, only steam drivenAFW operates

S3HF LOCA (RCP seal), no ECCS, no CSRSS3H LOCA (RCP seal) no ECC recirculation

SBO Station Blackout LOCA Loss of Coolant AccidentRCP Reactor Coolant Pump DCH Direct Containment HeatingPCS Power Conversion System ESF Engineered Safety FeatureCS Containment Spray CSRS CS Recirculation SystemATWS Anticipated Transient Without Scram LOOP Loss of Offsite Power

The details of the specific accident sequences are documented in NUREG/CR-5747, Estimate ofRadionuclide Release Characteristics into Containment Under Severe Accident Conditions (Ref. 17).

To determine whether a design basis LOCA was areasonable scenario upon which to base the timing ofinitial fission product release into the containment,various PRAs were reviewed to determine thecontribution to core damage frequency (CDF) resultingfrom LOCAs. This information is shown in Table 3.3.As can be seen from this table, LOCAs are a smallcontributor to CDF for BWRs, but can be a substantialcontributor for PWRs. Therefore, for PWRs a large

LOCA is considered a reasonable initiator to assumefor modeling the earliest appearance of the gap activityif the plant has not been approved for leak beforebreak (LBB) operation. For plants that have receivedLBB approval, a small LOCA (6" line break) wouldmore appropriately model the timing. For BWRs, largeLOCAs may not be an appropriate scenario for gapactivity timing. However, since the time to initial fuelrod failure is long for BWRs, even for large LOCAs,

NUREG-1465 6

use of the large LOCA scenario should not undulypenalize BWRs and will maintain consistency with theassumptions for the PWR. As with the PWR, for anLBB approved plant, the timing associated with a smallLOCA (6" line break) would be more appropriate.

In order to provide a realistic estimate of the shortesttime for fuel rod failure for the LOCA, calculationswere performed using the FRAPCON2,SCDAP/RELAP5 MOD 3.0, and FRAPT6 computercodes for two plants. The two plants were a Babcockand Wilcox (B&W) plant with a 15 by 15 fuel rod arrayand a Westinghouse 4-loop (M!) plant with a 17 by 17fuel rod array. For each plant, a sensitivity study was

performed to identify the size of the LOCA thatresulted in the shortest fuel rod failure time (Ref. 15).In both cases, the accident was a double-endedguillotine rupture of the cold leg pipe. The minimumtime from the time of accident initiation until first fuelrod failure was calculated to be 13 and 24.6 seconds forthe B&W and _ plants, respectively. A sensitivitystudy was performed to determine the effect of trippingor not tripping the reactor coolant pumps. The resultsindicated that tripping of the reactor coolant pumpshad no appreciable impact on timing. For a 6-inch linebreak, the time until first fuel rod failure is expected tobe greater than 6.5 and 10 minutes, respectively.

Table 33 Contribution of LOCAs to Core Damage Frequency (CDF)-Internal Events

Percent or CDFPercent of CDF caused by large LOCAs

Boiling Water Reactors caused by LOCAs (>6" line break)

Peach Bottom (NUREG-1150) 3.5 1.0Grand Gulf (NUREG-1150) 0.1 0.03Millstone 1 (Utility) 23 13

Pressurized W'ater Reactors

Surry (NUREG-1150) 15 4.3Sequoyah (NUREG-1150) 63 4.6Zion (NUREG-1150) 87 1.4Calvert Cliffs (IREP) 21 < 1Oconee-3 (EPRI/NSAC) 43 3.0

A comparison calculation was done using the TRAC-PF1 MOD I code, version 14.3U5Q.LG on the Wplant. This analysis indicated that the first fuel rodfailure would occur 34.9 seconds after pipe rupture, incontrast to the value of 24.6 seconds calculated usingSCDAPIRELAP. The reasons for the differencebetween the SCDAP/RELAPS MOD 3.0 andTRAC-PF1 MOD 1 are discussed in Reference 15.

The review of the FSARs for BWRs indicates that fuelfailures may occur significantly later, on the order ofseveral minutes or more. No calculations have beenperformed using the aforementioned suite of codes.

For determining the time of appearance of gap activityin the containment (i.e., initial fuel failure), whichcorresponds to the duration of the coolant activityphase and the beginning of the gap activity phase, itwould be appropriate to perform a plant specificcalculation using the codes described above. However,if no plant specific calculations are performed, theminimum times discussed above may be used to providean estimate of the earliest time to fuel rod failure.


3.3 Duration of Release PhasesSection 1.2 provided a qualitative discussion of therelease phases of an accident. This section providesestimated durations for these release phases.

The coolant activity phase begins with a postulated piperupture and ends when the first fuel rod has beenestimated to fail. During this phase, the activityreleased to the containment atmosphere is thatassociated with very small amounts of radioactivitydissolved in the coolant itself. As discussed in Section3.2 above, this phase is estimated to last about 25seconds for Westinghouse PWRs, and about 13 secondsfor B&W PWRs, assuming a large break LOCA. For asmaller LOCA (e.g., a 6-inch line break), such as would

7 NUREG-1465

be considered for a plant that has received LBBapproval, the coolant activity phase duration would beexpected to be at least 10 minutes. Although notspecifically evaluated at this time, CombustionEngineering (CE) PWRs would be expected to havecoolant activity durations similar to Westinghouseplants. For BWRs, the coolant activity phase would beexpected to last longer; however, unless plant specificcalculations are made, the durations discussed aboveare considered applicable.

The gap activity release phase begins when fuelcladding failure commences. This phase involves therelease of that radioactivity that has collected in thegap between the fuel pellet and cladding. This processreleases to containment a few percent of the totalinventory of the more volatile radionuclides,particularly noble gases, iodine, and cesium. During thisphase, the bulk of the fission products continue to beretained in the fuel itself. The gap activity phase endswhen the fuel pellet bulk temperature has been raisedsufficiently that significant amounts of fission productscan no longer be retained in the fuel. As noted inReference 16, a review of STCP calculated results forsix reference plants, PWRs as well as BWRs, indicatedthat significant fission product releases from the bulk ofthe fuel itself were estimated to commence no earlier

than about 30 minutes and 60 minutes for PWRs andBWRs, respectively, after the onset of the accident.However, more recent calculations (Ref. 19) for thePeach Bottom plant using the MELCOR codeindicated that the durations of the gap release for threeBWR accident sequences were about 30 minutes, aswell. On this basis, the duration of the gap activityrelease phase has been selected to be 0.5 hours, forboth PWRs and BWRs.

During the early in-vessel release phase, the fuel aswell as other structural materials in the core reachsufficiently high temperatures that the reactor coregeometry is no longer maintained and fuel and other'materials melt and relocate to the bottom of thereactor pressure vessel. During this phase, significantquantities of the volatile nuclides in the core inventoryas well as small fractions of the less volatile nuclidesare estimated to be released into containment. Thisrelease phase ends when the bottom head of thereactor pressure vessel fails, allowing molten coredebris to fall onto the concrete below the reactorpressure vessel. Release durations for this phase varydepending on both the reactor type and the accidentsequence. Tables 3.4 and 3.5, based on results fromReference 16, show the estimated duration times forPWRs and BWRs, respectively.

Table 3.4 In-Vessel Release Duration for PWR Sequences

Release DurationPlant Accident Sequence (Min)

Surry TMLB' (H) 41Surry S3B (H) 36Surry AG (L) 215Surry V (L) 104Zion TMLU (H) 41Zion S2DCR/S2DCF (H) 39Sequoyah S3HF/S3B (H) 46Sequoyah S3B1 (H) 75Sequoyah TMLB' (H) 37Sequoyah TBA (L) 195Sequoyah ACD (L) 73Oconee TMLB' (H) 35Oconee SPDCF (L) 84

*(H or L) Denotes whether the accident occurs at high or low pressure.

Based on the information in these tables, the staffconcludes that the in-vessel release phase is somewhatlonger for BWR plants than for PWR plants. This islargely due to the lower core power density in BWRplants that extends the time for complete core melt.Representative times for the duration of the in-vessel

release phase have been selected to be 1.3 hours and1.5 hours, for PWR and BWR plants respectively, asrecommended by Reference 17.

The ex-vessel release phase begins when molten coredebris exits the reactor pressure vessel and ends when

NUREG-1465 &

Table 3.5 In-Vessel Release Duration for BWR Sequences

Release DurationPlant Accident Sequence* (Min)

Peach Bottom TC2 (H) 66

Peach Bottom TC3 68

Peach Bottom TC1 (L) 97

Peach Bottom TB1ITB2 (H) 91

Peach Bottom V (L) 69

Peach Bottom S2E (H) 81

Peach Bottom TBUX (H) 67

LaSalle TB (H) 81

Grand Gulf TB (H) 122

Grand Gulf TC1 (L) 130

Grand Gulf TBS/TBR (L) 96

*(H or L) denotes whether the accident occurs at high or low pressure.

the debris has cooled sufficiently that significantquantities of fission products are no longer beingreleased. During this phase, significant quantities of thevolatile radionuclides not already released during theearly in-vessel phase as well as lesser quantities ofnon-volatile radionuclides are released intocontainment. Although releases from core-concreteinteractions are predicted to take place over a numberof hours after vessel breach, Reference 16 indicatesthat the bulk of the fission products (about 90%), withthe exception of tellurium and ruthenium, are expectedto be released over a 2-hour period for PWRs and a3-hour period for BWRs. For tellurium and ruthenium,ex-vessel releases extend over 5 and 6 hours,respectively, for PWRs and BWRs. The difference induration of the ex-vessel phase between PWRs andBWRs is largely attributable to the larger amount ofzirconium in BWRs, which provides additional chemicalenergy of oxidation. Based on Reference 17, theex-vessel release phase duration is taken to be 2 and 3hours, respectively, for PWRs and BWRs.

The late in-vessel release phase commences at vesselbreach and proceeds simultaneously with theoccurrence of the ex-vessel phase. However, theduration is not the same for both phases. During thisrelease phase, some of the volatile nuclides depositedwithin the reactor coolant system earlier during coredegradation and melting may re-volatilize and bereleased into containment. Reference 17, after a reviewof the source term uncertainty methodology used inNUREG-1150 (Ref. 7), estimates the late in-vessel

release phase to have a duration of 10 hours. This valuehas been selected for this report.

A summary of the release phases and the selectedduration times for PWRs and BWRs is shown forreference purposes in Table 3.6.

Table 3.6Release Phase Durations for PWRs and BWRs

Duration, Duration,PWRs BWRs

Release Phase (Hours) (Hours)

Coolant Activity 10 to 30 seconds 30 seconds'Gap Activity 0.5 0.5Early In-Vessel 1.3 1.5Ex-Vessel 2 3Late In-Vessel 10 10

Without approval for leak-before-break. Coolantactivity phase duration is assumed to be 10 minuteswith leak-before-break approval.

3.4 Fission Product Composition andMagnitude

In considering severe accidents in which the contain-ment might fail, WASH-1400 (Ref. 5) examined thespectrum of fission products and grouped 54 radionu-clides into 7 major groups on the basis of similarity inchemical behavior. The effort associated with the STCP

9 NUREG-1465

further analyzed these groupings and expanded the 7fission product groups into 9 groups. These are shownin Table 3.7.

Table 3.7 STCP Radionuclide Groups

Group Elements

1 Xe, Kr2 I, Br3 Cs, Rb4 Te, Sb, Se5 Sr6 Ru, Rh, Pd, Mo, 'T7 La, Zr, Nd, Eu, Nb, Pm, Pr, Sm, Y8 Ce, Pu, Np9 Ba

Similarly, low pressure sequences cause aerosolsgenerated within the RCS to be swept out rapidlywithout significant retention within the RCS, therebyresulting in higher release fractions from the core intocontainment.

Table 3.8 Revised Radionuclide Groups

Group Title Elements in Group

1 Noble gases Xe, Kr2 Halogens I, Br3 Alkali Metals Cs, Rb4 Tellurium group 'T, Sb, Se5 Barium, strontium Ba, Sr6 Noble Metals Ru, Rh, Pd, Mo, Tc, Co7 Lanthanides La, Zr, Nd, Eu, Nb, Pm,

Pr, Sm, Y, Cm, Am8 Cerium group Ce, Pu, Np

Both the results of the STCP analyses and theuncertainty analysis (using the results of theNUREG-1150 source term expert panel elicitation)reported in NUREG/CR-5747 (Ref. 17) indicate onlyminor differences between Ba and Sr releases. Hence,a revised grouping of radionuclides has been developedthat groups Ba and Sr together. The relativeimportance to offsite health and economicconsequences of the radioactive elements in a nuclearreactor core has been examined and documented inNUREG/CR-4467 (Ref. 20). In addition to theelements already included in Thble 3.7, Reference 20found that other elements such as Curium could beimportant for radiological consequences if released insufficiently large quantities. For this reason, group 7has been revised to include Curium (Cm) andAmericium (Am), while group 6 has been revised toinclude Cobalt (Co). The revised radionuclide groupsused in this report including revised titles and theelements comprising each group are shown in Table 3.8.

Source term releases into the containment wereevaluated by reactor type, i.e., BWR or PWR, from thesequences in NUREG-1150 and the supplementalSTCP calculations discussed in Section 3.1.

Releases into containment during the early in-vesselphase, prior to reactor pressure vessel failure, aremarkedly affected by retention in the RCS, which is afunction of the residence time in the RCS during coredegradation. High pressure in the RCS during coredegradation allows for longer residence time ofaerosols released from the core. This, in turn, permitsincreased retention of aerosols within the RCS andlower releases from the core into the containment.

The relative frequency of occurrence of high vs. lowpressure sequences were examined for both BWRs andPWRs. The results of this survey are shown inThble 3.9, and they indicate that a significant fraction ofthe sequences examined, in terms of frequency,occurred at low pressure. In addition, advanced PWRdesigns are increasingly incorporating safety-gradedepressurization systems, primarily to minimize thelikelihood of high pressure melt ejection (HPME) withits associated high containment atmosphere heat loadsand large amounts of atmospheric aerosols.

For these reasons, the composition and magnitude ofthe source term has been chosen to be representativeof conditions associated with low pressure in the RCSat the time of reactor core degradation and pressurevessel failure. Reference 17 provides estimates of themean core fractions released into containment, asestimated by NUREG-1150 (Ref. 7), for accidentsequences occurring under low RCS pressure and highzirconium oxidation conditions. These are shown inTables 3.10 and 3.11.

3.5 Chemical FormThe chemical form of iodine and its subsequentbehavior after entering containment from the reactorcoolant system have been documented inNUREG/CR-5732, Iodine Chemical Forms in LWRSevere Accidents (Ref. 18) and in ORNLITM-12202,"Models of Iodine Behavior in Reactor Containments,"(Ref. 21).

NUREG-1465 10

Table 3.9 Fraction of mean core damage frequency with high, intermediate, and low pressure sequences(internal events only unless otherwise noted)

Low Pressure atHigh Pressure at Intermed. press. Vessel Breach

Boiling Water Reactors Vessel Breach at vessel breach (<200 psi) No vessel breach

LaSalle-external events only 0.27 N/A 0.67 0.06

LaSalle-internal events only 0.19 N/A 0.62 0.19Grand Gulf 0.28 N/A 0.51 0.21Peach Bottom 0.51 N/A 0.41 0.08Pressurized Water Reactors

Surry 0.06 0.07 0.37 0.50Sequoyah 0.14 0.21 0.24 0.41Zion 0.03 0.15 0.72 0.10

Table 3.10 Mean Values or Radionuclides Into Containment for BW'Rs,Low RCS Pressure, High Zirconium Oxidation

Nuclide Early In-Vessel Ex-Vessel Late In-vessel

N.G. 1.0 0 0I 0.27 0.37 0.07Cs 0.2 0.45 0.03le 0.11 0.38 0.01Sr 0.03 0.24 0Ba 0.03 0.21 0Ru 0.007 0.004 0La 0.002 0.01 0Ce 0.009 0.01 0

Table 3.11 Mean Values of Radionuclide Releases Into Containment for PWRs,Low RCS Pressure, High Zirconium Oxidation

Nuclide Early In-Vessel Ex-Vessel Late In-vessel

N.G. 1.0 0 0I 0.4 0.29 0.07Cs 0.3 0.39 0.06,lb 0.15 0.29 0.025Sr 0.03 0.12 0Ba 0.04 0.1 0Ru 0.008 0.004 0La 0.002 0.015 0Ce 0.01 0.02 0

11 NUREG-1465

The results from Ref. 18 indicate that iodine enteringthe containment is at least 95% CsI with the remaining5% as I plus HI, with not less than 1% of each as I andHI. Once the iodine enters containment, however,additional reactions are likely to occur. In an aqueousenvironment, as expected for LWRs, iodine is expectedto dissolve in water pools or plate out on wet surfacesin ionic form as I-. Subsequently, iodine behaviorwithin containment depends on the time and pH of thewater solutions. Because of the presence of otherdissolved fission products, radiolysis is expected tooccur and lower the pH of the water pools. Without anypH control, the results indicate that large fractions ofthe dissolved iodine will be converted to elementaliodine and be released to the containment atmosphere.However, if the pH is controlled and maintained at avalue of 7 or greater, very little (less than 1%) of thedissolved iodine will be converted to elemental iodine.Some considerations in achieving pH control arediscussed in NUREG/CR-5950, "Iodine Evolution andpH Control," (Ref. 22).

Organic compounds of iodine, such as methyl iodide,CH31, can also be produced over time largely as aresult of elemental iodine reactions with organicmaterials. Organic iodide formation as a result ofreactor accidents has been surveyed in WASH-1233,"Review of Organic Iodide Formation Under AccidentConditions in Water-Cooled Reactors," (Ref. 23), andmore recently in NUREG/CR-4327, "Organic IodideFormation Following Nuclear Reactor Accidents,"(Ref.24). From an analysis of a number of containmentexperiments, WASH-1233 concluded that, consideringboth non-radiolytic as well as radiolytic means, no morethan 3.2 percent of the airborne iodine would beconverted to organic iodides during the first two hoursfollowing a fission product release. The value of 3.2percent was noted as a conservative upper limit and wasjudged to be considerably less, since it did not account,among other things, for decreased radiolytic formationof organic iodide due to iodine removal mechanismswithin containment. Reference 24 also included resultsinvolving irradiated fuel elements, and concluded thatthe organic iodide concentration within containmentwould be about 1 percent of the iodine releaseconcentration over a wide range of iodineconcentrations.

A conversion of 4 percent of the elemental iodine toorganic has been implicitly assumed by the NRC staff inRegulatory Guides 1.3 and 1.4, based upon an upperbound evaluation of the results in WASH-1233.However, in view of the results of Ref. 23 that aconversion of 3.2 percent is unduly conservative, avalue of 3 percent is considered more realistic and willbe used in this report. Where the pH is controlled at

values of 7 or greater within the containment,elemental iodine can be taken as comprising no morethan 5 percent of the total iodine released, and iodinein organic form may be taken as comprising no greaterthan 0.15 percent (3 percent of 5 percent) of the totaliodine released.

Organic iodide formation in BWRs versus PWRs is notnotably different. Reference 18 examined not onlyiodine entering containment as CsI; but also consideredother reactions that might lead to volatile forms ofiodine within containment, such as reactions of CsOHwith surfaces and revaporization of CsI from RCSsurfaces. Reference 18 indicates (Thble 2.4) that for thePeach Bottom TC2 sequence, the estimated percentageof iodine as HI was 3.2 percent, not notably less thanthe PWR sequences examined. While organic iodide isformed largely from reactions of elemental iodine, Ref.22 clearly notes that reactions with HI may beimportant.

Although organic iodine is not readily removed bycontainment sprays or filter systems, it is undulyconservative to assume that organic iodine is notremoved at all from the containment atmosphere, oncegenerated, since such an assumption can result in anoverestimate of long-term doses to the thyroid.References 23 and 24 discuss the radiolytic destructionof organic iodide, and Standard Review Plan Section(S.R.P.) 6.5.2 notes the above reference and indicatesthat removal of organic iodide may be considered on acase-by-case basis. A rational model for organic iodinebehavior within containment would consider both itsformation as well as destruction in a time-dependentfashion. Development of such a model, however, isbeyond the scope of the present report.

Clearly, where the pH is not controlled to values of 7or greater, significantly larger fractions of elementaliodine, as well as organic iodine may be expected withincontainment.

All other fission products, except for the noble gasesand iodine, discussed above, are expected to be inparticulate form.

3.6 Proposed Accident Source TermsThe proposed accident source terms, including theirtiming as well as duration, are listed in Thbles 3.12 forBWRs and 3.13 for PWRs. The information for thesetables was derived from the simplification of theNUREG-1150 (Ref. 7) source terms documented inNUREG/CR-5747 (Ref. 17). It should also be notedthat the rate of release of fission products into thecontainment is assumed to be constant during theduration time shown.

NUREG-1465 12

Table 3.12 BWR Releases Into Containment*

Gap Release*** Early In-Vessel Ex-Vessel Late In-Vessel

Duration (Hours) 0.5 1.5 3.0 10.0

Noble Gases* 0.05 0.95 0 0Halogens 0.05 0.25 0.30 0.01

Alkali Metals 0.05 0.20 0.35 0.01

Tellurium group 0 0.05 0.25 0.005Barium, Strontium 0 0.02 0.1 0

Noble Metals 0 0.0025 0.0025 0

Cerium group 0 0.0005 0.005 0Lanthanides 0 0.0002 0.005 0

* Values shown are fractions of core inventory.* See Table 3.8 for a listing of the elements in each group

*** Gap release is 3 percent if long-term fuel cooling is maintained.

Table 3.13 PWR Releases Into Containment

Gap Release*** Early In-Vessel Ex-Vessel Late In-Vessel

Duration (Hours) 0.5 1.3 2.0 10.0

Noble Gases* 0.05 0.95 0 0

Halogens 0.05 0.35 0.25 0.1

Alkali Metals 0.05 0.25 0.35 0.1

Tellurium group 0 0.05 0.25 0.005Barium, Strontium 0 0.02 0.1 0Noble Metals 0 0.0025 0.0025 0

Cerium group 0 0.0005 0.005 0Lanthanides 0 0.0002 0.005 0

Values shown are fractions of core inventory.See Table 3.8 for a listing of the elements in each group

* Gap release is 3 percent if long-term fuel cooling is maintained.

It is emphasized that the release fractions for thesource terms presented in this report are intended tobe representative or typical, rather than conservative orbounding values, of those associated with a lowpressure core-melt accident, except for the initialappearance of fission products from failed fuel, whichwas chosen conservatively. The release fractions are notintended to envelope all potential severe accidentsequences, nor to represent any single sequence.

Tibles 3.12 and 3.13 in this, the final report, weremodified from the tables in the draft report which weretaken from Table 3.9 and Table 3.10, for BWRs and

PWRs, respectively. The changes and the reasons forthese was as follows:

1. BWR in-vessel release fractions for the volatilenuclides (I and Cs) increased slightly whileex-vessel release fractions for the same nuclideswas reduced as a result of comments received andadditional MELCOR calculations available afterissuance of the draft report. The total I and Csreleased into containment over all phases of theaccident remained the same.

2. Release fractions for Te, Ba and Sr were reducedsomewhat, both for in-vessel as well as ex-vesselreleases, in response to comments.

13 NUREG-1465

3. Release fractions for the non-volatile nuclides,particularly during the early in-vessel phase werereduced significantly based on additional researchresults (Ref. 25) since issuance of NUREG-1150which indicate that releases of low volatilenuclides, both in-vessel as well as ex-vessel, havebeen overestimated. A re-examination in responseto comments received showed that the supposed"means" of the uncertainty distribution were inexcess of other measures of the distribution, suchas the 75th percentile. In this case, the 75thpercentile was selected as an appropriate measureof the release fraction. For additional discussionon this topic, see Section 4.4.

4. Gap activity release fractions were reduced from 5percent to 3 percent for accidents not involvingdegraded or molten core conditions, and wherelong-term fuel cooling is maintained. Seeadditional discussion below.

Based on WASH-1400 (Ref. 5), the inventory of fissionproducts residing in the gap between the fuel and thecladding is no greater than 3 percent except for cesium,which was estimated to be about 5 percent.NUREG/CR-4881 (Ref.16) reported a comparison ofmore recently available estimations and observationsindicating that releases of the dominant fission productgroups were generally below the values reported inReference 5. However, the magnitude of fissionproducts released during the gap release phase canvary, depending upon the type of accident. Accidentswhere fuel failures occur may be grouped as follows:

1. Accidents where long-term fuel cooling ismaintained despite fuel failure. Examples includethe design basis LOCA where ECCS functions,and a postulated spent fuel handling accident. Forthis category, fuel failure is taken to result in animmediate release, based upon References 5 and16, of 3 percent of the volatile fission products(noble gases, iodine, and cesium) which are in thegap between the fuel pellet and the cladding. Nosubsequent appreciable release from the fuelpellet occurs, since the fuel does not experienceprolonged high temperatures.

2. Accidents where long-term fuel cooling or coregeometry are not maintained. Examples includedegraded core or core-melt accidents, includingthe postulated limiting design basis fission productrelease into containment used to show compliancewith 10 CFR Part 100. For this category, the gaprelease phase may overlap to some degree withthe early in-vessel release phase. The releasemagnitude has been taken as an initial release of 3percent of the volatiles (as for category 1), plus an

additional release of 2 percent over the durationof the gap release phase.

3. Accidents where fuel failure results from reactivityinsertion accidents (RIA), such as the postulatedrod ejection (PWR) or rod drop (BWR) accidents.The accidents examined in this report do notcontain information on reactivity inducedaccidents to permit a quantitative discussion offission product releases from them. Hence, thegap release magnitude presented in Tables 3.12and 3.13 may not be applicable to fission productreleases resulting from reactivity insertionaccidents.

Recent information has indicated that high burnup fuel,that is, fuel irradiated at levels in excess of about 40GWD/MTU, may be more prone to failure duringdesign basis reactivity insertion accidents thanpreviously thought. Preliminary indications are thathigh bumup fuel also may be in a highly fragmented orpowdered form, so that failure of the cladding couldresult in a significant fraction of the fuel itself beingreleased. In contrast, the source term contained in thisreport is based upon fuel behavior results obtained atlower burnup levels where the fuel pellet remainsintact upon cladding failure, resulting in a release onlyof those fission product gases residing in the gapbetween the fuel pellet and the cladding. Because ofthis recent information regarding high burnup fuels, theNRC staff cautions that, until further informationindicates otherwise, the source term in Tables 3.12 and3.13 (particularly gap activity) may not be applicable forfuel irradiated to high burnup levels (in excess of about40 GWD/MTU).

With regard to the ex-vessel releases associated withcore-concrete interactions, according to Reference 17,there were only slight differences in the fissionproducts released into containment between limestonevs. basaltic concrete. Hence, the table shows thereleases only for a limestone concrete. Further, thereleases shown for the ex-vessel phase are assumed tobe for a dry reactor cavity having no water overlying anycore debris. Where water covers the core debris,aerosol scrubbing will take place and reduce thequantity of aerosols entering the containmentatmosphere. See Section 5.4 for further information.

3.7 Nonradioactive AerosolsIn addition to the fission product releases intocontainment shown in Tables 3.12 and 3.13, quantitiesof nonradioactive or relatively low activity aerosols willalso be released into containment. These aerosols arisefrom core structural and control rod materials releasedduring the in-vessel phase and from concrete decompo-sition products during the ex-vessel phase. A detailed

NUREG-1465 14

analysis of the quantity of nonfission product aerosolsreleased into containment was not undertaken. Preciseestimates of the masses of non-radioactive aerosolsreleased into containment are difficult to determine.

Reference 26 evaluated one PWR sequence (Sequoyah)and one BWR (Peach Bottom) sequence and calculatedin-vessel non-radioactive aerosol masses of 350 and 780kilograms, respectively, for the PWR and BWRsequences. The same reference calculated thatex-vessel aerosol masses (assuming a dry cavity) wouldbe higher, 3800 and 5600 kilograms, respectively, forthe PWR and BWR sequences investigated. However,these values, particularly for the ex-vessel releasephase, may be excessive. NUREG/CR-4624 (Ref. 27)examined several sequences for both PWRs and BWRsand calculated ex-vessel releases to containment ofabout 1000 and 4000 kilograms, respectively, for PWRsand BWRs. NUREG/CR-5942 (Ref.19), making use ofthe MELCOR code, calculated significantly lowerreleases during the ex-vessel phase of about 1000kilograms for the Peach Bottom plant.

In view of the wide diversity of calculated results, theNRC staff concludes that precise estimates of therelease of non-radioactive aerosols are not available atthis time. Because nonradioactive aerosol masses couldhave an effect upon the operation of certain plantequipment, such as filter loadings or sump perfor-mance, during and following an accident, however, theNRC staff concludes that the release of non-radioactiveaerosols should be considered by the designer usingmethods considered applicable for his design, and thepotential impact upon the plant evaluated.

4 MARGINS AND UNCERTAINTIES

This section discusses some of the more significantconservatisms and margins in the proposed accidentsource term given in Section 3. Briefly, the proposedrelease fractions have been developed from a completecore-melt accident, that is, assuming core melt withreactor pressure vessel failure and with the assumptionof core-concrete interactions. The timing aspects wereselected to be typical of a low pressure core-meltscenario, except that the onset of the release of gapactivity was based upon the earliest calculated time offuel rod failure under accident conditions. Themagnitude of the fission products released intocontainment was intended to be representative and,except for the low volatile nuclides, as discussed insection 4.4, was estimated from the mean values for atypical low-pressure core-melt scenario.

4.1 Accident Severity and lypeAs noted earlier in Section 2.2, this report discussesmean or average release fractions for all the releasephases associated with a complete core-melt accident,including reactor pressure vessel failure. The accidentselected is one in which core melt occurs at lowpressure conditions. A low pressure core melt scenarioresults in a relatively low level of fission productretention within the reactor coolant system, and aconsequently high level of release of fission productsfrom the core into containment during the earlyin-vessel release phase. Since the bulk of the fissionproducts entering containment do so during the earlyin-vessel release phase, selection of a low pressure coremelt scenario provides a high estimate of the totalquantity of fission products released into containment,as well as that during the early in-vessel release phase.

4.2 Onset of Fission Product ReleaseThe onset, or earliest time of appearance of fissionproducts within containment, has been selected on thebasis of the earliest time to failure of a fuel rod, given adesign basis LOCA. This is estimated to be from about13 to 25 seconds for plants that do not have leak-before-break approval for their reactor coolant systempiping, and it is expected to vary depending on thereactor as well as the fuel rod design. This value, whilerepresenting some relaxation from the assumption ofinstantaneous appearance, is nevertheless conservative.As noted in Reference 15, these estimates are valid fora double-ended rupture of the largest pipe, assume thatthe fuel rod is being operated at the maximum peakingfactor permitted by the plant Technical Specificationsand at the highest burnup levels anticipated, andassume that the emergency core cooling system (ECCS)is not operating. Use of more realistic assumptions forany of these parameters would increase estimated timesto fuel rod failure by factors of two or more. Neverthe-less, the use of conservative assumptions in estimatingfuel rod failure times is considered appropriate sincesuch failure times are likely to be used primarily inconsideration of the necessary closure time for certaincontainment isolation valves. Since it is important thatclosure of such valves be ensured before the release ofsignificant radioactivity to the environment, a conserva-tive estimate of fuel failure time and consequent onsetof fission product appearance is deemed appropriate.For plants with leak-before-break approval for theirreactor coolant system piping, a longer duration beforefuel clad failure is expected. However, other constraintsmay become the limiting factor on containmentisolation valve closure time.

4.3 Release Phase DurationsThe durations of the various release phases have beenselected primarily by examination of the values

15 NUREG-1465

available for the group of severe accident scenariosconsidered in Section 3. The durations of the earlyin-vessel and ex-vessel release phases differs for BWRsversus PWRs and reflect the differing core heatup ratesas well as the differing amounts of zirconium availableto supply chemical energy after core-melt. While theselected durations of the release phases are realistic,some conservatisms should be noted. The duration ofthe early in-vessel release phase for BWRs and PWRsis short and does not represent a probabilisticallyweighted average or mean value for the accidentsequences considered. This will introduce a givenquantity of fission products into containment in ashorter time than might be expected for a typicalsequence.

Similarly, the duration of the ex-vessel release phase,while considered realistic for the bulk of the fissionproducts being released, is short for releases oftellurium and ruthenium since, as noted in Section 3.3,release of these nuclides occurs over a longer time.

The selected release duration times have been chosenprimarily on the basis of simplicity, since an accuratedetermination of the duration of the release phasesdepends not only on the reactor type but also on theapplicable accident sequence, which varies for eachreactor design.

4.4 Composition and Magnitude ofReleases

The composition of the fission products was initiallybased on the grouping developed with the STCP, buthas been modified as discussed in Section 3.4.

The magnitudes of the fission products released intocontainment for the accident source term were selectedin the draft version of this report to be the meanvalues, using NUREG-1150 methodology, for BWRand PWR low-pressure scenarios involving highestimates of zirconium oxidation. The uncertaintydistributions for the in-vessel release and total releaseinto containment are displayed graphically in Appen-dix A. Bounding estimates for the releases intocontainment taken from Reference 17, using the STCPmethodology, are shown in Appendix B.

The release magnitudes for the low volatile fissionproducts were reduced significantly in the final report.This reduction was based upon recent experimentalresearch results (Ref. 25) since completion ofNUREG-1150, as well as a re-examination of theuncertainty distribution, in response to comments onthe draft report. Research on in-vessel phenomenaincludes the in-pile Severe Fuel Damage (SFD)experiments in the Power Burst Facility (PBF), further

examination of the Three Mile Island (TMI) accident,and the SASCHA out-of-pile tests. Ex-vessel insightsderive primarily from large scale tests performed aspart of the internationally sponsored ACE Program.Reference 25 notes that, based on the SFD experimentsas well as the TMI accident, in-vessel release fractionsfor cerium, for example, were about 104, compared tothe value of 10-2 cited in the draft report. Based onthese results, the NRC staff concludes that the lowvolatile release fractions cited in draft NUREG-1465are too high.

The uncertainty distributions were also examined toobtain additional insight. As can be seen from theuncertainty distributions in Appendix A, the range ofrelease estimates for the volatile nuclides, such as thenoble gases, iodine, cesium, and to some extenttellurium, spans about one order of magnitude. For thisgroup of nuclides, use of the mean value is areasonable estimate of the release fraction. In contrast,the range for the low volatile nuclides, such as barium,strontium, cerium and lanthanum, spans about 4 to 6orders of magnitude. For the latter group of nuclides,the mean value can be misleading, since it may be wellin excess of other measures of the distribution. This isillustrated in TIable 4.1 which tabulates the mean,median, and 75th percentile values for several lowvolatile nuclides released during the early in-vesselphase.

Table 4.1 Measures of Low Volatile In-Vessel ReleaseFractions

Nuclide Mean Median 75th percentile

Sr 0.03 0.001 0.006Ba 0.04 0.003 0.009La 0.002 0.00003 0.0003Ce 0.01 0.00006 0.0006

As can be seen from Thble 4.1, the mean value for thisgroup of nuclides is one to two orders of magnitudegreater than the median value, and is about 5 timesgreater than the 75th percentile of the distribution. Forthis group of nuclides, the mean is controlled by theupper tail of the distribution, and the details of thewhole distribution may be more indicative of theuncertainty than the "bottom line" results, such as amean value. Because of this, the final version of thisreport has chosen not to use the mean value inestimating releases for the non-volatile nuclides. Whilethe median value might be selected as an alternate, itfails to provide an appreciation of the range of valueslying above it. Since this report is intended forregulatory applications, the intent is to avoidunder-estimation of potential releases or offsite doses,without undue conservatism. Hence, for the final

NUREG-1465 16

report, the 75th percentile value has been selected forthe low volatile nuclides on the basis that it boundsmost of the range of values, without undue influence bythe upper tail of the distribution.

Uncertainties, particularly in understanding andmodeling core melt progression phenomena, can affectthe duration of the early in-vessel release phase,including the timing of reactor pressure vessel failure.An increase in duration of the early in-vessel phase canlead to increased releases of volatile fission productsduring the early in-vessel phase and a concomitantreduction during the ex-vessel phase. An increase induration of the early in-vessel phase, however, alsoprovides additional time for fission product removalwithin containment by natural processes or fissionproduct cleanup systems.

Upper bound estimates, tabulated in Appendix B,indicate that virtually all the iodine and cesium couldenter the containment. Similarly, for tellurium, upperbound estimates indicate that as much as abouttwo-thirds of the core inventory of tellurium could bereleased into containment. Hence, for this importantgroup of radionuclides (iodine, cesium, and tellurium),the upper bound estimates of total release intocontainment are approximately 1.5 times the meanvalue estimates.

For the lower volatility radionuclides such as bariumand strontium, upper bound estimates range fromabout 50 to 70% of the core inventory released intocontainment. Almost all of this is estimated to bereleased as a result of core-concrete interactions. Incontrast, mean value estimates range from 15 to 25%.Hence, in this case, the upper bound estimates areabout two to three times the mean values.

Finally, for the refractory nuclides such as lanthanumand cerium, the upper bound estimates indicate thatabout 5% of the inventory of these nuclides couldappear within containment, whereas the mean valueestimate indicates only about 1% released.

PRAs have indicated that, considering the magnitudesof the radioactive species estimated to be released tothe environment for severe reactor accidents, theradionuclides having the greatest impact on risk aretypically the volatile nuclides such as iodine andcesium, with tellurium to a somewhat lesser degree.The uncertainty distributions for this group ofradionuclides is also the smallest, as shown in thegraphical tabulations of Appendix A. Hence, our abilityto predict the behavior and releases for this group ofnuclides is significantly better than for other fissionproduct groupings.

Mean value estimates selected for the in-containmentaccident source term provide reasonable estimates forthe important nuclides consisting of iodine, cesium, andtellurium. These estimates show a relatively low degreeof uncertainty and are unlikely to be exceeded by morethan 50%. Uncertainty increases in estimating releasesfor the remaining nuclides.

4.5 Iodine Chemical FormThe chemical form of iodine entering containment wasinvestigated in Reference 18. On the basis of this work,the NRC staff concludes that iodine enteringcontainment from the reactor coolant system iscomposed of at least 95% cesium iodide (CsI), with nomore than 5% 1 plus HI. Once within containment,highly soluble cesium iodide will readily dissolve inwater pools and plate out on wet surfaces in ionic form.Radiation-induced conversion of the ionic form toelemental iodine will potentially be an importantmechanism. If the pH is controlled to a level of 7 orgreater, such conversion to elemental iodine will beminimal. If the pH is not controlled, however, arelatively large fraction (greater for PWRs than BWRs)of the iodine dissolved in containment pools in ionicform will be converted to elemental iodine.

5 IN-CONTAINMENT REMOVALMECHANISMS

Since radioactive fission products within containmentare in the form of gases and finely divided airborneparticulates (aerosols), the principal mechanism bywhich fission products find their way from the reactorto the environment with an intact containment is vialeakage from the containment atmosphere. The specificfission product inventory present in the containmentatmosphere at any time depends on two factors: (1) thesource, i.e., the rate at which fission products are beingintroduced into the containment atmosphere, and(2) the sink, the rate at which they are being removed.Aspects of the release and transport of fission productsfrom the core into the containment atmosphere werepresented in Section 3.

Mechanisms that remove fission products from theatmosphere with consequent mitigation of thein-containment source term fall into two classes:(1) engineered safety features (ESFs) and (2) naturalprocesses. ESFs to remove or reduce fission productswithin the containment are presently required(Criterion 41 in Appendix A of 10 CFR Part 50) andinclude such systems as containment atmospheresprays, BWR suppression pools, and filtration systemsutilizing both particulate filters and charcoal adsorptionbeds for the removal of iodine, particularly in elemen-tal form. Natural removal includes such processes as

17 NUREG-1465

aerosol deposition and the sorption of vapors onequipment and structural surfaces.

The draft version of this report contained a discussionof some of the more important fission product removalmechanisms, including some quantitative results. Thesenumerical results were intended to be illustrative of thephenomena involved and were not intended to beapplied rigorously, however. It was recognized that thedata and illustrations used in the draft might not beapplicable to all situations.

In recognition of this, the NRC staff undertook toexamine, with contractor assistance, improvedunderstanding of fission product removal mechanisms.At this time, this effort is still underway. Rather thanprovide numerical values that may be inapplicable, thisreport will provide references, where available, so thatthe reader may utilize improved methodologies toobtain results that apply to the situation at hand.

5.1 Containment SpraysContainment sprays, covered in Standard Review Plan(SRP) Section 6.5.2 (Ref. 28), are used in many PWRdesigns to provide post-accident containment cooling aswell as to remove released radioactive aerosols. Spraysare effective in reducing the airborne concentration ofelemental and particulate iodines as well as otherparticulates, such as cesium, but are not effective inremoving noble gases or organic forms of iodine. Thereduction in airborne radioactivity within containmentby a spray system as a function of time is expressed asan exponential reduction process, where the sprayremoval coefficient, lambda, is taken to be constantover a large part of the regime. Typical PWRcontainment spray systems are capable of rapidlyreducing the concentration of airborne activity (byabout 2 orders of magnitude within about 30 minutes,where both spray trains are operable). Once the bulk ofthe activity has been removed, however, the spraybecomes significantly less effective in reducing theremaining fission products. This is usually accountedfor by either employing a spray cut-off, wherein thespray removal becomes zero after some reduction hasbeen achieved, or changing to a much smaller value oflambda to reflect the decreased removal effectivenessof the spray when airborne concentrations are low.

SRP Section 6.5.2 (Ref. 28) provides expressions forcalculating spray lambdas, depending on plantparameters as well as the type of species removed. Inaddition, SRP 6.5.2 currently suggests that thecontainment sump solution be maintained at values ator above pH levels of 7, commencing with sprayrecirculation, to minimize revolatilization of iodine inthe sump water. Current guidance states that

containment spray systems be initiated automatically,because of the instantaneous appearance of the sourceterm within containment, and that the spray durationnot be less than 2 hours. In contrast, the revised sourceterm information given in Section 3 suggests that spraysystem actuation might be somewhat delayed forradiological purposes, but that the spray systemduration should be for a longer period of about 10 ormore hours. Because sprays are effective in rapidlyremoving particulates from the containmentatmosphere, intermittent operation over a prolongedperiod may also provide satisfactory mitigation.

The spray removal coefficient for particulates appearsparticularly important in view of the informationpresented in Section 3, which indicates that most fissionproducts are expected to be in particulate form. Thespray removal coefficient (X) is derived from thefollowing equation from Standard Review PlanSection 6.5.2

x =3hFE2VD

h = Fall height of spray dropsV = Containment building volumeF = Spray flowE/D = the ratio of a dimensionless collectionefficiency E to the average spray drop Diameter D.EJD is conservatively assumed to be equal to10/meter for spray drops 1 mm in diameter changingto 1/meter when the aerosol mass has beendepleted by a factor of 50.

Using values typical for PWRs, the formulation given inSRP 6.5.2 estimates particulate removal rates to be onthe order of 5 per hour. Nourbakhsh (Ref. 29) exa-mined the effectiveness of containment sprays, asevaluated in NUREG-1150 (Ref. 7), in decontamin-ating both in-vessel and ex-vessel releases. Powers andBurson (Ref. 30) have developed a more realistic, yetsimplified, model with regard to evaluating theeffectiveness of aerosol removal by containment sprays

5.2 BWR Suppression PoolsBWRs use pressure suppression pools to condensesteam resulting from a loss-of-coolant accident. Prior tothe release to the reactor building, these pools alsoscrub radioactive fission products that accompany thesteam. Regulatory Guide 13 (Ref. 2) suggests notallowing credit for fission product scrubbing by BWRsuppression pools, but SRP Section 6.5.5 (Ref. 31) wasrevised to suggest allowing such credit. The pool waterwill retain soluble, gaseous, and solid fission productssuch as iodines and cesium but provide no attenuationof the noble gases released from the core. The ReactorSafety Study (WASH-1400, Ref. 5) assumed adecontamination factor (DF) of 100 for subcooled

NUREG-1465 18

suppression pools and 1.0 for steam saturated pools.Since 1975 when WASH-1400 was published, severaldetailed models have been developed for the removalof radioactive aerosols during steam flow throughsuppression pools.

Calculations for a BWR with a Mark I containment(Ref. 27) used in NUREG-1150 (Ref. 7) indicate thatDFs ranged from 1.2 to about 4000 with a median valueof about 80. The suppression pool has been shown to beeffective in scrubbing some of the most importantradionuclides such as iodine, cesium, and tellurium, asthese are released in the early in-vessel phase. TheNRC staff is also presently reviewing fission productscrubbing by suppression pools to develop simplifiedmodels.

If not bypassed, the suppression pool will also beeffective in scrubbing ex-vessel releases. Suppressionpool bypass is an important aspect that places an upperlimit on the overall performance of the suppressionpool in scrubbing fission products. For example, if aslittle as 1% of the fission products bypass thesuppression pool, the effective DF, taking bypass intoaccount, will be less than 100, regardless of the pool'sability to scrub fission products.

Although decontamination factors for the suppressionpool are significant, the potential for iodinere-evolution can be important. Re-evolution of iodinewas judged to be important in accident sequenceswhere the containment had failed and the suppressionpool was boiling. There is presently no requirement forpH control in BWR suppression pools. Hence, it ispossible that suppression pools would scrub substantialamounts of iodine in the early phases of an accident,only to re-evolve it later as elemental iodine. It maywell be that additional materials likely to be in thesuppression pool as a result of a severe accident, suchas cesium borate or cesium hydroxide and core-concretedecomposition products, would counteract anyreduction in pH from radiolysis and would ensure thatthe pH level was sufficiently high to precludere-evolution of elemental iodine. Therefore, if credit isto be given for long-term retention of iodine in thesuppression pool, maintenance of the pH at or above alevel of 7 must be demonstrated. It is important tonote, however, that this is not a matter of concern forpresent plants since all BWRs employ safety-relatedfiltration systems (see Section 5.3) designed to copewith large quantities of elemental iodine. Hence, evenif the suppression pool were to re-evolve significantamounts of elemental iodine, it would be retained bythe existing downstream filtration system.

5.3 Filtration SystemsESF filtration systems are discussed in RegulatoryGuide 1.52 (Ref. 32) and are used to reduce the

radioactive aerosols and iodine released duringpostulated accident conditions.

A typical ESF filtration system consists of redundanttrains that each have demisters to remove steam andwater droplets from the air entering the filter bank,heaters to reduce the relative humidity of the air, highefficiency particulate air (HEPA) filters to removeparticulates, charcoal adsorbers to remove iodine inelemental and organic form, followed finally byadditional HEPA filters to remove any charcoal finesreleased.

Charcoal adsorber beds can be designed, as indicated inRegulatory Guide 1.52, to remove from 90 to 99% ofthe elemental iodine and from 30 to 99% of the organiciodide, depending upon the specific filter train design.

Revised insights on accident source terms, given inSection 3, may have several implications for ESFfiltration systems. Present ESF filtration systems arenot sized to handle the mass loadings of non-radioactive aerosols that might be released as a resultof the ex-vessel release phase, which could producereleases of significant quantities of nonradioactive aswell as radioactive aerosols. However, if ESF filtrationsystems are employed in conjunction with BWRsuppression pools or if significant quantities of waterare overlaying molten core debris (see Section 5.4),large quantities of nonradioactive (as well asradioactive) aerosols will be scrubbed and retained bythese water sources, thereby reducing the aerosol massloads upon the filter system.

A second implication of revised source term insights forESF filtration systems is the impact of revisedunderstanding of the chemical form of iodine withincontainment. Present ESF filtration systems presumethat the chemical form of iodine is primarily elementaliodine, and these systems include charcoal adsorberbeds to trap and retain elemental iodine. Assuming thatpH control is maintained within the containment, a keyquestion is whether charcoal beds are necessary. Twoquestions appear to have a bearing on this issue andmust be addressed, even assuming pH control. Theseare (1) to what degree will Csl retained on particulatefilters decompose to evolve elemental iodine? and (2)what effect would hydrogen bums have on the chemicalform of the iodine within containment? Based onpreliminary information, Csl retained on particulatefilters as an aerosol appears to be chemically stableprovided that it is not exposed to moisture. Exposure tomoisture, however, would lead to CsI decompositionand production of iodine in ionic form (1), which inturn would lead to re-evolution of elemental iodine.Although ESF filtration systems are equipped withdemisters and heaters to remove significant moisturebefore it reaches the charcoal adsorber bed, an

19 NUREG-1465

additional concern is that the demisters themselves maytrap some CsI aerosol.

In conclusion, present ESF filtration systems, whileoptimized to remove iodine, particularly in elementalform, have HEPA filters that are effective in the -

removal of particulates as well. Although such filtrationsystems are not designed to handle the large massloadings expected as a result of ex-vessel releases, whenthey are used in conjunction with large water sourcessuch as BWR suppression pools or significant waterdepths overlaying core debris, the water sources willreduce the aerosol mass loading on the filter systemsignificantly, making such filter systems effective inmitigation of a large spectrum of accident sequences.

5.4 Water Overlying Core DebrisExperimental measurements (Ref. 33) have shown thatsignificant depths of water overlying any molten coredebris after reactor pressure vessel failure will scruband retain particulate fission products. The question ofcoolability of the molten debris as a result of wateroverlying it is still under investigation. A major factorthat may affect the degree of scrubbing is whether thewater layer in contact with the molten debris is boilingor not.

Results from Ref. 33 indicate that both subcooled aswell as boiling water layers having a depth of about3 meters had measured DFs of about 10. A recent study(Ref. 34) performed for the NRC has provided asimplified model to determine the degree of aerosolscrubbing by a water pool overlying core debrisinteracting with concrete.

5.5 Aerosol DepositionSince the principal pathway for transport of fissionproducts is via airborne particulates, i.e., aerosols, thissubject is discussed in some detail. Aerosols are usuallythought of as solid particulates, but in general, the termalso includes finely divided liquid droplets such aswater, i.e., fog. The two major sources of aerosols arecondensation and entrainment. Condensation aerosolsform when a vapor originating from some high-temperature source moves into a cooler region wherethe vapor falls below its saturation temperature andnucleation begins. Entrainment aerosols form when gasbubbles break through a liquid surface and dragdroplets of the liquid phase into the wake of the bubbleas it leaves the surface. In general, condensationparticles are smaller in size (submicron to a fewmicrons), while entrainment particles are usually larger(1.0-100 microns). Once airborne, both types ofaerosols behave in a similar manner with respect toboth natural and engineered removal processes.

There are four natural processes that remove aerosolsfrom the containment atmosphere over a period oftime: (1) gravitational settling, (2) diffusiophoresis,(3) thermophoresis, and (4) particle diffusion. (Particlediffusion is less important than the first three processesand will not be discussed further.) All particles fallnaturally under the force of gravity and collect on anyavailable surface that terminates the fall, e.g., the flooror upper surfaces of equipment. Both diffusiophoresisand thermophoresis cause the deposition of aerosolparticles on all surfaces regardless of their orientation,i.e., walls and ceiling as well as the floor.Diffusiophoresis is the process by which water vapor inthe atmosphere 'drags' aerosol particles with it as itmigrates (diffuses) toward a relatively cold surface onwhich condensation is taking place. Thermophoresisalso causes aerosol particles to move toward anddeposit on colder surfaces but not as a result of massmotion. Rather, the decreasing average velocity of thesurrounding gas molecules tends to drive the particledown the temperature gradient until it traverses theinterface layer and comes into contact with the surfacewhere it sticks.

Aerosol agglomeration is another natural phenomenonthat has an influence on the rates at which the removalprocesses described above will proceed. Agglomerationresults from the random inelastic collisions of particleswith each other. The process brings about a gradualincrease in average particle size resulting in more rapidgravitational settling. Three phenomena contribute toparticle growth by agglomeration: (1) Brownian motion,(2) gravitational fall, and (3) turbulence. Brownianagglomeration is caused by particle collisions resultingfrom random 'buffeting' by high-energy gas molecules.Gravitational agglomeration results from the fact thatsome particles fall faster than others and thereforetend to collide with and stick to other slower fallingparticles on their way down. Finally, rapid variations ingas velocity and flow direction in the atmosphere, Le.,turbulence, tend to increase the rate at which particlecollisions occur and thus increase the average particlesize. It is to be expected that, as agglomerationadvances, the size of the particle will increase, and itsshape can be expected to change as well. These latterfactors have a strong influence on the removalprocesses.

The agglomeration and aerosol removal processes alldepend critically upon the thermodynamic state andthermal-hydraulic conditions of the containmentatmosphere. For example, the condensation onto andevaporation of water from the aerosol particlesthemselves have strong effects on all of theagglomeration and removal processes. Water condensedon aerosol particles increases their mass and makesthem more spherical; both of these effects tend toincrease the rate of gravitational settling. Some

NUREG-1465 20

aerosols, such as CsI and CsOH, are hygroscopic andabsorb water vapor even when the containmentatmosphere is below saturation. As with condensation,hygroscopicity also increases the rate of deposition.

Because of its importance to fields such as weather andatmosphere pollution, the behavior of aerosols hasbeen under study for many decades. A number ofcomputer codes have been developed to specificallyconsider aerosol behavior as it relates to nuclearaccident conditions. The most complete mechanistictreatment of aerosol behavior in the reactorcontainment is found in CONTAIN, a computer codedeveloped at Sandia National Laboratories under NRCsponsorship for the analysis of containment behaviorunder severe accident conditions. The aerosol modelsin the NAUA code are very similar to those used inCONTAIN; NAUA was developed at theKernforschungszentrum, Karlsrhue, F.R.G., and wasused for aerosol treatment in the NRC STCP. Thereare a number of other well-known aerosol behaviorcomputer codes, but these two are the most widely usedand accepted throughout the international nuclearsafety community.

The rate at which gravitational settling occurs dependsupon the degree of agglomeration at any particulartime (i.e., the average particle size) as well as the totalparticle density m (mass per unit volume). Thus, as inmost cases where the decrement of a variable isproportional to the variable itself, one can expect anexponential behavior. The gravitational settling processis quite complex and depends upon a large number ofphysical quantities, e.g., collision shape factor, particlesettling shape factor, gas viscosity, effective settlingheight, density correction factor, normalized Browniancollision coefficient, gravitational acceleration, andparticle material density. The only variable in this listthat is independent of the plant, the accident scenario,and the atmospheric thermal-hydraulic conditions is theconstant of gravitation. It follows that no single DF canbe ascribed to cover the entire range of plant designs,accident scenarios, and source materials. An effort isunder way to establish a set of simplified algorithmsthat can be used to provide a set of specific ranges ofatmosphere conditions. This effort is still underway atthis time.

6. REFERENCES1. U.S. Nuclear Regulatory Commission; "Reactor

Site Criteria," Title 10, Code of FederalRegulations (CFR), Part 100.

2. U.S. Nuclear Regulatory Commission;"Assumptions Used for Evaluating the PotentialRadiological Consequences of a Loss of Coolant

Accident for Boiling Water Reactors," RegulatoryGuide 1.3, Revision 2, June 1974.

3. U.S. Nuclear Regulatory Commission;"Assumptions Used for Evaluating the PotentialRadiological Consequences of a Loss of CoolantAccident for Pressurized Water Reactors,"Regulatory Guide 1.4, Revision 2, June 1974.

4. JJ. DiNunno et al., "Calculation of DistanceFactors for Power and Test Reactor Sites,"Technical Information Document (11D}-14844,U.S. Atomic Energy Commission, 1962.

5. U.S. Nuclear Regulatory Commission; "ReactorSafety Study An Assessment of Accident Risks inU.S. Commercial Nuclear Power Plants,"WASH-1400 (NUREG-75/014), December 1975.

6. J. A. Gieseke et al., "Source Term Code Package:A User's Guide," NUREG/CR-4587 (BMI-2138),prepared for NRC by Battelle Memorial Institute,July 1986.

7. U.S. Nuclear Regulatory Commission; "SevereAccident Risks: An Assessment for Five U.S.Nuclear Power Plants," NUREG-1150, December1990.

8. M.R. Kuhlman, DJ. Lehmicke, and R.O. Meyer,"CORSOR User's Manual," NUREG/CR-4173(BMI-2122), prepared for NRC by BattelleMemorial Laboratory, March 1985.

9. H. Jordan, and M.R. Kuhlman, "TRAP-MELT2User's Manual," NUREG/CR-4205 (BMI-2124),prepared for NRC by Battelle MemorialLaboratory, May 1985.

10. D.A. Powers, J.E. Brockmann, and A.W. Shiver,"VANESNA A Mechanistic Model of RadionuclideRelease and Aerosol Generation During CoreDebris Interactions with Concrete,"NUREG/CR-4308 (SAND 85-1370), prepared forNRC by Sandia National Laboratories, July 1986.

11. P.C. Owczarski, A.K. Postma, and R.I. Schreck,'Technical Bases and User's Manual for thePrototype of SPARC-A Suppression Pool AerosolRemoval Code," NUREG/CR-3317 (PNL-4742),prepared for NRC by Battelle Pacific NorthwestLaboratories, May 1985.

12. W.K. Winegardner, A.K. Postma, and M.W.Jankowski, "Studies of Fission Product Scrubbingwithin Ice Compartments," NUREG/CR-3248(PNL-4691), prepared for NRC by Battelle PacificNorthwest Laboratories, May 1983.

21 NUREG-1465

13. H. Bunz, M. Kayro, and W. Schock, "NAUA-Mod4: A Code for Calculating Aerosol Behavior inLWR Core Melt Accidents," KfK-3554,Kernforschungszentrum Karlsruhe Germany,1983.

14. R.M. Summers, et al., "MELCOR 1.8.0: AComputer Code for Nuclear Reactor SevereAccident Source Term and Risk AssessmentAnalysis," NUREG/CR-5531 (SAND 90-0364),prepared for NRC by Sandia NationalLaboratories, January 1991.

15. K.R. Jones, et al, '"Tming Analysis of PWR FuelPin Failures," NUREG/CR-5787 (EGG-2657),prepared for NRC by Idaho National EngineeringLaboratory, September 1992.

16. H.P. Nourbakhsh, M. Khatib-Rahbar, and R.E.Davis, "Fission Product Release Characteristicsinto Containment Under Design Basis and SevereAccident Conditions," NUREG/CR-4881(BNL-NUREG-52059), prepared for NRC byBrookhaven National Laboratory, March 1988.

17. H.P. Nourbakhsh,: "Estimates of RadionuclideRelease Characteristics into Containment UnderSevere Accidents," NUREG/CR-5747(BNL-NUREG-52289), prepared for NRC byBrookhaven National Laboratory, November 1993.

18. E.C. Beahm, C.F. Weber, and T.S. Kress, "IodineChemical Forms in LWR Severe Accidents",NUREG/CR-5732 (ORNLTM-11861), preparedfor NRC by Oak Ridge National Laboratory, April1992.

19. JJ. Carbajo, "Severe Accident Source TermCharacteristics for Selected Peach BottomSequences Predicted by the MELCOR Code,"NUREG/CR-5942 (ORNLnTM-12229), preparedfor NRC by Oak Ridge National Laboratory,September 1993.

20. DJ. Alpert, D.I. Chanin, and LT. Ritchie,"Relative Importance of Individual Elements toReactor Accident Consequences Assuming EqualRelease Fractions."NUREG/CR-4467, preparedfor NRC by Sandia National Laboratories, 1986.

21. C.F. Weber, E.C. Beahm and T.S. Kress, "Modelsof Iodine Behavior in Reactor Containments,"ORNrITM-12202, Oak Ridge NationalLaboratory, October 1992.

22. E.C. Beahm, R.A. Lorenz, and C.F. Weber,"Iodine Evolution and pH Control,"NUREG/CR-5950, (ORNLErM-12242), prepared

for NRC by Oak Ridge National Laboratory,December 1992.

23. A.K. Postma, and R.W. Zavadowski, "Review ofOrganic Iodide Formation Under Accident

. Conditions in Water-Cooled Reactors,"WASH-1233, U.S. Atomic Energy Commission,October 1972.

24. E.C. Beahm, W.E. Shockley, and O.L. Culberson,"Organic Iodide Formation Following NuclearReactor Accidents," NUREG/CR-4327,(ORNLITM-9627), prepared for NRC by OakRidge National Laboratory, December 1985.

25. D. J. Osetek, "Low Volatile Fission ProductReleases During Severe Reactor Accidents,"DOE/ID-13177-2, prepared for U.S. Departmentof Energy by Los Alamos Technical Associates,October 1992.

26. M. Silberberg et al., "Reassessment of theTechnical Bases for Estimating Source Terms,"NUREG-0956, July 1986.

27. R.S. Denning, et al., "Radionuclide ReleaseCalculations for Selected Severe AccidentScenarios: BWR Mark I Design,"NUREG/CR-4624, Vol. 1, prepared for NRC byBattelle Memorial Institute, July 1986.

28. U.S. Nuclear Regulatory Commission:"Containment Spray as a Fission Product CleanupSystem," Standard Review Plan, Section 6.5.2,Revision 2, NUREG-0800, December 1988.

29. H.P. Nourbakhsh,: "In-Containment RemovalMechanisms," Presentation to NRC staff January3, 1992, Brookhaven National Laboratory, January1992.

30. D.A. Powers and S.B. Burson, "A SimplifiedModel of Aerosol Removal by ContainmentSprays," NUREG/CR-5966, (SAND92-2689),prepared for NRC by Sandia NationalLaboratories, June 1993.

31. U.S. Nuclear Regulatory Commission: "PressureSuppression Pool as a Fission Product CleanupSystem," Standard Review Plan, Section 6.5.5,NUREG-0800, December 1988.

32. U.S. Nuclear Regulatory Commission: "Design,Testing, and Maintenance Criteria for PostaccidentEngineered-Safety-Feature Atmosphere CleanupSystem Air Filtration and Adsorption Units ofLight-Water-Cooled Nuclear Power Plants,"Regulatory Guide 1.52, Revision 2, March 1978.

NUREG-1465 22

33. J. Hakii et al., "Experimental Study on AerosolRemoval Efficiency for Pool Scrubbing UnderHigh Temperature Steam Atmosphere,"Proceedings of the 21st DOE/NRC Nuclear AirCleaning Conference, August 1990.

34. D.A. Powers and J.L Sprung, "A Simplified Modelof Aerosol Scrubbing by a Water Pool OverlyingCore Debris Interacting With Concrete,"NUREG/CR-5901, (SAND92-1422), prepared forNRC by Sandia National Laboratories, November1993.

23 NUREG-1465

APPENDIX AUNCERTAINTY DISTRIBUTIONS

NUREG-1465 24

a)

U)

a)cor.

0to

a)V)co.2

P.0

i:- r1

U =

(b) Low Zirconium Oxidation (High Zr Content in the Melt)Uncertainty Distributions for Total Rdeases Into Containment PWR,.Low RCS Pressure, Lime-stone Concrete, Dry Cavity, Two Openings After VB, FPART = 1.

25 NUREG-1465

en

0)

-Z

0

0)

-jC.)

to

lk:

(b) Low Zirconium Oxidation (High Zr Content in the Melt)Uncertainty DistributIons for Total Releases Into Containment PWR. Low ICS Pressure, BasalticConcrete, Dr Cavity, Two Openings After YB, FPART = 1.

NUREG-1465 26

i0

inU 10-3

-51Clo f.2 10 sth

10~

107

l-B

10

10 -

(a) High Zirconium Oxidation (Low Zr Content in the Melt)

10 --Lc .. r{s1 Th I

10

-Jh

(b) Low Zirconium Oxidation (High Zr Content in the Melt)Uncertainty Distributions fbor Total Releass late Contalatnent BWER, Low Pressure Fast StatlonBlackout, LUmestone Concrete, Dry Pedestal, bou DrywvH Temperature, FPART = 1.

_0-27 NUREG-1465

APPENDIX B

STCP BOUNDING VALUE RELEASES

NUREG-1465 28

- Updated Bounding Value of Radionuclide Releases Into the ContainmentUnder Severe Accident Conditions for PWRs

n NVU

Hiah RCS Low RCSPressure Pressure

ST ePIeh RCSPressurn

urmeCont

NG 1.0 1.0 0. a

1 0.30 0.75 0.10 0.

Cs 0.30 0.75 .0.10 0.

To 0.20 0.50 0.05 0.

Sr-Ba 0.003 0.01 0.01 0.

Ru 0.003 0.01 0.05 o.C

La-Co 5 x105 1.5 x 10'4 0.01 0.

Release 40 minutesDuration

(' All entries are fractions of Initial core Inventory.

(b) Assuming 100% of the core participate In CCI.

(') Except for To and Ru where the duration Is extended to five hours.

STr (b)

stone Icreto e

15

15

40

40

005

05

2 hourste'

3asalticoncrete

0.

0.15

0.15

0.30

0.15

0.005

0.05

STEV

Hlah RCS Low RCSPressure Pressure

0. 0.

0.05 0.02

0.02 0.02

0.02 0.01

10 hours -

C

m

0dI.-

z

I-

Updated Bounding Value of Radionuclide Releases Into the ContainmentUnder Severe Accident Conditions for BWRs

MG

Cs

To

I r8-

Ru

La-Ca

ReleaseDuration

eTv (a

Hbh PCS Low RCSPressure Pressure")

1. 1.

0.50 0.75

0.50 0.75

0.10 0.15

0.003 0.01

0.003 0.01

5x10' 1.5x104

1.5 hours

Hith RCSPressure

0.

0.10

0.10

0.05

0.01

0.05

* 0.01

aTMXV!2Llmestone BasalticConcrete Concrete

0. 0.

015 0.15

0.15 0.15

0.50 0.30

0.70 0.30

0.005 0.005

0.10 0.10

3 hours(d)

High RCS Low RCSPressure Pressure (b

0. 0.

0.10 0.02

0.05 0.01

0.02 0.02

10 hours

"' All entries are fractions of Initial core Inventory.

I' High pressure ATWS wre also considered In this category.

"' Assuming 100% of the core participate In CCI.

I' Except for To and Ru where the duration Is extended to six hours.

NRC FORM 335 U.S. NUCLEAR REGULATORY COMMISSION . REPORT NUMBER12 a9i (Al,~.d by NA C. Add Vo.I. $.g. Rjw.

32C013202. BIBLIOGRAPHIC DATA SHEET umkfe I 'a(SM inS fCtomS on the t.rain

2. TITLE AND SUBTITLE NUREG- 1465

Accident Source Terms for Lischt-Water Nuclear Power Plants3. OATE REPORT PUBLISHED

MONTH Y YEAR

February 19954. FIN OR GRANT NUMBER

S. AUTHOR(S) 6. TYPE OF REPORT

L. Soffer, S. B. Burson, C. M. Ferrell, R. Y. Lee, J. N. Ridgelv 7.PERIOD COVEREO rneNjr"i

U.r s~Antlt r%,ul-Tnm, , N% %/fl inARFCCi'&fl 1%f no nif.Y_ At~h as'hf ax~ I, 'VM. .W U E. U'MPAl'. . MU ( qo yL m O. mbfg ,ujt nrw p v

-E d maia 4ad&oj

Division of Systems TechnologyOffice of Nuclear Regulatory ResearchU.S. Nuclear Regulatory CommissionWashington. DC 20555 -0001

9. SPONSORING ORGANIZATION -NAME AND ADDR ESS (If NRC. rp -Sw'v u *t egonrcMt, eofdoe NRC Ories on. Oftivo potion,. U.S. SIKOCMAeotjrorr Commni'o.,an -wln odk

Same as above

10. SUPPLEMENTARY NOTES

11. ABSTRACT (Ie_ or ir

In 1962 the U.S. Atomic Energy Commission published TID-14844, "Calculation of Distance Factors forPower and Test Reactors" which specified a release of fission products from the core to the reactorcontainment for a postulated accident involving "substantial meltdown of the core". This "source term", thebasis for the NRC's Regulatory Guides 1.3 and 1.4, has been used to determine compliance with the NRC'sreactor site criteria, 10 CFR Part 100, and to evaluate other important plant performance requirements.During the past 30 years substantial additional information on fission product releases has been developedbased on significant severe accident research. This document utilizes this research by providing morerealistic estimates of the "source term" release into containment, in terms of timing, nuclide types,quantities and chemical form, given a severe core-melt accident. This revised "source term" is to be appliedto the design of future light water reactors (LWRs). Current LWR licensees may voluntarily proposeapplications based upon it.

12. KEY WORbSIOES/MPTORS (fL AM"" fet uIil.w 13. AVAILAE)ITY STATEMENT

Unlimited14. SECURIITY CLASSIFICATION

ITM, PWj

Severe Accident Source Term UnclassifiedCore Meltdown 1Th rARpenDesign Basis Accident UnclassifiedTID-14844 Replacement 15. NUMBER OF PAGES

Core Fission Product ReleasesIS. PRICE

NPC FORM 33 m(2"

Federal Recycling Program

accident source terms for light-water nuclear power plants

Documents