ML20072F659

Loss of DHR Sequences at Browns Ferry Unit One - Accident Sequence Analysis

ML20072F659
Person / Time
Site:	Browns Ferry
Issue date:	06/30/1983
From:	Cook D, Greene S, Harrington R, Hodge S OAK RIDGE NATIONAL LABORATORY
To:	NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES)
References
CON-FIN-B-0452, CON-FIN-B-452 NUREG-CR-2973, ORNL-TM-8532, NUDOCS 8306280080
Download: ML20072F659 (180)

Text

-

NUREG /CR-2973 ORNL/TM-8532 g-OAK RI D G E '-

4 NATIONAL..

-LABORATORY Loss of DHR Sequences at i

- O:

Browns Ferry Unit One Accident Sequence Analysis D. H. Cook S. R. Greene R. M. Harrington S. A. Hodge I

p' w.

l l

s 9

Prepared for the U.S. Nuclear Regulatory Commission Office of Nuclear Regulatory Research Under Interagency Agreements DOE 40-551-75 and 40-552-75 r.

15 8306280080 830630 PDR ADOCK 05000259 P

PDR

4 Yl r

Printed in the United States of America. Available fiom National Technical Information Service U.S Department of Commerce 5285 Port Royal Road, Springfield, Virginia 22161 Available from GPO Saks Propam Division of Technical Information and Document Control U S. Nuclear Regu!arory Commission

)

Washington, D C. 20555 TNs report was prepared as an account of work sponsored by an agency o' the

• ~

United States Governerent Neither the U nited S tates Governmen t nor any agency thereof, nor any of their employees, mdkes any warranty. erpress or impi.ed. or assumes any segM hab:htv or respons.biig for the accuracy comp.eteness, or i

)

usefuMess of any inforrrat,on. apparatas p rod s.c t or process disclosed or represents tnat its use wou!d not infringe prvately cwned rights Re'erence ne e.n to any specific commerc alproduct, process ce semce by trade name. trademars.

manufacturer, or otherwise. does not necessardy consttute or imply its endorsement, recommendat:un. or fa ormg by the United States Government or v

any agency thereof The v ews and opm.cns of authors espressed nereen do not necessardy state or ref tect those of theUnitec StatesGovernment or any agency thereof e

b 9

E

1 NUREG/CR-2973 ORNL/TM-8532 Dist. Ca te gory RX, 1S 4

"' y>

Contract No. W-7405-eng-26

)

Engineering Technology Division Instrumentation and Controls Division LOSS OF DHR SEQUENCES AT B2 OWNS FERRY UNIT ONE -

ACCIDENT SEQUENCE ANALYSIS D. H. Cook R. M. Harrington S. R. Greene S. A. Hodge Manuscript Completed - April 14, 1983 Date Published May 1983 C

4 Notice: This document contains information of a preliminary nature.

It is subj ect to revision or correction and therefore doe s not represent a final report.

Prepared for the U.S. Nuclear Regulatory Commission Office of Nuclear Regulatory Research Under Interagency Agreements DOE 40-551-75 and 40-552-75 l

NRC FIN No. B0452 Prepared by the cf',

Oak Ridge National Laboratory Oak Ridge, Tennessee 37830 I "a operated by 1

UNION CARBIDE CORPORATION for the DEPARTMENT OF ENERGY l

111 CONTENTS j

EaLe.

p

SUMMARY

............................................................ vii 1

ABSTRACT...........................................................

1.

INTRODUCTION 1

2.

INITIATING EVENTS 6

2.1 Transients with Unifcrm Pool Heatup 6

2.2 Transients with Pool Thermal Stratification 7

7 2.3 Stuck-Open Relief Valve 8

2.4 LOCA......................................................

3.

TRANSIENTS WITH LOSS OF DHR: CALCULATIONS ASSUMING UNIFORM 9

POOL TEMPERATURE...............................................

9 3.1 Introduction 10 3.2 Summary and Conclusions 10 3.3 Detailed Results..........................................

.f--

o c qr 3.3.1 Reactor vessel water level 11 12 3.3.2 Reactor vessel pressure 3.3.3 Suppression pool temperature and water level 13 3.3.4 Primary containment atmosphere pressure and 14 t empe ra ture 3.3.5 ECCS pump net positive suction head (NPSH) 16 3.3.6 Reactor building environmental considerations 17 4.

NORMAL RECOVERY FROM AND MITIGATION OF LOSS OF DHR FUNCTION 33 33 4.1 Summa ry 4.2 Minimum Number of RHR Coolers Necessary for Pool 34 Cooling 35 4.3 Mitigation Measures 4.3.1 Primary Containment Venting 35 36 4.3.2 Alternate Pool Cooling 5.

TRANSIENTS WITH LOSS OF DHR: CALCULATIONS PERFORMED FOR THE CASE OF THERMAL STRATIFICATION Ih THE SUPPRESSION POOL 42 42 5.1 Introduction..............................................

42 5.2 Summa ry 1

42 1

5.3 Detailed Results 43 5.3.1 Input and assumptions

. 3.2 Transient pool temperature distribution 43 5.3.3 Effect of pool temperature distribution on primary containment pressure build-up Ad

iv

.P_g.12 6.

LOSS OF DHR WITH STUCK-OPEN RELIEF VALVE 49 49 A

Introduction 6.1 gg{

49 6.2 Summary and Conclusions 6.3 Detailed Results 50 t

6.3.1 Reactor vessel water level 50 6.3.2 Reactor ves sel pres sure 51 6.3.3 Suppression pool temperature and water level 51 6.3.4 Primary containment pressure and tasperature 52 6.3.5 ECCS pump net positive suction head (NPSH) 52 7.

STATIC OVERPRESSURIZATION CONTAINMENT FAILURE MECHANISMS AND PHENOMENA 62 7.1 Introduction 62 7.2 BFNP Primary Containment Structural Design 62 7.3 Drywell Line r Gap Construction.............................

63 7.4 Sta tic Overpressure Containment Failure Mechanisms 64 66 7.5 Pos,t Containment Failure Phenomena 7.6 Long-Term Core Cooling Requirements and Syst em 68 Capabilities 7.7 Impact of Post Containment Failure Phenomena on

't 70 g

Inj ection Availability

~

7.8 Loss of DHR Accident Containment Failure 71 Phenomena - Summary 7.9 Loss of DHR Post Containment Failure Event 72 Sequence - Implications 8.

ANALYSIS OF CONTAINMENT FAILURE AND POST CONTAINMENT FAILURE LOSS OF DHR EVENT SEQUENCE 85 8.1 Introduction 85 8.2 Loss of DHR with Loss of Inj ection Following Containment 85 Failure...................................................

9.

IMPLICATIONS OF RESULTS

.......................................101 9.1 Ins trumenta ti on

..........................................101 9.2 Operator Preparedness

....................................102 9.3

System Design

............................................103 9.4 Reconsideration of IREP Study Conclusions

................105 9.4.1 Operation of the CRD hydraulic system pump........ 10 6 9.4.2 Effect of containment backpressure

................107 9.4.3 Content of condensate storage tank

................108 9.4.4 RHR sy s t em minimum fl ow bypa s s val ve s

............. 10 9

y

_Page a

9.4.5 Use of the SBCS system to control primary con-tainment pressure

................................. 109 9.4.6 Requirements for the EECW system 109 9.4.7 Other considerations

............................... 110 9.4.8 Recommendations

................................... 111 10.

CONCLUSIONS

.................................................. 115 10.1 Itemized Conclusions

................................... 115 10.2 Uncertainties in the Analysis

.......................... 117 Appendix A.

DESCRIPTION OF 'IHE BROWNS FERRY UNIT ONE RESIDUAL HEAT REMOVAL SYSTEM

.................................. 125 A.1 Introduction

.................................... 125 A.2 LPCI Operational Mode 125 A.3 Primary Containment Cooling Operational Mode 126 A.4 Reactor Vessel Shutdown Cooling Mode 127 A.5 Special RHR System Features 128 Appendix B.

MODIFICATIONS TO BWR-LACP FOR THIS STUDY 133 B.1 RHR Heat Exchangers

.............................. 133 B.2 Pump Net Positive Suction Head (NPSH) 134 hp B.3 Drywell Coolers................................. 135 B.4 Torus Room Temperatures.........................

136 B.5 Containment Leakage and Containment Vent Flow

............................................ 137 B.6 Evaporation from Surface of Suppression Pool 139 B.7 Flow Rate of Vessel Water Inj ection by the CRD Hydraulic System............................ 140 Appendix C.

MODELING FOR LOCALIZED HEATING OF PRESSURE SUPPRES-l SION POOLS

........................................... 143 C.1 Introduction

.................................... 143 C.2 Phenomenology................................... 143 i

C.3 Pressure Suppression Pool Model 144 C.4 Plume Transport Analysis........................ 146 Appendix D.

MARCH CODE INPUT 157 Appendix E.

ACRONYMS AND SYMBOLS................................. 165 e

i I

.-,_._.m_

m_

l vil

SUMMARY

This study describes the predicted response of Unit I at the Browns 7,

Ferry Nuclear Plant to an extended post-shutdown loss of decay heat re-moval (DHR) capability. The postulated loss of DHR involves the prolonged i

loss of the power conversion system (PCS) and both the pressure suppres-sion pool cooling and the reactor vessel shutdown cooling operational modes of the residual heat removal (RHR) system. With the decay heat re-moval capabilities of the PCS and the RHR system unavailable, the reactor decay heat energy would be concentrated in the pressure suppression pool.

The loss of DHR accident sequences have been selected for the Severe Accident Sequence Analysis (SASA) study presented in this report because they constitute six of the eight dominant accident sequences leading to core melt which have been identified for Browns Ferry Unit One by the NRC's Interim Reliability Evaluation Program (IREP). The IREP study is a probabilistic risk assessment (PRA) whose function is to attempt to con-sider all possible accident sequences at a nuclear plant using event tree and f ault tree methodology for the purpose of identifying the more prob-able, or dominant, sequences. The SASA approach, on the other hand, is to examine a particular category of accident sequences in far greater depth than would be possible in a PRA study.

The purpose of the SASA studies is to pre-determine the probable course of the identified dominant severe accidents so as to establish the timing and the sequence of events for use in the unlikely case that one of these accidents might actually occur.

The SASA studies also produce rec-ommendations concerning the implementation of better system design and U3,

better emergency operating instructions and operator training.

In the interest of efficiency, it is desirable that the SASA effort be directed toward the dominant accident sequences identified by the IREP or other PRA studies as in the case of the Loss of DHR accident sequences at Browns Ferry Unit One.

The basic initiating events for a Loss of DHR sequence include a re-actor scram, closure of the main steam isolation valves (MSIVs) so that the main condenser cannot function as a heat sink, and subsequently, fail-urc of the RHR system to provide either suppression pool cooling or re-actor vessel shutdown cooling.

The steam produced by decay heat is re-lieved from the reactor vessel by the safety / relief valves (SRVs) and is condensed in the pressure suppression pool. The suppression pool tempera-ture increases monotonically and the resulting increase of pressure in the primary containment ultimately threatens containment integrity.

Reactor vessel water level can be maintained during the early stages of a loss of DHR accident sequence by operation o.

either the high pres-sure coolant injection (HPCI) or reactor core isolation cooling (RCIC) pumps. The control rod drive (CRD) hydraulic system pump injects water into the reactor vessel at a rate of 0.0038 m8/s (60 gpm) if the scram is reset and 0.011 m /s (170 gpm) when a scram signal is in ef fect. All 8

three pumps take suction on the condensate storage tank, and operating procedures provide that there would be an initial supply of water in the tank sufficient to last well beyond the time of containment failure in a loss of DHR accident sequence.

I

viii The BWR-LACP code developed at Oak Ridge National Laboratory for BWR analysis has been used for the analysis of the sequence of events before core uncovery. The assumed containment f ailure pressure has been taken from a recent study conducted at the Ames Laboratory which predicts fail-d ure of the Browns Ferry steel containment by static overpressurization at g*

0.910 MPa (117 psig) and that the failure would occur at the juncture of the cylindrical and spherical geometries in the drywell.

The rate of pressure increase in the primary containment during a loss of DHR sequence depends to some extent on the nature of the initiat-ing event.

If the scram is caused by a transient event and at least one pump and basic piping loop of the RHR system is available for circulation and mixing of the suppression pool water, then the suppression pool can be treated as a well-mixed volume of water undergoing a uniform pool heatup.

An example fitting this case would be a loss of offsite power combined with a f ailure of the RHR service water (RHRSW) system; the RHR system would remain available for circulation of the suppression pool water but there would be no cooling flow to the secondary side of the RHR heat ex-changers. The discharge of each RHR loop enters the pool through an elbow which is aligned so that the ef fluent flows axially in the torus to pro-mote mixing, and experiments have shown that the operation of one RHR pump will effectively eliminate thermal stratification in the pressure suppres-sion pool.

For the case of a loss of DHR accident sequence with RHR pump opera-tion and uniform heatup of the pressure suppression pool, the containment pressure reaches the assumed static overpressurization failure point of a

0.910 MPa (117 psig) af ter 35 h.

In the interim, events at several impor-tant milestones determine the temporal plant response, g,

The drywell pressure is 0.108 MPa (1.1 psig) at the inception of the accident. Afte 1 h of suppression pool heatup with cooling unavailable, the drywell pressure reaches 0.115 MPa (2 psig). This is a scram setpoint and also causes the diesel generators, the standby gas treatment system, the high pressure coolant inj ection (HPCI) system, and the RHRSW pumps assigned to the emergency equipment cooling water (EECW) system to start.*

Also, the valves included in groups two, six, and eight of the primary containment isolation system (PCIS) are automatically shut to isolate the drywell and torus.

Even though all control rods would have been inserted at the incep-tion of the accident, the scram signal generated by high drywell pressure at the 1 h point is particularly important to the course of the loss of DHR accident sequence. This is because the control rod drive (CRD) hy-draulic system injection into the reactor vessel increases from 0.004 m /s 8

(60 gpm) to 0.011 m /s (170 gpm) when the scram inlet valves are opened 8

pursuant to a scram signal.

Since the drywell pressure remains above 0.115 MPa (2 psig) throughout the loss of DHR sequence after 1 h, the operator cannot reset the scram signal during this period and the inj ec-i tion to the vessel would remain at the higher rate.

The CRD hydraulic

• It should be noted that all of these events with the exception of HPCI system actuation would occur at the inception of the accident se-quence if the initiating event were a loss of offsite power.

in system pump takes suction on the condensate storage tank and thus the flow does not depend on the status of the pressure suppression pool.*

The operator would control reactor vessel level with the RCIC system during the initial stages of the loss of DHR accident sequence so the larger capacity HPCI system would not be needed and the HPCI turbine would be manually tripped shortly af ter its automatic initiation on high drywell pressure.

The emergency operating instructions require the operator to begin reactor vessel depressurization when the pressure suppression pool ten-perature reaches 498C (120*F) and this also occurs at the 1 h point. The depressurization proceeds at a rate corresponding to a 55.5'C/h (100*F/h) cooldown of the reactor vessel and is completed at the 3.5 h point.

The r e-after, the operator maintains reactor vessel pressure at about 0.689 MPa j

(85 psig) which is suf ficient to run the RCIC turbine when necessary.t l

After the 4 h point, the reactor decay heat has decreased suffi-ciently so that all required reactor vessel makeup injection is supplied by the CRD hydraulic system pump and all other vessel inj ection is termi-nated. The reactor vessel water level increases slowly over the next several hours until at the 8.6 h point, the operator must begin to throt-t1e the CRD hydraulic pump discharge to prevent overfill of the reactor vessel.

Although injection by the RCIC system pump is not required af ter the 4 h point, this system would remain available for a significant period of time thereaf ter until it was isolated on high temperature [366.5 K (200*F)] in the torus room at about the 13 h point. The RCIC turbine high exhaust pressure trip setpoint of 0.276 MPa (25 psig) in the wetwell would l

be reached soon thereafter, at about the 14 h point.

The Icw pressure ECC systems (RHR and core spray) would remain available thereaf ter for l

injection to the reactor vessel from the condensate storage tank as long as the reactor vessel remains depressurized.t The primary containment design pressure of 0.487 MPa (56 psig) would be exceeded at the 21.5 h point. At the 24 h point, the pressure in the drywell would exceed 0.550 MPa (65 psig) and the SRVs could no longer be remote-manually operated as necessary to keep the reactor vessel depres-surized.

The reactor vessel would therefore repressurize, reaching the l

j

• If offsite power is not available, the spare CRD hydraulic pump can be operated with power from a diesel generator.

tThe pressure suppression pool temperature exceeds the maximum design l

lube oil cooler inlet temperature [60'C (140*F)] for the RCIC (and HPCI) system at the 1.6 h point.

Since the lube oil is cooled by the water be-ing pumped, RCIC pump suction should be kept in its normal alignment, i.e.,

to the condensate storage tank.

It should be noted that operation t

of the HPCI system becomes questionable af ter the 2-1/2 h point, when the indicated suppression pool level exceeds +7 in. and the suction of the HPCI booster pump is automatically (and irreversibly) shifted to the heated pressure suppression pool.

e tOperator action would be required to realign the suction of these systems from the pressure suppression pool to the, condensate storage tank.

-,.,,e.e---,,-r,--my---,

r- - - - - -- - - - - - - -. - -

-,,w

,---s---,--~_,--,,_,,-,-.-,,,,,~-,,-v,-

e,-

.-,y-,,,,,,,,,-o,,,em ve ~,,,

a--

.,---m-e,,--v

.y

i x

i setpoint [7.72 MPa (1105 psig)] for automatic actuation of the lowest-set SEY at about the 28 h point.'

The pressure in the primary containment would reach the assumed f ail-are pressure'of 0.910 MPa (117 psig) at the 35 h point.

The reactor ve s-sol would have been pressurized during the seven hour period immediately

,I.

preceeding containment f ailure with the pressure controlled by automatic actuation of the lowest-set SRV and the water level enintained by opera-tion of the CRD hydraulic system pump.

At the time of containment fail-ure, the temperature of the pressure suppression pool is 446 K (343*F) and the temperature in the drywell abnosphere is 500 K (440'F).t The sequence of events af ter containment f ailure is uncertain.

Th e physical integrity of the primary system might be lost because of a vio-lent displacement of the drywell during blowdown.

The capability for suf-l ficient reactor vessel injection to keep the core covered might be lost l

because of the harsh environmental conditions in the reactor building com-bined with an inability to depressurize the reactor vessel so that the low-i l

pressure inj ection systems located outside of the reactor building could be used.

Thus the possibilities range from a large-break LOCA with loss of injection to continued adequate core cooling and consequently, no severe accident.

For this study, it has been assumed that the integrity of the primary system is maintained but all reactor vessel inj ection cap-ability is lost.

This is the approach adopted by the Reactor Safety Study (WASH-1400) and subsequent PRAs.

The MARCH code has been used for the analysis of the depressurization of the primary containment and the subsequent events.

The MARCH computa-tions were initiated just before the primary containment pressure reached the f ailure level, with initial conditions provided by the results of the j

BWR-LACP code at the 34 h point.

Based on recent analytics1 work at the AMES laboratory, the primary containment is assumed to f ail in the dry-l well, at the j uncture of the cylindrical and spherical portions of the 2 (10 ft2).

liner with a failure area of 0.929 m The MARCH results predict primary containment failure at the 351/4 h point and all water injection to the reactor vessel is assumed to cease at this time.

As previously discussed, the primary system is assumed to maintain its integrity during and af ter the primary containment blowdown, and a pressurized boilof f of the water in the reactor vessel at the time of containment f ailure follows. Because of the large inventory of water in the reactor vessel that must be boiled away through the relief valves and the low level of decay heat this long af ter shutdown, core uncovery does not occur until about 21/2 h af ter the loss of injection.

The onset

• No coolant is lost from the reactor vessel during the repressuriza-l tion and the level swell caused by heating of the water would cause the operator to keep the CRD hydraulic pump of f during most of the repressuri-zation.

(The mass of water in the vessel remains constant but the density decreases.)

TThe drywell coolers are lost early in the accident sequence as a result of automatic load shedding when the core spray actuation signal of a combination high drywell pressure-low reactor vessel pressure occurs.

Drywell heating is accelerated during the latter part of the accident se-quence when the reactor vessel has repressurized.

zi of fuel melting occurs about I h later, or 38-3/4 h af ter the inception of the accident.

3 The results of this study illustrate the characteristically slow nature of the loss of DHR accident sequence and the very long time avail-able for the operator to take corrective action.

One purpose of this work has been to determine if additional informa-tion and calculations might af fect the conclusion of the IREP study that Loss of DHR accident sequences constitute a major portion of the total risk of core melt at Browns Ferry Unit 1.

This assignment is a natural and intended function of the SASA program, since this task involves a de-tailed consideration of a specific set of accident sequences.

The PRA done under the auspices of the IREP program identifies the Loss of DHR sequences as dominant as a result of an attempted considera-tion of all possible accident sequences at Browns Ferry Unit 1.

With such a broad scope of study, available RHR system cross-ties between units were neglected and several other simplifying assumptions were necessarily made. These include:

1.

Reactor vessel injection by the CRD hydraulic system was ne-

glected, 2.

The containment was assumed to remain at atmospheric pressure during the heatup of the pressure suppression pool, 3.

The ample source of cool water (not af fected by pressure suppres-sion pool heatup) available to the ECCS systems from the condensate stor-age tank was neglected,*

4.

It was assumed that the RHR system function totally fails if the minimum flow bypass valves provided for pump protection do not close, and 5.

The analysis does not include consideration of the use of the standby coolant supply system, which can be used if necessary in a loss of DHR accident sequence to periodically inj ect river water into the reactor vessel directly or into the drywell or wetwell spray headers as a means to reduce the pressure in the primary containment and thereby avoid contain-ment failure.

Since removal of water from the pressure suppression pool can be accomplished in several ways, especially if the wetwell is pres-surized, river water spray would be an ef fective long-term heat removal mechanism to substitute for the normal decay heat removal functions.

With the simplifying assumptions employed in the IREP study, all re-actor vessel water inj ection capability is lost when the pressure suppres-sion pool water temperature reaches 355 K (180*F), about 5 h af ter the inception of the loss of DHR accident sequence, and core uncovery occurs shortly thereafter. However, the sequence of events determined by the more detailed analysis presented in this report shows that the reactor vessel water injection capability can be maintained at least until the containment fails by overpressurization, more than 24 h af ter the incep-tion of the accident sequence. This allows much more time for corrective action by the operators.

Thus the IREP study treatment of assumptions (1) l

• The IREP study did not recognize that the RHR system and the core spray system pumps can take suction on the condensate storage tank, or e

that the condensate storage tank normally holds enough water to maintain the core covered beyond the point of containment f ailure by over pressur-ization.

,~

- - - - ~ ~-

,,,,----w

,,,y

..-,,-,, - ~ - -. -,,

,m,,,--,,.-

-..-.riv,.,.

-n

-.-.-,,.,,,,-,,--,,,---y,

xii through (5) above might have caused the loss of DHR accident sequences to unrealistically appear to constitute the majority of the dominant core melt sequences.

Accordingly, it is recommended that the order of dominant sequences t

established by the IREP study be reconsidered because it is probable that this will lead to a significant reduction in the core melt frequency as-signed to the loss of DHR sequences. For example, a probability should be assigned as to whether or not the CRD hydraulic system is available during the accident sequence rather than assuming that it is not available, which is tantamount to assigning a 100% f ailure probability to this important system.

4 9

9 I

l t

LOSS OF DHR SEQUENCES AT BROWNS FERRY UNIT ONE -

ACCIDENT SEQUENCE ANALYSES D. H. Cook R. M. Harrington S. R. Greene S. A. Rodge ABSTRACT This study describes the predicted response of Unit One at the Browns Ferry Nuclear Plant to a postulated loss of de-cay heat removal (DHR) capability following scram from full power with the power conversion system unavailable.

In acci-dent sequences without DER capability, the residual heat re-moval (RHR) system functions of pressure suppression pool cooling and reactor vessel shutdown cooling are unavailable.

Conse quently, all decay heat energy is stored in the pressure suppression pool with a concomitant increase in pool tempera-ture and primary containment pressure.

With the assumption that DHR capability is not regained during the lengthy course of this accident sequence, the containment ultimately fails by overpressurization. Although unlikely, this catastrophic f ailure might lead to loss of the ability to inj ect cooling water into the reactor vessel, causing subsequent core un-covery and meltdown.

The timing of these events and the ef-fective mitigating actions that might be taken by the opera-tor are discussed in this report.

1.

INTRODUCTION This is the third report in a series of accident studies concerning

.the BWR 4 - MK I containment plant design.* These studies have been con-ducted at Oak Ridge National Laboratory with the full cooperation of the Tennessee Valley Authority (TVA), using Unit 1 at the Browns Ferry No-clear Plant as the model design. These studies have been done under the auspices of the Severe Accident Sequence Analysis (SASA) program, spon -

sored by the Containment Systems Research Branch of the Division of Acci-dent Evaluation within the Nuclear Regulatory Research arm of the Nuclear Regulatory Commission. The purpose is to pre-determine the probable course of each of a series of severe accidents so as to establish the tim-l ing and the sequence of events; this information would be of use in the unlikely case that one of these accidents might actually occur.

These studies also produce recommendations concerning the implementation of bet-ter system design and better meergency operating instructions and operator training to further decrease the probability of such an event.

l l

• Previous reports concern Station Blackout (NUREG/ CR-2182) and Scram Discharga Volume Break (NUREG/CR-2672).

-.--.. r-~r

,,n,.,,~..-.._,.-,__m

.,.---,.y

,,--_-_..~.-_.r..-.-e

2 The Browns Ferry Nuclear Plant is located on the Tennessee River between Athens and Decatur, Alabama.

Each unit of this three-unit plant comprises a Boiling Water Reactor (BWR) steam supply system designed by 6

the General Electric Company with a maximum power authorized by the op-erating license of 3293 MW(t) or 1067 net MW(e). The General Electric Company and the TVA performed the contruction. Unit 1 began commercial i

operation in August 1974, Unit 2 in March 1975, and Unit 3 in March 1977. The primary containments are of the Mark I pressure suppression pool type and the three units share a secondary containment of, the con-trolled leakage, elevated release design.

Each unit occupies a separate reactor building located in one structure underneath the common refueling floor.This report presents an analysis of the sequence of events during a prolonged loss of decay heat removal (DHR) capability following a scram l

at Unit 1 of the Browns Ferry Nuclear Plant. This accident ca tegory was selected for analysis because it is included in six of the eight dominant accident sequences identified for Browns Ferry Unit 1 by the Interim Reliability Evaluation Program (IREP).1.1 The postulated loss of DHR in-volves the loss of the power conversion system

• and both the pressure suppression pool cooling and the reactor vessel shutdown cooling modes of I

the residual heat removal (RHR) system. With the RRR decay heat removal capability unavailable, the reactor decay heat energy would be concen-trated in the pressure suppression pool. The pressure suppression pool response depends to some extent on the manner in which the decay heat j

energy is introduced; Chap. 2 provides a discussion of the general classi-fication of initiating events.

Loss of DBR accident sequences have been previously considered in Probabilistic Risk Assessment (PRA) studies such as the Reactor Safety Study (WASH-1400). These studies.have treated the pressure suppression pool as a well-mixed volume of water. There is some justification for this approach, since operation of the RHR system pumps (even without the heat exchanger function to remove heat from the flow) would provide good -

l pressure suppression pool mixing during the general pool heatup. The response of Browns Ferry Unit 1 af ter a scram with loss of DHR function and uniform pool heatup is presented in Chap. 3 of this report.

Given that the normal modes of decay heat transfer to the plant cool-ing water systems are not available, there is still the opportunity for the operator to use ingenious methods to remove decay heat from the over-all plant.

Methods for mitigation and normal recovery from the loss of DHR function are discussed in Chap. 4.

If not even one pump and basic piping loop of the RHR system is available to induce suppression pool mixing, then the effect of thermal stratification in the pool water will cause containment f ailure by over-pressurization earlier than would be predicted using the assumption of uniform pool.heatup. The results of analyses of containment response without the assumption of a well-mixed. pool are discussed in Chap. 5.

Two of the IREP-identified Browns Ferry dominant sequences involving loss of DHR capability include a stuck-open relief valve in the initiating

• Loss of the Power Conversion System means that decay heat cannot be removed via the main condensers.

y

, 3 --

7

=- -~ w

'r*q

,,e

/;p

'OEdl,}[

k; i j

.t f;,?

N j f '

3.

l h,(

b event W With a sttek-opfu,scelief$, valve, all of the decay e a ts en'ergy from the reactor vesse,1* isfrptsmitted into the suppression poolca,t,one'loca-tion, and the reactec'tessel' is depressurized a t the time whtr.' the con-discussed in Chap.pc' dent sIquences with a stuck-open relief valve. are tainment falls.#. #

i i

f.

', (.j d

i As shown by the dets ei SASA program analysis p, ovide (d,in.this work, such a long' time is rquiref;for pressure'suppressicy potd heatup to the point where contaissent istlure vould occur by'cyerN@rurir.ation and there is ad much opportunity f 6r. equipment repair and so many-mitigating actio'as av'allable to the oper3 tilts staff that it Is doubtis1 th a t l o s s-of-DHR aE.hidedt sequences should. be eligible for inclusinfin the category of "desinant se,quences" leading to core uncovery,and[rsel'ti'ng at Bh0ME 1 con-tailament plants. Neve r,th'el e s s, this study includes consideration of the pos/Ible /Scver0 Accident l phases of a Loss of DHR accidents.- A Severe Acci-dent'by (efinition,Is arg c+1 dent that in jhe absence,of 'fi.e'ctive correc-e tive action by the operating staff proceeds through core nacovery, core meltdcy and the releassrof fission products from ths' fuel. The events in'the Severe Accident /st'ases of a prolonged Loss-of-DHR event have been a

9'

[

analyzed by# application of thegMUtrH code and are described in Chapters 7 and 8 t

o

~

1 The implica[iIns of t$e resuits of this study are disegssed in

/

(

(hap.

9.'

The disegssipa includes an evaluation of the available instra-s 4

f j

sentation, the level or op9eator' training, the existing emergency proce-dures, and,the loveralt.systen de~ sign from the star.dpoint of req'airements j

,I for mitigatlon,of thigraccider.t. The final portion of Chapter 9 provides-a disection Of the reed fo</ rei.nnsideration of the IREP study findings in ligh6f1he reatitg of this work.//

is

,. j i 6 The conclusions of this gtud:/and a. brief discussioA ot the uncer-e tainties involired uso discussed in Chag,10.

A sloples schemaIic diascan of the reactor vesseYinjection systems considered ~ in this study is provided in Fig. 1.1. With the exception of thgcong1roddrive-(CRD)hydranlic system, all n,f thspe inj ection sys-tems can taken suction on either the conJensate storage tank or the pres-sure suppression thatlegiEedI o' pool and have inj ection capabilities much larger than t replace' the water boiled to steam and" lost f rom the ves -

aef thrqugh the SRVs after a scram. The CRD hydraulic pump. injects con-f dansate storage tank water into the reactor vessel, at: s, rate #of 0.0038 l

m /s (60*gpe) uqJer normal, operating conditions. This' flow increases when 8

l'.

s' scram'is'in effect,1.8 to about 0.0070,m8/a (100 gpm) while the reactor i-vessel is pressurized and 0.011 m /s (170'gpa) when the vessel is depres-s i

surized; d I,

/.

An understanding of the RHR system is important to the consideration of the general category of loss of DHR accidents and the necessary infor-jaation cc{ncerpf a's this important system is provided in Appendip A.

7 Appendix @ contains a description of the additions to. the. computer program EWR-IACP made for this study;,this is the code develope ~d by R. M.

Harrinaton at 40RNL to model' operator actions and the associated primary I

c.

/

k 3

4 i

8These-are (1)' anticipated transient with loss of the power conver-

',, ion system ind' Q) loss of offsite power.

s 4

a s

~~

/

a-f'

.gf.t f

s 5

u' a7 l

J

+

!y

+

L '

5 q?

4 system and reactor building response during the period prior to core un-covery in accident sequences at Browns Ferry.

Appendix C provides a discussion of the computer code developed by D. H. Cook at ORNL as a dissertation project to provide a realistic model of suppression pool heatup in a BWR Mark I containment system, with con-sideration of thermal stratification and localized pool heating.

The MARCH code input for the Severe Accident phases of this study is provided in Appendix D.

A listing of acronyms and symbols used in the report is provided with definitions in Appendix E.

The primary sources of plant-specific information used in the prepa-ration of this report were the Browns Ferry Nuclear Plant (BFNP) Final Safety Analysis Report (FSAR), the USNRC BWR Systems Manual, the BFNP Hot License Training Program Operator Training Manuals, the BFNP Unit 1 Tech-nical Specifications, the BFNP Emergency Operating Instructions, and vari-ons other specific drawings, documents, and manuals obtained from the Tennessee Valley Authority. The experience gained from two plant visits j-in connection with previous studies was also applied in this ef fort.

The setpoints for automatic equipment response used in this study are

^

the currently established safety limits.

In many cases these differ slightly from the actual setpoints used for instrument adjustment at the BFNP because the instrument adjustment setpoints are established so as to provide margin for known instrument error.

This study could not have been conducted on a realistic basis without the current plant status and the extensive background information provided by the Tennessee Valley Authority. The assistance and cooperation of TVA personnel at the Browns Ferry Nuclear Plant, at the Training Simulator, and at the Engineering Support Offices la Chattanooga and Knoxville are gratefully acknowledged.

l References for Section 1 t

1.1 S. E. Mays et al., " Interim Reliability Evaluation Program: Analysis of the Browns Ferry, Unit 1, Nuclear Plant," NURFB/CR-2802, EGG-2199, July 1982.

1.2 S. A. Hodge et al., "SBLOCA Outside Containment at Browns Ferry Unit One - Accident Sequence Analysis," NUREG/CR-2672, Volume 1, ORNL/TM-8119/V1 (November 1982), Sect. E.3.

l

[

r---.-

__ ~

5 ORPL-OwG 83-8481 e

NORMAL SOURCC 1

1 FECDWATER LINE E-l CORE v

E PECIRCUL ATION Qg Y

CRO MECHANISM ASSEM8U ES NORMAL - 60 gpm O

pgqqqqqjSCRAM - 170 gn CR0 ng, HYDRAUUC (4 PUMPS)

W 40,000 gpm p

RCIC

~~~~'*******,

CORE SPRAY 600 gpm CONDENSATE (4 P M )

STORAGE 12.S00 gpm TANK l

l

]/s\\,

i s.

~.,

A j

qE S 9 O OL URE SUPPRESS \\OH Fig,. 1.1.

Simplified diagram of reactor vessel injection systems.

4 l

l

~~

e

~_...

I t

6 l

2.

INITIATING EVENTS

}

Loss of the decay heat removal (DHR) function means that the residual heat removal (RHR) functions of pressure suppression pool cooling and reactor vessel shutdown cooling are unavailable during an accident so-quence 'in which the power conversion system (PCS) is also not available.

Thus decay heat cannot be removed to the RHR service water system via the heat exchangers in the RHR system nor to the main circulating water system via the main condensers.

Under these circumstances, all decay energy is passed from the reactor vessel through the safety / relief valves (SRVs) to the pressure suppression pool.

Since a large amount of energy is absorbed in the pressure suppres-sion pool over a long period of time, the pressure suppression pool ten-perature steadily increases. As the temperature of the pool upper layer increases above 373 K (212'F), the pressure of the primary containment drywell-wetwell combination begins to increase significantly and, unless i

successful operator action is taken to restore the DBR function or to vent I

the containment, the pressure will ultimately reach the failure pressure of the drywell.

• The pressurization rate of the primary containment during a Loss of DER accident sequence depends on the nature of the initiating event, Therefore it is convenient to the purposes of this study to group all i

initiating events into four classes according to their ef fect on suppres-sion pool heatup.

Each of these classes is discussed in the following sec-tions.

t 2.1 Transients with Uniform Pool Heatuo The accident initiators in this category lead to Loss of DHR accident sequences in which the pressure suppression pool can be treated as a well-mixed volume of water undergoing a uniform pool heatup.

This requires that at least one loop of the RHR system be operable for circulation and mixing of the pressure suppression pool water, even though the heat re-moval function of the loop is not available.t Loss of DBR accident sequences with the assumption of uniform pool heatup are discussed in Chap. 3.

• As discussed in Ref. 2.1, static overpressure [0.910 MPa (117 psig)]

is expected to cause f ailure of the primary containment in the drywell at the cylinder-sphere interf ace.

tFor example, the failure might be in the RHR service water system, leaving the RHR system fully available for suppression pool circulation.

Experimental results discussed in Ref. 2.2 show that the mixing induced by operation of one RHR loop will ef fectively eliminate thermal stratifica-tion in the pressure suppression pool.

This mixing is enhanced by a

(

piping elbow within the torus that directs the RHR pump discharge in a l

direction parallel to the torus axis, i.e., around the circumference of l

the pool.

This elbow will be installed within the Unit I torus during the refueling outage to begin in March, 1983.

It has been assumed to be in l

place for the calculations performed in this study.

. _ _ ~. _ _, _. -... _. _ _ _ _,.__.

=.

_=..

l 7

2.2 Transients with Pool Thermal Stratification l

If in addition to the loss of DHR function, no RHR pump and basic piping loop is available for suppression pool circulation and mixing, then significant thermal stratification will occur in the pool water during the heatup.

[See the discussion in Appendix C].

Since the temperature of the i

i upper layer of the suppression pool water will be significantly higher than the pool bulk average temperature, the containment pressurization rate will be higher and the drywell failure pressure will be reached earlier.

It should be noted that operator action to manually operate the re-actor vessel relief valves, alternating among the 13 valves as required to distribute the relief valve discharge evenly around the circumference of the pressure suppression pool, would also be ef fective in providing a more uniform pool heatup. However, manual relief valve actuation is only poss-ible when the available control air pressure is 0.172 MPa (25 psi) or more higher than the pressure in the drywell.

Since the average drywell con-trol air pressure is 0.722 MPa (90 psig), manual relief valve actuation will not be possible-af ter the drywell pressure has reached 0.550 MPa (65 psig). This will occur in every Loss of DHR accident sequence that pro-coeds to containment failure, which is expected to occur at 0.910 MPa (117 psig).s.2 When the relief valves can no longer be manually operated, the reactor vessel will repressurize to the setpoint [7.722 MPa (1105 psig)]

)

for automatic actuation of the lowest-set relief valve and this relief valve will repeatedly actuate thereaf ter.

l Loss of DHR accident sequences analyzed with consideration of the effect of thermal stratification in the pressure suppression pool are dis-cussed in Chap. 5.

2.3 Stuck-ocen Relief Valve The third broad category of Loss of DHR accident sequences involves transients with a stuck-open relief valve.

These cases are represented in two of the eight dominant sequences identified for Browns Ferry Unit 1 by the IREP study *.s and therefore analyses of the follow-on accident sequences have been included in this study.

The Loss of DHR accident sequences with a stuck-open relief valve differ from those discussed in Chaps. 2.1 and 2.2 because the reactor ves-set remains depressurized to a pressure about 0.396 MPa (50 psi) above containment pressure throughout the latter part of the accident sequence.

This alters the characteristics of the energy addition to the suppression

pool, i.e.,

there is a relatively slow continuous discharge from the T-quencher (located at the tailpipe terminus of the stuck-open valve) in-stead of the intermittent bursts of high steam flow which occur in the other cases during the periods when the reactor vessel is pressurized.

Loss of DHR sequences with a stuck-open relief valve are discussed in Chap. 6.

_ _, __,_____.~,

8 2.4 LQCA The fourth broad category of Loss of DHR accident sequences comprises those sequences associated with a loss of coolant accident (LOCA) in the d rywel l. These accident sequences were not included among the dominant sequences identified by the IREP study and therefore are not considered in detail here.

Nevertheless, it should be noted that with an intermediate or large LOCA in the drywell, the decay heat energy enters the pressure suppression pool through the 96 downcomers rather than through the relief valve T-quencher discharge devices.* Since the T-quenchers are located about 3.20 m (10.5 f t) beneath the surf ace of the pool and the downconers discharge Just 1.04 m (3.4 f t) below the surface of the pool, it is probable that thermal stratification would be more severe in the case of a loss of DRR function following a LOCA in the drywe11.8.4 References for Section 2 2.1 L. G. Greimann, et al., " Reliability Analysis of Steel Containment Strength," NUREG/CR-2442, June 1982.

2.2 B. J. Patterson, " MARE I Containment Program Monticello T-quencher Thermal Mixing Test Final Report Task Number 7.5.2," NEDO-24542 Class I, August 1979.

2.3 S. E. May s, et al., " Interim Reliability Evaluation Program: Analy-sis of the Browns Ferry, Unit 1, Nuclear Plant," NUREG/CR-2802, BGG-2199, July 1982.

2.4 E. W. Wong and H. S. Yao, " MARK I Containment Program Downcomer Re-duced Submergence - Functional Assessment Report Task Number 6.6,"

NEDO-21885 Class I, June 1978.

• For a small LOCA in the drywell, the entry to the pressure suppres-sion pool would be divided between the T-quenchers and the downconers.

l

___._._--,_.-__.,-_.,_.._m

_..,. - _ ~ _ _

9 3.

TRANSIENTS WITH LOSS OF DHR: CALCULATIONS ASSUMING UNIFORM POOL Tr' APERATURE 4

3.1 Introduction The defining system failures for the Loss of DHR accident sequence occur af ter an initiating incident and successful reactor scram, whereby the Main Steam Isolation Valves (MSIVs) close, the condenser cannot func-tion as a heat sink, and the RHR system is unable to provide either sup-pression pool cooling or shutdown cooling. The steam produced by decay j

heat is relieved from the reactor vessel by the SRVs and is condensed in the suppression pool. The temperature of the uncooled suppression pool increases monotonically, leading to escape of steam from the pool surface and therefore to a pressure buildup which eventually causes high pressure f ailure of the drywell.

The vessel wster injection function is not initially impaired, and it j

is assumed that the operators would act to maintain reactor vessel water level near the normal 14.25 m (561 in.) above vessel zero. Manual control i

of the SRVs is also initially unimpaired, and the operators would control reactor vessel pressure according to the emergency operating instructions which require initiation of a 56*C/h (200'F/h) depressurization before suppression pool temperature exceeds 49'C (120*F).

If the RHR system can be operated to circulate the suppression pool water,* It is assumed that the operators would do so.

The calculations reported in this section were performed with the ORNL-developed BWR-LACP code. Appendix B gives detailed input assumptions and discusses sequence-specific modifications which were necessary to ade-quately model these sequences. The most significant modeling assumption for the results reported in this section is that the temperature of the suppression pool is uniform throughout the pool.

It is known that hotter pool water tends to rise to the top of the pool and that the water in the vicinity of a discharging SRV T-quencher is hotter than bulk pool tempera-ture.8.1 However, with at least one RHR pump operating (without heat re-moval) to circulate the pool water, both of these ef fects would be mini-mixed. Modifications are planned for the next (March 1983) refueling 4

outage to equip the Unit 1 pool cooling discharge lines with elbow and fittings which wil) discharge horizontally along the circumference of the torus to promote circulation and mixing of the whole pool.

In addition, the Browns Ferry emergency operating instruction for main steam isolation valve (MSIV) closure requires that operators alternate their selection of relief valves in order to minimize local temperature buildup in the vicin-ity of a discharging T quencher.

If the suppression pool is not mixed by the operation of at least one RHR pump, the net ef fect of locally higher temperatures would he a more rapid buildup of primary containment pressure than reported in this sec-tion.

The results reported in Chaps. 5 and 6 were calculated using a e

• Albeit without the RHR heat exchanger function to provide pool cooling.

A

)

.w.,, -.-. -

.m.,

,,--._---,,,-.-,-_--.-,-,,,.m_,.

n~_..

,_-,._--,_.-.,,,_,.,w,

,..,-,-.__,-__..%--_m,

=

4 10 special suppression pool model which can be used to calculate the tempera-ture as a function of location in the pool when the water is not well-mixed.

3.2 Summary and Conclusions Following accident initiation, the operators would maintain reactor vessel water level by control of the Reactor Core Isolation Cooling (RCIC) the 0.011 m /s (170 spm) inj ection provided by 8

system. Af ter about 4 h, the Control Rod Drive (CRD) hydraulic system pump is suf ficient to main-tain vessel level, without the aid of the higher capacity RCIC system.

Dependence of the vessel injection function upon the status of the pres-8 (362,000 gal) supply sure suppression pool is avoided because the 1370 m of cooling water stored in the Unit 1 Condensate Storage Tank (CST) is sufficient to last throughout the sequence.

As the suppression pool temperature increases, steam escaping from the surf ace of the pool increases the primary containment pressure until, 35 h after accident initiation, the 0.91 MPa (117 psig) failure pressure 8.8 of the drywell is exceeded. The calculations reported in this section are terminated af ter 35 h - no attempt has been made to model events af ter drywell f ailure with the BWR-LACP code.

If the flow area of the drywell rupture (the weak point in the dry-well is at the intersection of the spherical and cylindrical segments 8.8) were small, then the subsequent energy release would be spread over a long period of time. This would minimize the disruptive ef fect of drywell failure on safety systems in the reactor building and drywell; it is pos-sible that the vessel water injection could be maintained and that there would be no core damage at any time during the sequence.

If the flow area of the drywell rupture were large then a great amount of energy would be released over a short period of time with poten-tially catastrophic ef fects on safety systems in the reactor building and drywell. The subj ect of drywell f ailure modes is discussed in detail in Chap. 7.

The MARCH calculations reported in Chap. 8 were performed with the assumption that vessel water injection fails af ter drywell failure, leading to core uncovery and severe core damage.

3.3 Detailed Results Figures 3.1-3.13 show the BWR-LACP results for reactor coolant system and primary containment variables throughout the entire 35 hours4.050926e-4 days <br />0.00972 hours <br />5.787037e-5 weeks <br />1.33175e-5 months <br /> before primary containment f ailure. Table 3.1 summarizes maj or events during the first 35 hours4.050926e-4 days <br />0.00972 hours <br />5.787037e-5 weeks <br />1.33175e-5 months <br />.

f i

e-

-..-y.,.%

,y--

,m_----

c_.--w,.

-,y-. - -, -.

m--,-

+.-,

11 3.3.1 Reactor vessel water level Figure 3.1 shows water level in the downconer region of the reactor l

vessel.* Figure 3.2 shows total vessel injection flow rate and the total amount of water injected is shown in Fig. 3.3.

Throughout the Loss of DBR accident sequence the pref erred source of water for injection into the reactor vessel would be the Unit 1 Condensate Storage Tank (CST). The normal water volume of the CST is 1370 ms (362,000 gal) and this enount of water is assumed to be present at the beginning of the accident.

It is possible, but not likely, that there could be signifi-cantly less volume at the beginning of the accident.

The main condenser hotwell draws its make-up from a standpipe within the CST and flow through the standpipe could conceivably reduce the supply of condensate to 511 m s

(135,000 gal). For this to occur would require a breach of the condensate system because the Browns Ferry operating procedures require replenishment of the CST (from the 1420 m8 (375,000 gali domineralized water storage tank) upon receipt of a CST low level alarm (which corresponds to a volume of 1301 m8 (344,000 gal).

If CST level cannot be rapidly restored follow-ing a low-low level alarm at 579 m8 (153,000 gal), the procedure requires an orderly shutdown of the unit.

The suppression pool might be used instead of the CST as a source of water for vessel injection by the FCCS and RCIC systems; however, the HPCI and RCIC turbines depend on the pumped water for cooling of their lube oil.

The recommended maximum water temperature for long term operation is 60*C (140'F) (see Browns Ferry FSAR, Amendment 24, Section Q14.1-4), and this temperature is exceeded in the pressure suppression pool about 2.0 h j

into the loss of DBR sequence.

The calculation represented in Figs. 3.1 through 3.13 was initialized 30 s after reactor scram, with the MSIVs closed and with reactor vessel water level at 12.7 m (500 in.) above vessel zero.

Section 10.2 discusses the uncertainty in this assumed value of initial vessel level.

One cycle of High Pressure Coolant Inj ection (HPCI) system actuation, initiated by the operators before level decreased to the setpoint for automatic initia-tion, is required to bring level back to the normal range; after this and for the next four hours, the 0.038 m /s (600 spa) RCIC system is more thsn 8

adequate to maintain normal vessel level.

The control rod drive (CRD) hydraulic system pumps water into the reactor vessel throughout the sequence.

Following reset of the initiating scram (i.e. if the scram condition has cleared), the CRD hydraulic system inj ection drops to 0.0038 m /s (60 gpm). After 1 h,t the drywell pressure 8

i exceeds the 0.115 MPa (2 psig) high drywell pressure scram setpoint, caus-ing the 185 CRD scram inlet and outlet valves to open.

This second scram l

does not affect the already fully inserted control rods, but with the scram inlet valves open, the CRD hydraulze system inj ection flow to the reactor vessel increases to 0.011 m8/s (170 spa).

• This is the level which the control room instruments are de signed to indicate.

To read the full range of level variation shown on Fig. 3.1, operators would have to consult the wide range instruments.

I tThis assumes continuous operability of the drywell coolers (see Sect. 3.3.4).

l

i 12 Af ter about 4 h, the CRD hydraulic system is providing all the vessel water injection and the RCIC system is no longer needed.

Several hours later (8.6 h after event initiation) the full amount of CRD vessel injec-tion is more than enough and the operators have to take action to prevent excessively high vessel water level. This could be accomplished either by intermittent CRD pump operation, or by throttling the CRD hydraulic pump di schar ge. The Unit 1 "A" CRD hydraulic pump can only be throttled by l

local-manual control of the discharge valve.

The "B" pump (which serves as a spare pump for both Units 1 and 2) can be throttled from the control room by remote-manual control of its discharge valve; therefore, the operators would most likely switch to the "B" pump in order to maintain i

continuous control of injection flow from the main control room.

Late in the sequence, when remoto-manual control of the SRVs is lost j

(see Sect. 3.3.2) the reactor vessel undergoes repressurization and for a period of 4 h, while vessel pressure is building to the automatic SRV ac-tuation pressure, no steam is lost from the vessel.

During this period the approximately constant mass of water in the vessel undergoes a thermal j

6 :nsnsion of about 22%, 'swe111ng to well above the normal range (some l

water would overflow into the main steam lines) even though all vessel injection is cut-of f during most of the repressurization.

3.3.2 Reactor vessel crossure Figures 3.4 and 3.5 show reactor vessel pressure and total steam flow l

to the suppression pool.

Since the MSIVs are closed, steam produced in the reactor vessel is l

relieved to the suppression pool through the SK(s and through the RCIC and HPCI turbine exhaust during the periods when these turbines are operating.

The lowest-set SRVs would actuate automatically at 7.72 MPa (1105 psig) and reclose af ter vessel pressure has been reduced by about 5%.

However, the Browns Ferry emergency operating instructions require the operator to minimize automatic SRV actuations by remote-manual operation of a single SRV at a pressure slightly lower than the setpoint for automatic actuation so as to reduce vessel pressure by about 20% instead of 5%.

This not only j

minimizes the total number of valve actuations but also allows the opera-tors to alternate their selection of SRVs around the suppression pool such that local pool heatup in the neighborhood of a discharging T quencher is minimized.

When pool temperature reaches 498C (120*F), the emergency operating instructions require that the operators initiate a 558C/h (100*F/h) de-pressurization of the reactor vessel, with the final target pressure below 1.48 MPa (200 psig). This depressurization rate is achieved at first by intermittent, and then by contine : operation of a single SRV.

The final pressure attained is about 0.69 MPa (85 psis) well above the isolation pressure of the RCIC turbine steam supply line.* The history of the re-actor vessel pressure during the accident sequence is shown in Fig. 3.4.

After depressurization, with vessel pressure in the neighborhood of 0.69 MPa (85 psig), the steam production rate is nearly in balance with

• The RCIC system automatically isolates if the reactor vessel pres-sure drops below 0.448 MPs (50 psig).

i 13 the capacity of a single SRV, so that, with one SRV remaining open, vessel pressure does not decrease further. Although not apparent on Fig. 3.4, the operators would occasionally close the open SRV and simultaneously open another SRV to direct the discharge to another part of the pool.

In order to enable remote-manual SRV operation, the pressure of the drywell control air must exceed the drywell pressure by at least 0.17 MPa (25 psid). The Group II isolation on high drywell pressure (Groups VI and VIII isolations are also triggered by the high drywell pressure signal) which occurs af ter about 1 h would' isolate the drywell control air suc-tion, thereby compromising long-term remote-manual operability of the SRVs (and also operation of the drywell coolers, whose discharge dampers re-quire control air to remain open). This situation cobid, however, be remedied because the operators are required by emergency operating in-l structions to valve-in the station control air, which is maintained at a pressure

• very close to that of the drywell control air.

Station control air compressors A and D can be powered by the diesels in the event of loss i

of offsite power, but they would have to be restarted locally after the load-shedding which would occur af ter about 2 h due to combined low re-actor vessel pressure and high drywell pressure.

Af ter about 24 h, the drywell pressure exceeds 0.55 MPa (65 psig),

and there is no longer the pressure differential required for remote-manual SRV actuation.

The open SRV therefore closes, and cannot be opened in response to operator action. Af ter the reactor vessel repressurizes to 7.72 MPa (1105 psig), the lowest-set SRV would begin automatic actuation as shown in Fig. 3.5 (this mode of SRV operation does not require control air).

It would be desirable to maintain a depressurized reactor vessel throughout this sequence in order to minimize heat losses to the drywell atmosphere. The design temperature of many components in the drywell (including the SRV remote-manual actuation solenoids) is 13 88C (281*F).

Additionally, when vessel pressure is low, both high and low pressure in-jection systems would be able to function if necessary.

3.3.3 Sucoression pool temperature and water level Suppression pool temperature (Fig. 3.6) and water level (Fig. 3.7) increase steadily throughout the Loss of DHR sequence except during pe-riods when there is neither SRV discharge nor any HPCI or RCIC turbine ex-i haust into the pool. The T-quencher underwater stema discharge device has replaced the ramshead design at Browns Ferry for SRV steam discharge.

The T quenchers can produce smooth condensation at pool temperatures up to i

near-saturation without the instability phenomenon known to occur with the ramshead device.

The calculations reported in this section assume that complete con-densation will occur if the pool is at least 1.1*C (2*F) subcooled.

Un-certainties associated with this assumption are discussed in Sect. 10.2.

The rate of increase of pool temperature (Fig. 3.6) is greatest during the first several hours of the sequence because the production of decay heat is greater during this period and also because of the reactor

• Average station control air pressure is about 0.724 MPa (90 psig).

f

..-..mm-

. -. - - _. _ -.. _..... - _,..,, _ _,., _ -, _. - - _, - _,.. -. _ _ _. _ _. -. _ _ ~,. _,,, -

14 vessel depressurization, which begins af ter 1 h [when pool tasperature reaches 49'C (120'F)]. When the reactor vessel has been depressurized about 3 h af ter sequence initiation, the pool tamperature has increased to 71'C (160*F).

When pool temperature reaches 100*C (212*F) one might expect that.

condensation of the SRV discharge or turbine exhaust steam would cease.

l This does not happen because by this time the total wetwell pressure has lacreased from atmospheric pressure to 0.18 NPa (11 psig) and the corre-sponding saturation temperature at the surf ace of the suppression pool is 117'C (242'F) instead of 100'C (212*F). The pool heating during the Loss of DER sequence is slow enough such that the evaporation of water vapor from the pool surface can contribute to the total (nitrogen plus water vapor) pressure over the pool.

In this manner, the pool remains slightly subcooled during the heatup and continues to condense 100% of the SRV i

steam discharge.

At the time when the drywcil failure pressure is exceeded (35 h of ter sequence initiation), the suppression pool water temperature has increased to 173*C (343*F).

Suppression pool water level (Fig. 3.7) increases not only because warmer water is less dense, but also because of the additional mass of condensed steam in the pool.

At the end of the sequence (35 h), the water i

level has increased by 1.37 m (4.5 ft).

Plant instrumentation would in-dicate a level lower by about 10% as this measurement is not temperature compensa ted.

The suppression pool-to-drywell vacuum breakers would at this point be partially submerged [i.e., water level about 0.15 m (6 in.)

above the bottom of the 0.46 m (18 in.) valves], but should still be able to function to keep drywell pressure from being significantly below wet-well pressure. The calculations reported in this section assume that the wetwell-to-drywell vacuum breakers are unimpaired; thus, whenever wetwell pressure exceeds drywell pressure by more than 3.45 kPa (0.5 psid), the vacuum breakers will open and equalize the pressures. The vacuum breakers open numerous times during the sequence because there is a significant i

amount of net mass transfer (water vapor plus nitrogen) from suppression pool atmosphere to drywell atmosphere.

l l

3.3.4 Primary containment stuoschere cressure and temocrature The major driving force which af fects primary containment pressure (Fig. 3.8) is the vapor pressure of the suppression pool water, which in-creases from 4.8 kPa (0.7 psia) to 848 kPa (123 psia) as the pool is heated from 32*C (90*F) to 17 3

• C d 2 ' during the 35 h Loss of DHR se-quence before containment failure.

2ne drywell pressure and suppression pool pressure (not shown) remain very close because of the action of the 12 wetuell to drywell vacuum breakers, which prevent the wetwell pressure from exceeding drywell pressure by more than 3.45 kr.. (0.5 psid).

The temperature of the wetwell atmosphere (no. shown) is held very close to suppression pool water temperature (Fig. 3.6) throughout the se-quence by combined convective and evaporative heat transf er from the sur-face of the slowly heated pool. The drywell temperature (Fig. 3.9) is determined by competing influences:

the hot surfaces of the reactor ves-sel and piping, the cool surfaces of heat sinks such as the 2.86 cm (1.125 l

i l

l l

15 in.) thick steel drywell liner, the influx of hotter nitrogen and steam from the wetwell atmosphere, and heat removal by the drywell coolers.

The design heat removal capacity of the drywell coolers is about 1.5 NW, but the actual heat removal rate depends on drywell atmosphere tem-perature and humidity. During a Loss of DHR sequence the coolers help to control not only drywell temperature but also primary containment pressure by condensing part of the steam wh.ich flows from the wetwell airspace to the drywell. The drywell coolers rr.a continuously and are available af ter

~

accident initiation.

Following a loss of offsite power initiator, the drywell coolers can be powered from an emergency diesel generator bus.

After receipt of the core spray initiation signal,* the diesels shed nonessential loads includ-ing the drywell cooler blowers.t There is system logic which would pre-vent the operators from subsequently restarting the blowers from the con-trol room. The coolers can, however, be restarted and operated by utiliz-ing local handswitches on the 480 V shutdown boards and motor control centers.

Since the Browns Ferry emergency operating instructions do not pro-vide explicit procedures for restart of the drywell coolers under emer-gency conditions, and because loss of of f-site power is a potential loss i

of DHR initiator, two different calculations of primary containment pres-sure and temperature have been performed.

One calculation (Figs. 3.8 and 3.9) assumes that the drywell coolers run only up to the time (2 h after i

j event init.'ation) of the core spray initiation signal; the other calcula-i tion (Figs. 3.10 and 3.11) assumes that the coolers are restarted after load shed, and continue to run until the blowers f ail (17 h) due to the combined deleterious effects of high drywell pressure and temperature [the assumed f ailure temperature is 93*C (200*F)]. The variables shown on Fiss. 3.1 through 3.7 and discussed in the preceding subsections were cal-culated assuming loss of the drywell coolers af ter 2 h.

The effect of the drywell coolers on the performance of these variables af ter 2 h is negli-gible; therefore, no discussion of the effect of drywell cooling was pro-vided in the corresponding subsections (3.3.1, 3.3.2, or 3.3.3).

As shown by' a comparison of Figs. 3.10 and 3.8, extended operation of the drywell coolers can delay by about 2.5 h the eventual high pressure f ailure of the drywell. A comparison of Figs. 3.11 and 3.9 shows that the drywell temperature is also lower for the case with extended operation of the drywell coolers and this would have the beneficial ef fect of main-taining the temperature-sensitive equipment in the drywell below the 13 8'C (281*F) long-term and the 163*C (325'F) short-term design temperatures for an additional period of about 3 h.

This equipment includes the solenoid valves which are necessary for remote manual operation of the SRVs.

l

• The core spray system pumps automatically start upon a combination of high drywell pressure and low reactor vessel pressure.

l tThe logic provides that the drywell cooler loads are shed if there is a core spray initiation signal and if diesels are running and loaded, l

l i

l,-

-.~. -.-.--- -.-.-.

16 i

I 3.3.5 ECCS caso net oositive suction head (NPSH)

Pump NPSH* is of special concern during the Loss of DHR sequence due to the need to pump the very hot suppression pool water. For example, the I

most direct way to recover from the Loss of DHR accident would be to re-gain the suppression pool cooling mode of the RHR system. The success of recovery would depend on whether the RHR system could pump without severe cavitation if the pool is heated to 100*C (212*F) or more.

j In plant testing at Browns Ferry has shown that the RHR pumps can operate down to about 65% of the manuf acturers recommended minimum NPSH with the following consequences:

10% degradation of developed pumping head, acceptable pump motor vibration, but severe audible cavitation.

l This would not jeopardize short-term operation although impeller cavita-tion damage would be expected in the long-term. The in plant. tests did not include reduction of NPSH to the point at which short-term pump opera-tion would be j eopardized by sudden and severe loss of developed head and/or severe vibration.

Fig. 3.12 shows the calculated NPSHf with one RHR pump operating to circulate suppression pool watert at 0.63 m /s (10,000 spa) throughout 8

the pool heatup. For the case in which the drywell coolers cease opera-tion af ter 2 h (see discussion in Section 3.3.4), the NPSH is greater than 90% of the manufacturer's recommended minimum at all times. Thus no diffi-culty with pump operation should occur. For the case of extended drywell cooler operation (curve 2), the RHR pumps could probably not function at full flow af ter 14 h (840 min) since the NPSH would be below the degraded region explored by the Browns Ferry tests, and attempted operation could result in pump motor failure and/or loss of all pump developed head.

When the drywell coolers operate, more water vapor escapes from the pool surf ace, mixes with the wetwell atmosphere above the pool, then flows through the vaccum breakers to the drywell where much of it is condensed.

l This process lowers the total pressure in the primary containment and tends to wash the nitrogen out of the pool atmosphere so that af ter 14 h there is only steam and water lef t in the suppression chamber and satura-tion conditions exist. Nitrogen is also washed out of the wetwell atmo-sphere when the drywell coolers are not operating, but to a much lesser extent.

The NPSR margin for acceptable RHR pump operation can be extended by operator action, throttling the flow as necessary to reduce the RHR pump discharge from the rated flow of 0.63 m8/s (10,000 spm). With reduced flow, there is a slight decrease in the required NPSH at the pump inlet

• NPSH is the static plus velocity pressure at the pump inlet, less l

the vapor pressure of the fluid being pumped, expressed in equivalent head of the fluid being pumped. The manufacturers minimum recommended NPSH is based upon a 3% decrease in developed head but no significant audible a

cavitation, tThe calculation takes into account the increased depth of water in the pool (Fig. 3.7) which increases the NPSH at the RHR pump suction by about 0.76 m (2.5 f t) at the 14 h point.

iBut without pool cooling.

i l

l

. _ _, - - _ _.. __~-_ _ -

17 according to the informa tion supplied by the pump manufacturer.* A much more important effect is that the reduction in flow serves to increase the actual NPSH available at the pump inlet by reducing the f rictional pres-sure losses incurred in the suction piping from the pressure suppressica pool. During the Browns Ferry RER system tests, the RHR pump continued to operate well as the NPSH was lowered to 5 m (16.4 f t) with full flow. At 80% of full flow, the RHR pumps were observed to perform well at an NPSH of 4.33 m (14.2 f t).

Calculations have been performed to determine the lowest NPSH which might be encountered during the long-tern Loss of DHR accident sequence it RHR flow were reduced to 80% of normal. The lowest calculated NPSH is 4.94 m (16.2 f t) which is higher than the region at which the pumps were demonstrated to be operable during the Browns Ferry tests.

Therefore, if i

r the RHR pump discharge were throttled, the RHR pumps could be operated to provide pool circulation throughout the Loss of DHR sequence even in the case of extended drywell cooler operation.

3.3.6 Reactor buildina environmental considerations This section provides an evaluation of the ef fect of the hot, uninsu-lated torus on the temperature of the reactor building atmosphere.

An-other concern might be the heating effect over long periods of operation of the ECCS pumps, which are located in the corner rooms of the reactor building basement; however, excessive building air temperature from these sources is prevented by the ECCS room coolers, the ventilation flow main-tained by the Standby Gas Treatment ( SGT) systan,t and the thick concrete walls acting as heat sinks.

The Browns Ferry pressure suppression chamber is located in a room (torus roon) which occupies the central portion of the reactor building l

basement. The torus room is 11.6 m (38 f t) from floor to ceiling and is essentially closed except for four open 1.83 m (6 f t) height manways lead-ing to the reactor building corner rooms, and several relatively small openings in the ceiling (such as the annular space between pipes which extend into the torus room through the drywell personnel access room).

As the suppression pool temperature increases, the surface of the torus begins to lose heat by radiation to the thick concrete walls and by natural convection to the torus room air.

Hotter air would tend to rise, and stratify at the top of the torus room. There would be a net circula-tion of air into the torus room from the basement corner rooms that would exit into the ventilation ductwork and also into the drywell personnel access room.

This circulation of air from the torus room would be rela-tively small (see Browns Ferry FSAR, Section 5.2.6.3) and not capable of transporting a large amount of heat to any of the major floor areas of the reactor building.

e

• The required NPSH is 7.93 m (26 f t) at full flow and 7.62 m (25 f t) at 80% of full flow.

tThe SGT system is automatically actuated when the drywell pressure reaches 0.115 MPa (2 psig).

l l

l l

l 18

{

The temperature of the toras room air and concrete have been cal-culated to estimate the rate of heat loss from the torus. The results indicate that the most significant heat loss is by radiant heat transfer l

to the concrete walls, ceiling, and floor.

The torus room air temperature remains approximately mid-way between concrete and torus surf ace tempera-tures. Figure 3.13 shows the average torus aurf ace temperature, the torus room air temperature, and the surf ace tamperature of the concrete sur-roundings. Torus room air temperature exceeds the 93*C (200*F) isolation setpoint of the HPCI and RCIC turbine steam lines after 13 h.*

By this time, suf ficient reactor vessel inj ection is being provided tolely by the I

CRD hydraulic pumps, so this isolation would not be a serious problem. At I

the end of the 35 h calculation, when the drywell is predicted to f ail by overpressurization, the torus surface is at about 166*C (330*F), the torus room air is 1478C (297'F) and the surface of the concrete has been heated to 132*C (269'F).

l

(

References for Section 3 3.1 B. J. Patterson, " Mark I Containment Progrma - Monticello T-Quencher Thermal Mixing Test Final Report," GE NEDO-24542 Class I, August 1979.

l 3.2 L. G. Greinann et al., " Reliability Analysis of Steel Containment Strengtu," NUREG/CR-2442, June 1982.

• The HPCI system can be isolated by temperature sensors located in I

both the torus room and in the HPCI room.

The RCIC system can also be isolated by temperature sensors in the torus room. These sensors are po-sitioned to detect steam leaks from the HPCI and RCIC turbine steam lines.

l l

1 I

19 Table 3.1.

Timetable of events for unmitigated loss of DHR with uniform pool heatup (h)

Event O

Initiating reactor trip followed by MSIV closure and f ailure of both pool cooling and shutdown cooling modes of the RHR system.

j 1

High drywell pressure scram at 0.115 MPa (2 psig). Diesel gener-ators and SGT3 automatically initiated. Drywell control air compressors isolated. Operators valve station control air into drywell control air header.

1 Pool temperature exceeds 4980 (120*F) - operators begin con-trolled depressurization of reactor vessel.

2 Core spray initiation signal [ reactor vessel pressure <3.21 MPs (465 psia) and drywell pressure >0.115 MPa (2 psig)] causes load shedding if loss of offsite power is still in effect.

Operators must use local control stations to restore diesel power to station control air compressors (A and D) and drywell coolers.

2 Suppression pool temperature exceeds the 60*C (140*F) recommended maximum temperature for cooling of RCIC and HPCI lube oil.

4 CRD hydraulic system provides sufficient reactor vessel inj ection - no RCIC system operation af ter this time.

8.6 Operators must begin to throttle CRD hydraulic system pump to avoid overfilling the reactor vessel.

13 HPCI and RCIC system steam supply line isolation caused by high

[93*C (200*F)] torus room temperature.

14 RCIC turbine high exhaust pressure trip at containment pressure

>0.28 MPa (25 psis).

21.5 Drywell design pressure [0.49 MPa (56 psig)] exceeded.

23.5 SRVs become inoperative in remote-manual mode because drywell pressure exceeds 0.55 MPa (65 psig).

35 Drywell fails when internal pressure exceeds 0.91 MPa (117 psig).

Suppression pool temperature has increased to 173*C (343*F).

8

20 ORNL-DWG 83-4222 ETD g

{

REPRESSURIZATION COMPLETE-g-

LEVE L SWE LL CAUSED BY HEATUP DURING g~

CRD REPRESSURIZATION w

INJECTION REDUCED g.

n 5

f g#

I l

a

d. J

-CRD

-CRD

>u u

CRD INJECTION

, INJECTION MODE RESTO ED RESTO r

s-d RCiC CRD W'

$a AND CRD HYDRAULIC e

W HYDRAULIC SYSTEM g

SYSTEM ONLY g

......................................................O..P O.F A CT I V E F U E L T

n BOTTOM OF ACTIVE FUEL o 160 260 36o 400 s60 soo 760 e60 960 tobo sibo 12bo in in in ishc i7bo th th 2 abo 2ico TIME (MIN)

Fig. 3.1.

Unmitigated Loss of DHR - reactor vessel water lebl.

b

21 e

e ORNL-DWG 83-4223 ETD x,

• ---SINGLE HPCI INJECTION TERMINATED BY OPERATORS g.

A-R-

M-In N' A g. E@*

e En-5 o g-S C' h -

is%

3 e_ d w-y B d 2 l--

U1 I"

W

~

OPERATOR ADJUSTMENTS TO RCIC FLOW CRD HYDRAULIC SYSTEM e-FLOW TERMINATED DURING USE OF 'RCIC h

VESSEL REPRESSURIZATION DISCONTINUED e-N PERIODIC ADJUSTMENTS TO CRD p

aLPdi HYDR AULIC SYSTEM F LOW

}

o-o-

C Inc 200 300 400 500 600 700 800 900 :000110012001300140015001600170018C0 IF 3XX) 2100 TIME (MIN)

Fig. 3.2.

Unmitigated Loss of DHR - total rate of inj ection flow.

22 ORNL-DWG 83-4224 ETD l-

~

g_

R INITIAL VOLUME OF WATER IN CONDENSATE STORAGE TANK f

5

$~

T (3

E~ k-

~

L1 id b-5 d

/

y k-

/

'=

e-NO WATER INECTED WATER USAGE GECREASES DURING REPRESSURIZATION h Sih.. AFTER DEPRESSURIZATION IS COMPLETE g gj a

y V*

~

k'

~

O' O

0 100 200 300 100 500 00 700 800 900 ITC 1100 ITO 1300 240015001600170010001900 2000 2100 TIME (MIN)

Fig. 3.3.

unmitigated Loss of DHR - total volume inj ected into reactor vessel.

e

23 ORNL-DWG 83-4225 ETD e-U

~

-DEPRESSURIZATION DUE TO OPERATION OF HPCI 8.

gOPERATOR CONTROLS PRESSURE USING 1 SRV i

[ OPERATORS INITIATE DEPRESSURIZATION x

BECAUSE OF HIGH POOL TEMPERATURE AUTOMATIC g.

~

ACTUATION f

OF 1 SRV

• ~

a DEPRESSURIZATION AT 55.5 K/h (100 F/h) p@

l BY OPERATOR ACTUATION OF 1 SRV

{s e

a i

b8-VESSE L ta"g y'

R EPR ESSU RIZES 0

89 l

r o

E

'f cr:

N"~ E j

Hg' f

m, I

REMOTE-MANU AL OPERATION OF

"~

y-l' SECOND SRV TO HIGH DRYWELL PRESSURE-SilVs NO LONGER POSSIBLE DUE OPENED SECOND SRV

~~

CLOSED o

8 -.

~

l STEAM VENTED CONTINUOUSLY THROUGH 1 SRV" a.

o 0 100 200 300 400 500 600 700 000 900 1000 11dC ;200 1300 14'00 1500.600 '.700 ;800

" 2000 2100 TIME IMINI Fig. 3.4.

Unmitigated Loss of DHR - reactor vessel steam pressure.

1 e

24 ORN L-DWG 834226 ETD k'

O, AUTOMATIC

~

SRV ACTUATION 1

8 k~

)

t 8-UG-M C

I:

p; pa DEPRESSURIZATION IN PROGRESS l

k OPERATOR MAINTAINS DESIRFD h

~~

3 I

RATE WITH 1 SRV 3 g,' d s-il d

E s_ E i

N W

Il s

n --

o j;

S W

/

J i

l l

i C,Of h

SECOND SRV OPENED

>a" i

M TO MAINTAIN DESIRED l

(

DEPRESSURIZATION RATE

! g..

1 NO SRV ACTUATION gJ

\\

DURING

'

• SECOND SRV CLOSED REPRESSURIZATION S-N' e

~

l:

l SINGLE SRV OPEN CONTINUOUSLY O-O o too s 360 a s6o c6c 76o ebo 960 i000it'aoidoo1300ticoisoo:50o1700tonoisco2aco2:00 TIME (MIN)

Fig. 3.5.

Unmitigated Loss of DHR - total steam flow from reactor vessel.

4

25 ORNL-DWG 83-4227 ETD E

8-NO POOL HEATUP DURING C-l' REPRESSUR'ZATION

~~

E G a__

y y

yR N E-h M

U D.

• - n gn' v

n d

0 E

r El fi B-S g.

E M~

HEATUP RATE DECREASES E

S.

'N AFTER REACTOR VESSEL En M DEPRESSURIZATION M_

COMPLETE

_7 -

S-N' O 160 200 360 4$0 500 660 760 800 900 1000 1100 1200 1300 liba 1sb0 16b0 17C0 18bD l b 20b0 2100 TIME (MINI Fig. 3.6.

Unmitigated Loss of DHR - suppression pool temperature.

a

26 ORNL-DWG 83-4228 ETo 8

w.

NO SRV DISCH ARGE DURING REACTOR VESSEL U~

REPRESSURIZATION g_

i-5~

3 9~ W a

o d e-5 8~

ti.- d

,8 " a$

e 0'~,'

/

i "$c LOWER RATE OF STE AM CONDENSATION

@" W ~

AFTER DEPRESSURIZATION COMPLETE

~J t y?

1

-J 2

'O 160 200 300 400 550 600 750 e00 9b0 10N 11N 12N 13N lib 1500 1600 1700 1800 1900 E C 2100 TIME (MIN)

Fig. 3.7.

Unmitigated Loss of DHR - suppression pool water level.

27 e

e ORNL-OWG 83-4229 ETD E

C-M~.DRYWELL FAILURE PRESSURE g_

~

RATE OF PRESSURE INCREASE SLOWED DURING REACTOR VESSEL REPRESSURIZATION g.

En-2" b

~ g $. g.

3*b h

h M,. G& D.R YW E L L..D E..S.I G N P R E SSU R E..........

g d

') O o.

g e4 d '

4 d

'd 4

E,'

g n-e I

nw R

n.

DRYWELL COOLi!RS TRIPPED ON HIGH DRYWELL PRESSURE (AND NOT RESTORED BY OPERATORS)

O-O-

0 100 20 300 100 500 600 700 800 900 10001100120013001430150016001700180C 1900 2000 2100 TIME (MIN)

Fig. 3.8.

Unmitigated Loss of DHR - drywell pressure (with early trip of drywell coolers).

l O

l

28 e

OmNL-DWG83-4239 ETD El i

O 9

C 8 n-e ACCELERATED RATE OF TEMPERATURE INCRE ASE u **

8.1 DUE TO INCRE ASING REACTOR VESSEL Y

E TEMPER ATURE DURING REPRESSURIZATION a 3%

o a

3 h w 3< u g.

DRYWELL DESIGN TEMPER ATURE

• i ~ 0 (281 F) 0 I

t b

h E t2 F 7%

/

a 8

'j p rj r

gM a

a 4

3' We - DRYWELL COOLERS TRIP ON Si HIGH DRYWELL PRESSURE H

im 2co a e sm sm 7m em a m noc is in um :sm am um tu in zo 2:ce o

TIPI (NINI Fig. 3.9.

Unmitigated Loss of DHR - drywell ' temperature (with early trip of drywell coolers).

b i

= - ~ - -

29 A

i ORNL-DWG 83-4230 ETD S

2-g _.

DRYWELL FAILURE PRESSURE R-

~

0-RATE OF PRESSURE INCREASE E h-E~

SLOWED DURING REACTOR h..

M VESSEL REPRESSURIZATION e-8 S*- u v

m

.)

.3

$g DRYWELL DESIGN PRESSURE g e-u LJ Q.

x a

a 8 a e-d~

d 6

x xa,<

x e

N

" r,.

DRYWELL COOLERS Fall g_

l o-o i 200 300 400 500 600 700 000 900 1000 1100 1:00 1300 1900 1500 1600 1700 1800 1900 2000 2100 0 100.

TIME (MIm Fig. 3.10.

Unmitigated Loss of DHR - dowell pressure (drywell cool-ers operated until failure).

4

30 o

(

ORNL-DWG83-4232 ETO 0

b' B-E E-S g_

W W

R g

ACCELERATED RATE OF TEMPERATURE INCREASE t c

c DUE TO INCREASING REACTOR VESSEL e

6 60' TEMPERATURE DURING REPRESSURIZATION

/

h a-5 f

uH H

y U 8-U

.....?...........'.......... [...d b

h e

a E

gM~

J a

Y %-Y g

p g-DRYWELL COOLERS FAIL %

J c) a E-3 0 100 j

E

, 200 300 400 500 600 700 800 900 1000110012001R 3 4JD 15001600170018001900 2000 2100 TIMC inlN)

Fig. 3.11.

Unmitigated Loss of DHR - drywell temperature (drywell coolers operated until failure).

. c 9

e O

m

./

, z f:

Ti

'f ' ;', b i"

- 5

,/

ij s

i+

, i. 'i

- /

31 43

/

l-

<r p

4 g

2 i

1 e sf r p

!h v

e

,6

• • /

,/

/

, # 'i ORNL-DWG 83-4233 ETD

,; 8 O

CURVE 1: DRYWELL COOLERS TRIP ON CORE SPR AY INITI ATION

,' ', 7 SIGNAL AND ARE NOT RESTARTED L

CURVE 2: DRYWELL COOLERS OPERATE UNTil FAILURE O ERS

~ !

TRIPPED j

S- (CURVE 1)

_9-n.

E U

1 l,,

f S- {

r z g_

n.

a.

5.- 5

..M.AN.U F.AC..T U.R.E R..S....

g.

g, y

y MINIMUM NPSH 2

x x

NPSH INCR E ASES N

- DUE TO SLIGHTLY

(-

HIGHE R WETWELL k

PRESSURE CAUSED DRYWELL BY INCRE ASED E-COOLERS D RYWELL TEM-FAIL PERATURE DURING VESSEL REPRES-n-

SURIZATION o-n-

400 500 600 700 800 900 10001:00 la 13001400150016C0170018001900 "E00 21CC 0 100 200, 300.

TIME IMIN)

Fig. 3.12.

Unmitigated Loss of DHR - RHR pump net positive suction head.

O

32 i

ORNL-DWG 82-19309 18 0 17 0 346 16 0 326 poou TORUS AIR 15 0 306 SURFACE 14 0 286 O

Lt.

o 13 0

- o 266 p 120 - @ 246 3

110

- D 226 Y

10 0 20

.........gy;,4c.gIggy.g Rc gg w

80

- ww 16 6 w

70 CONCRETE 60 14 6 SURFACE 50 12 6 40 10 6 3

30 86 O

4 8

12 16 20 24 28 32 36 TIME (h) l l

Fig. 3.13.

Average torus room temperature during the Loss of DHR accident sequence.

l I

O

33 4.

NORMAL RECOVERY FRON AND MITIGATION OF LOSS OF DER FUNCTION 4.1 S pgtagy 3

Recovery of either the main condenser or the RHR system at any tir.e in the period of more than 24 h before containment f ailure would preclude further increase in suppression pool temperature and therefore also pre-vent the eventual primary containment f ailure.

Both the main condensate pumps and the RHR system can be powered from the diesel generators in the event of loss of off-site power.

Any one of the four Unit 1 RHR heat exchangers is capable of prevent-ing the suppression pool from reaching excessive temperatures.

In addi-tion (see Appendix A.5.5), two of the four Unit 2 RHR pumps and heat ex-changers can be aligned to cool the Unit 1 pool, and there are procedures and training to instruct operators in the use of this option.

Suppression pool temperature and primary containment pressure con-tinue to increase in an unmitigated loss of DER accident.

When drywell pressure reaches half of design pressure [0.241 NPa (20 psig)] the opera-tors would initiate primary containment sprays (see Appendix A.3). This action is required by both the Browns Ferry emergency operating instruc-tions and the Emergency Procedure Guidelines.4.1 The spraying of hot suppression pool water

• would result in a ten-porary pressure decrease (the magnitude of which would depend on circum-stances prevailing at the time of initiation), but would not be capable of preventing the eventual containment overpressure f ailure.

The maximum pressure decrease due to use of containment sprays would occur if the pool were uncirculated (as discussed in Chap. 5).

In this case the thermal stratification effect would result in hotter water et the surface of the pool and cooler water at the bottom. The RHR pump suction is near the bottom of the pool.

Initiation of containment sprays with a stratified suppression pool would therefore cause a significant temporary pressure decrease, but the spraying proce ss itself would promote some pool mixing and the pool would continue to heat due to continued condensation of the SRV discharge; therefore, the containment pressure would continue to in-crease and the time of containment f ailure would be only slightly post-poned.

The initiation of containment sprays (even using the hot suppression pool water) would have the significant long term beneficial effect of minimizing the temperature of the drywell atmosphere, which exceeds 204*C (400*F) in the latter stages of the Loss of DER accident (Figs. 3.9 and 3.11).

There are two non-standard operational strategies that could prevent the eventual primary containment f ailure that would otherwise be the re-salt of an unmitigated loss of DER accident.

One would be to open the drywell and/or suppression pool vent lines early in the sequence, to pre-vent an ultimate catastrophic containment f ailure.

The other would be to

• Use of the RHRSW system to spray cool river water into the primary I

containment is not considered here.

=

34

^

provide alternative pool cooling by a f eed-and-bleed maneuver which would feed cool river water into the torus and rej ect heated water from the torus. These operational strategies have not been evaluated by either the utilities or the NRC, so it is unlikely that they would be employed under accident conditions.

4.2 Minimum Number of RHR Coolers Necessary for Pool Coolinn To determine if one RHR cooler would adequately cool the suppression pool, a calculation, very similar to those reported in Sect. 3, was per-formed with the same input except that one RHR heat exchanger was assumed to start af ter 0.5 h and run continuously thereaf ter with nominal RHR flow of 0.63 m /s (10,000 gym) and RHR service water (RHRSW) flow of 0.28 m /s 8

s (4,500 gpm). The design RHR heat exchanger heat transfer coefficients j

(which include substantial fouling allowances on both tube and shell side) were used. The calculated peak pool temperature (Fig. 4.1) was 82*C j

(179'F), occurring after 10 h.

In order to determine the sensitivity of this re sult to conservative assumptions, the calculation was repeated with RHR heat exchanger effectiveness degraded by 10%, and with the nominal ANS (1979) standard decay heat (with actinides) increased by a factor of 1.2 during the first 900 s and by a f actor of 1.1 af ter 900 s.*

These conser-vatisms increased the peak pool temperature from 82'C (179'F) at 10 h to 88'C (191*F) at 11 h.

Either of these peak temperatures would be accept-able. The more conservative result is similar to a calculation performed by the TVA (in response to 10 CFR 50 Appendix R - Fire Protection require-ments) which also showed that a single RRR heat exchangert is adequate to keep pool temperature below 93*C (200'F).

An initial delay of an hour or two before starting suppression pool cooling would have little effect on peak pool temperature since the heat transfer in the RHR heat exchanger increases in direct proportion to pool temperature. A more substantial delay would lead to undesirably high pool temperatures and possibly to an insuf ficient RHR pump net positive suction head (NPSH). As discussed in Sect. 3.3.5, the NPSH of the RER pumps 'can be maintained above the recommended minimum by decreasing RHR flow from full capacity to 0.5 m8 /s (8000 spm). The RHR pumps and heat exchangers are designed for water temperatures up to 177'C (350*F) and pressure to j

3.21 MPa (450 psig), so they should be able to function at any time during an unmitigated loss of DHR sequence before primary containment f ailure.

• These multiplying f actors were chosen to confirm with those used in an existing TVA calculation.

tin the TVA calculations, the heat exchanger was started af ter 1 h, but it was assumed that the associated RHR pump was also being used to provide vessel water inj ection.. Therefore, pool cooling was interrupted periodically to allow for alignment to vessel inj ection followed by an assumed 10 min delay for realigament to pool cooling.

- -.,.. _ ~

s 1

]

35 i

4.3 Mitimation Measures 4.3.1 Primary containment ventina This section investigates the possibility that primary containment venting could prevent the drywell over pressure f ailure which eventually i

occurs in an unmitigated loss of DER accident.

The primary containment ventilation and inerting systems are shown in Fig. 4.2.

The Browns Ferry design includes two 5.1 cm (2 in. ) lines (one from the drywell and one from the wetwell) that come together into a com-non line (of the seme size) that is connected directly to Standby Gas Treatment (SGT) System ductwork.

The flow of primary containment atmo-sphere into these lines is controlled by globe valves 84-19 and 84-20 as shown on Fig. 4.2.

These vent lines can be used for minor pressure ad-justments during normal operation, but they are also designed for high pressure use in the containment reinerting operation which might be re-quired after a loss of coolant accident.

The control valves automatically isolate when drywell pressure exceeds 0.115 MPa (2 psig), but the opera-tors can over-ride the isolation f rom the main control room.

The 5.1 cm (2 in.) vent lines are obviously not large enough to hold primary containment pressure near atmospheric during an extended Loss of DER accident sequence.

In order to be able to prevent containment failure they would have to be capable of venting the total decay heat steam pro-duction at some pressure below containment f ailure pressure.

As reported in Chap. 3, it takes about 35 h for drywell pressure to exceed 0.91 MPa (117 psig), at which time the decay heat steam production is about 7.3 kg/s (16 lb/s). The drywell and wetwell vents, combined, can pass only about-1.14 kg/s (2.5 lb/s) at this elevated pressure.

To determine the maximum effect of this venting path, the Loss of DER calculations of Chap.

3 were repeated with the assumption of continuous venting through these line s during the accident sequence.

The results showed that drywell f ail-t are would be delayed by about 4 h (i.e. delayed f rom 35 to 39 h.)

The Browns Ferry design also includes two 46 cm (18 in.) lines (one

. f rom the drywell, and one from the wetwell) that can vent the primary con-tainment directly to the main ventilation system ductwork.

The flow through these lines is controlled by 46 cm (18 in.) butterfly valves 64-29 and 64-32 as shown on Fig. 4.2.

These ventilation system lines are used during shutdown when the primary containment is being inerted with nitro-gen prior to startup or when the nitrogen inerted primary containment atmosphere is being purged with air prior to personnel entry.

They are not intended for use under high pressure [ procedures require primary con-tainment pressure below 0.103 MPa (0.25 psis) prior to venting]. The 46 cm (18 in.) buttefly valves isolate when drywell pressure exceeds 0.115 MPa (2 psig); the operators cannot over-ride the isolation signal.

If'the operators, during the early part of a Loss of DER accident, were to open the 46 cm (18 in.) butterfly valves before they were auto-natically shut and held shut on a high drywell pressure isolation signal, j

the primary containment pressure could be held very pear atmospheric pres-sure and subsequent f ailure of the drywell by over pressu,rization would be prevented.

As reported in Sect. 3.3.1, the high drywell pressure signal would occur 1 h after event initiation.

However, if the wetwell-to-dry-well AP compressor were not operated following the initiating event and

36 the 5.1 cm (2 in.) vents were opened to slow the pressure rise, then it would take about 3.8 h, instead of 1 h, for the high drywell pressure sig-nal to occur (see the discussion of Item 8 in Sect.10.2). Therefore, the operators would have, at most, 3.8 h in which to consider whether to initiate the option of venting through the 46 cm (18 in.) lines.

The suppression pool temperature reached in a loss of DER sequence with a vented primary containment at atmospheric pressure would be limited to 100* C (212*F). Af ter the pool reached saturation, no more SRV exhaust steam would be condensed in the pool; the steam would escape from the pool surf ace and would have to be vented from the primary containment. The amount of decay heat-generated steam being produced at the time the pool temperature reaches 100*C (212*F) is about 10 kg/s (22 lb/s), or 16.5 m8s (35,000 cfa) at atmospheric pressure. This is the maximum steam venting rate that would occur during the accident sequence with vented contain-ment, and is about 25% above the SGT system blower capacity, but well be-low the main ventilation system blower capacity. Therefore, the main ventilation system would have to be operational to avoid a steam environ-The 16.5 m /s (35,000 cfa) of steam would 8

ment in the reactor building.

flow at a bulk velocity of 101 m/s (333 ft/s) in the 46 cm (18 in.) vent l ine. This high velocity would indicate some possibility of damage to the ductwork, and the attendant risk of the release of steam into the reactor building.

The possibility of ventilation system damage caused by releasing the decay heat-generated steam through the reactor building ventilation system must be weighed against inaction and the attendant risk of severe damage to the primary system caused by the subsequent catastrophic primary con-tainment f ailure by overpressurization.

The possibic post containment f ailure phenomena are discussed in Chap. 7.

4.3.2 Alternative noo1 coolina This section investigates the possibility that 'a feed-and-bleed sup-pression pool cooling method could limit pool temperature sufficiently to prevent the eventual containment f ailure otherwise caused by an unmiti-gated Loss of DER accident. The direct addition of cool water with the removal of a corresponding amount of heated pool water

• would cool the pool in a manner similar to that of the closed-cycle pool cooling mode of the RHR system. No emergency operating instruction (E0I) for such a procedure currently exists; the following brief analysis is only intended to establish the feasibility of the approach and to offer suggestions as to possible methods of Laplementation.

In an unmitigated loss of DER accident, pool temperature slowly in-crease s until containment f ailure.

If a sufficient flow of feed-and-bleed pool cooling could be started, the trend of increasing pool temperature would be stopped or reversed. The heat removal by feed-and-bleed cooling is directly proportional to the difference between the feed temperature

• In order for pool water level to remain constant, the bleed flow rate would have to exceed the feed flow rate because the bleed flow would include both the f eed flow and the condensed SRV discharge.

Nyw--w---,%_.n- - _ _ _

37

[ assumed to be 32*C (90'F) for the calculations presented here] and the bleed (i.e. bulk pool) temperature. As pool temperature increases, the temperature difference between the feed and the bleed temperatures becomes greater, so less cooling flow is required to remove the heat associated with the decay heat generated steam production.

This effect is quantified in Table 4.1, which specifies the suppres-sion pool temperatures (taken from the Sect. 3 calculations) reached with no pool cooling, along with the results of a simple heat-balance calcula-tion of the minimum flow of feed-and-bleed cooling that would prevent fur-ther temperature increase.

For example, in the first 5 h following the loss of DER, the suppression pool temperature would (without any pool cooling) increase from 32*C (90*F) to 87'C (189'F).

If a feed-and-bleed cooling flow of at least 0.13 m8/s (2020 gym) were begun at 5 h, there 1

I would be no further increase in bulk pool temperature *.

The RHR service water (RHRSW). pumps might be operable in a loss of DHR accident; if so, they could be used to feed river water directly into 4

the suppression pool (see also Appendix A.5.2).

Each of the four Unit 1 RHRSW pumps can pump 0.28 m8/s (4500 gym) against a head of 0.93 MPa (120 psig), and could therefore accomplish feeding even with primary containment pressurized to the drywell failure point of 0.91 MPa (117 psig). The RHRSW pumps can be powered from the diesel generators, so off-site power is not necessary for this operation.

The river water would be fed into the pool via the 46 cm (18 in.) recirculation pump test line which is also used to discharge cooled water into the pool when the RHR system is in the normal pool cooling mode.t The RHR drain pumps are used for routine suppression pool level ad-l justment.

Each of the two drain pumps can pump 0.05 m8s (800 spa), with a developed head of 46 m (150 f t), to the main condenser hotwell or and/or l

to the radwaste system.

These pumps might be operable in a loss of DER sequence, and, if so, they could remove hot water from the pool for feed-and-bleed cooling.

An alternative pool cooling strategy, based on 0.1 a /s (1600 sym) of s

RHRSW feed of river water with the RHR drain pumps used to bleed the same amount of hot pool water, would, if initiated after 7 h, prevent pool ten-perature from exceeding about 99'C (210*F) (see Table 4.1).

If this same cooling were initiated before 7 h, the peak pool temperature would be lower.

The risks attendant to feed-and-bleed pool cooling would, in general, be in proportion to how much time was allowed to elapse before beginning l

the procedure.

For example, af ter 12 h without pool cooling, the pool would have reached a temperature of 119'C (2478F) and a pressure of 0.25 MPa (21 psig).

If this hot water were bled f rom the pool, its pressure

(

would have to be lowered to atmospheric pressure at some point, and about 3.5% would flash to steam.

This flashing might occur inside the piping

• If the same flow of feed were initiated but without bleed flow, the pool temperature increase woald be prevented but the water level of the pool would increase and become excessively high af ter several hours.

tAs noted in Sect. A.5.2, the RHRSW pump discharge can also be di-rected into the containment spray headers.

It is assumed here that this path is not available.

+v.ae,

,r,v.+

e--no,-m

,,,,,-..,-,-,n.~.-

-, - - - -, - - - -, - -, ~, -..., -,,,,.,

,,-w,,-,,,---w-~.,,-.

w---w--,v.n,,---,---..---.,n.

38 system used for the bleeding operation.

If so, the steam might contact cool water residing in the piping before initiation of the bleed, result-ing in violent condensation and possibly pipe rupture. The postulated rupture would then cause a flooding and steam environment hazard, with the seriousness of the hazard depending on break size and location (the re-actor zone would be the worst location).

In any feed-and-bleed cooling technique, the sheer volume of bleed created would cause difficulty.

How ev er, the pool heatup is slow and there should be sufficient time during a Loss of DER accident sequence to develop a practical strategy. The ideal receptors for the bleed would be the two 1893 as (500,000 gal) pressure suppression pool water storage ta nks.

If the pool water cannot be transf erred to these tanks, then it might be moved into the Unit 2 pressure suppression pool or some flooding might be acceptable if it could be I.inited to a non-critical location.

The risks of feed-and-bleed pool cooling would have to be weighed against the risk of letting the Loss of DER accident sequence proceed to primary containment f ailure by overpressurization, which is discussed in Chap. 7.

Reference for Section 4 4.1 "Energency Procedures Guideline s BWR 1 through 6," NEDO-24934, Revi-sion 2, Prepublication draf t, May, 1982.

e j

9 4

e s---

a-r-

e r---,

-,-,,r

,eme,,- m w

,,,,,w-,

r --

e

1 39 Table 4.1.

Minimum feed flow required in a J

feed-and-bleed cooling schema to prevent suppression pool temperature increase during Loss of DER accident Time elapsed Pool temperature"

'I",'$eh*

since event reached without initiation any cooling Im,/s (gPa))

(h)

[*C (*F)]

2 61 (141) 0.32 (5000) 3 73 (163) 0.20 (3130) 5 87 (189) 0.13 (2020) 8 102 (216) 0.09 (1400) 12 119 (247) 0.06 (1010) 16 133 (272) 0.05 (810)

1. Taken f rom the calculation reported in Sect.

3.3.3.

bMinimum required to prevent f urther increase in pool temperature. No flow until corresponding elapsed time, then continuous flow thereaf ter at the indicated rate.

1. If suppression pool water level is to be main-tained constant, the volumetric bleed flow would have to exceed by 10-20% the f eed flow. Bleed flow would include both the condensed SRV discharge as well as the f eed flow, dCalculated by simple heat balance, assuming 32*C (90*F) feed, and decay heat steam production given by 1979 ANS Standard Decay Heat (with Acti-nides, and without 1.1 conservatism f actor).

e m--,

,..._....-..,,_~._,-w

40 ORNL-0WG 83 4234 ETD S,

e-

?-

C U S-

~

u u

$ l$

eaJ$

5E e

Id u

o e_ d o ~ O o

AT TIME 600 MIN. TEMPER ATURE 5

PEAKS AT 82 C(179 F) 0

=38- $ g.,

S~

RHR HEAT E

E EXCHANGER L-M STARTED a

ve m

3-8~

TEMPERATURE INCREASE MOST RAPID DURING REACTOR VESSEL DEPRESSURIZATION E'

6'o 150150 240 350 360 420 450 550 650 650 750 750 840 9dC 950 d20108011'40120012' 0122013' 01440 O

G 8

TIME IMIN)

Fig. 4.1.

Suppression pool temperature with one RHR heat exchanger in operation.

D e

e 4p 4

29 G

8 DN LO 3

BI S

T LN A

R C

OI w

TT O

CN AE EV

.M R

k O

a

~ - I R

S F

E LE I7 0

R 1

1 2

2

\\

4 4

4 I

ie 6

s M

4 E

2 T

S 8

c Y

7 8

s S

m

~

N,G

+

. O ""

4 N

8 M

e E

t il i

T s

DS R

9 AY y

1 E

1 CS

]

s I

I gs.

P 4

1T 6

t 4

e g

I le s

n i

4 a

o t

8

_ 4-h7 0

O 2

r e

x Wd 6

9 n

1 4

1 1

8 f-i

[

6 6

'3,7 6

M ir

)

7 7

~i D E G

n A

I N

M TS E

a C Y aTI T

S NR S

v

)[

E Y

n A

N 8

I S

o i

5 t

4 2

8 a

x l

i e

tnev t

n e

e mn i

a tn 6

o 2

c

-X yram ir

)

P 4

,J 2

3 m

1 3

4 6

4 4

6 6

3'

[k 4TI 2

) J 0

q j 4

x m '2

(,N i

9 4

2g I

I 2

m"8 g

3 1T M31 F

R 463 I

S O

OT S

TGS R

PM O

m'B C

4 m"8 01=

3 lb 3

1T 461I 4

e N

RG OI T ON TS T I

, AU S

D O2 T

C LA AI T H T

G L

I S

EU Nx RS E

W i

I!i1

-1ij

}

)ii!i'li!

ii

{!

1kil:

jIJ!i

)!

42 l

5.

'IRANSIENTS WITH LOSS OF DER: CALCULATIONS PERFORMED FOR 'IHE CASE OF 'IEERNAL STRATIFICATION IN 'IHE SUPPRESSION POOL 5.1 Introduction The accident sequence considered in this section is exactly the same as that considered in Sect. 3, except that the assumption of uniform pool temperature is not made. The uniform pool temperature assumption is equivalent to the assumption that the RHR pumps are operational (even though the pool cooling is not) and that they are used to circulate and mix the pool.

Without f orced circulation of the pool, the average water temperature in the locality of a discharging T quencher (" local" temperature) will be higher than the average temperature of the whole pool (" bulk pool" ten-per ature ). In addition, there will be thermal stratification throughout the pool (i.e., the temperature of the water at the surf ace of the pool will be higher than the bulk pool temperature). The Monticello T quencher tests

.1 showed that local temperature exceeds bulk pool temperature by as s

much as 24*C (43*F) during extended SRV discharge, and that there is con-siderable thermal stratification throughout the pool.

During a loss of DER accident sequence the higher surf ace and local temperatures will increase the rate of escape of steam from the pool to the wetwell atmosphere, and thereby decrease the time required for pres-sure to build to the point of primary containment f ailure. To study these effects, a detailed thermal-hydraulic model of the suppression pool was developed (see Appendix C for a description) and used to calculate the spatial temperature distribution of the pool during the Loss of DHR accident seque nce. _ This model was utilized independently for the calcu-lations reported in this chapter.

The transient input to the pool model j

consisted of the SRV flow, the total wetwell pressure, and the evaporation rate from the pool surface throughout the sequence.

This input was taken from the BWR-LACP calculations of Chap. 3.

5.2 Summary When there is no forced circulation of the suppression pool, an un-mitigated loss of DHR could lead to primary containment f ailure by over-pressurization as early as 28 h af ter reactor shutdown. This is seven I_

hours sooner than for the case with uniform suppression pool temperature.

l 5.3 Detailed Results l

l The purpose of this section is to provide an estimate of how much sooner the containment would f ail in a Loss of DHR sequence if the RHR system is not circulating the pool so that the assumption of uniform pool temperature is not valid.

The subsections that follow specify the input assumptions for the pool temperature distribution calculation, present the

43 calculated temperature distribution, and evaluate the effect of the ten-perature distribution on the time of containment f ailure.

5.5.1 Innut and assuantions j

i The thermal-hydraulic model described in Appendix C was used to cal-culate the temperature distribution of the pool during the first 15 h of the 35-h-long base-case

• Loss of DER sequence (described in detail in Sec-tion 3.3). A longer calculation was not attempted because of numerical difficulty associated with the very low sustained SRV discharge rate reached after 15 h.

In addition, the model is not presently programmed to calculate the condensation and transport of T quencher discharge during the operating mode reached af ter about 24 h, in which the reactor vessel has repressurized and a single SRV is discharging intermittently, but at a high flow rate, into a pressure suppression pool which is close to satura-tion.

Nevertheless, the 15 h of available calculational results provide considerable insight into pool behavior yithout forced circulation, and allow an estimate to be made of the effe:t on containment f ailure time.

The special suppression pool model was run independently for the 15 h calculation period.

The following information was input as a function of time: rate of discharge of steam from the SRVs to the pool (Fig. 3.5),

total pressure in the wetwell (essentially equal to the drywell pressure shown in Fig. 3.8), and the rate of evaporative steaming from the surf ace of the suppression pool (not shown - is equal to about seven percent of the total decay heat steaming rate at the 15 h point).

It was not nece s-sary to specify to the pool model which of the 13 SRVs were actuating because the model assumes that all the SRV discharge occurs through the same T quencher at a constant, fixed location in the pool throughout the calculation. This assumption introduces an element of conservatism into the results between 0 and 24 h because during the first ~24 h of the Loss of DER sequence, the operators are able to rotate their selection of SRVs around the pool as specified by the emergency operating procedures. After 24 h, the remote-manual SRV actuation capability is lost, and the SRV discharge af ter this time would occur by automatic actuation of the sin-gle, lowest-set SRV in the group of four SRVs that have an individually-set actuation pressure of 7.72 MPa (1105 psig). The discharge after this time would therefore enter the pressure suppression pool from the same T quencher.

I 5.3.2 Transient cool temocrature distribution The special suppression pool model divides the toroidal pool into 16 equal angular sognents which correspond to the 16 bays of the vetwell

• Chapter 3 considers two Loss of DER sequences: one with and one j.

without operator restart of the drywell coolers af ter their automatic trip on high drywell pressure early in the sequence.

The sequence without operator restart of the coolers is the base case - see Section 3.3.4 for more discussion of this point.

44 i

torus. The water in each of the 16 bays is further divided into four vertical regions at initially equal depth increments.

A temperature is calculated for each of the 64 regions. The bulk pool temperature is the mass-weighted average of all 64 pool temperatures.

The surf ace tempera-ture is the average temperature of the 16 surf ace nodes. Local tempera-l ture applies to the region between 1.14 m (3.75 f t) and 2.13 m (7.0 f t) in

'the bay of SRV T quencher discharge [the centerline of the T quencher is 1.52 m (5 ft) from the bottom of the normally ~4.6 m (15 ft) deep pool].

The local temperature is a weighted average of both the very hot water within the plume of condensed SRV discharge as well as cooler water flow-ing toward the T quencher before it makes contact with the discharging steam.

Figure 5.1 presents the bulk pool temperature, the average surf ace temperature, and the average temperature of the bottom regions of the pool.

The difference between surf ace and bulk temperatures reaches a maximum of 13*C (23'F) af ter 4 h, and then declines steadily, reaching 5'C (9'F) at the end of the 15 h calculation period.

This behavior is ex-pected since the rate of SRV discharge is also declining throughout the period after 4 h (i.e.,

the depressurization of the reactor vessel is cam-plete, and decay heat is decreasing).

Local pool temperature is plotted in Fig. 5.2.

For comparison pur-poses, bulk pool temperature and average surf ace temperature are also shown in Fig. 5.2.

The difference between bulk and local temperature is 10*C (18'F) a t 2.75 h, and declines steadily thereafter.

This shows that natural circulation is effectively distributing the decay heat energy from the bay of discharge into the other 15 bays around the circanference of the suppression pool.

5.3.3 Ef fect of cool temperature distribution on crimary containment oressure build-un The excessive pressure buildup in the Loss of DHR sequence is due to the escape of steam from the suppression pool, which can occur either by evaporative steaming f rom the heated surf ace of the pool or by failure of the pool water to condense all of the SRV T quencher discharge.

In the uniform pool temperature results of Chapter 3, evaporation was the primary means for the buildup of primary containment pressure.

The thermal stratification which occurs when there is no circulation of the pool would eccelerate the evaporative steaming from the pool surf ace.

However, this effect alone could hasten primary containment f ailure by only about two 4

hours, since the average surface temperature (Fig. 5.1) is only 5'C (9'F) i above bulk pool temperature (and decreasing) after 15 h.

Complete condensation of the T queacher discharge can only occur if the temperature of the water surrounding the T quencher is sufficiently below the saturation temperature.

NUREG-0783 (Ref. 5.2) specifian that a minimum subcooling* of 118C (20*F) should exist to ensure stable condensa-s [42 tion of SRV T quencher discharge at rates not exceeding 205 kg/m8

• 1he subcooling is defined as the dif ference between local tempera-ture (see Sect. 5.3.2) and saturation temperature at the T quencher depth, and thus is increased by the increase in local saturation temperature pro-vided by the ~34.5 kPa (5 psid) static overpressure of the water above the T quencher.

.~-

. ~. - -...

45 lb/(ft8 s)], and that local temperature should not exceed 93*C (200*F) for SRV T quencher discharge rates exceeding 460 kg/m8 s [94 lb/(ft8 s)].

These limits reflect the current extent of experimental determination of the conditions necessary for stable condensation.

If local temperature exceeds either of these limits, there will not necessarily be unstable or i

incomplete condensation; on the other hand, stable condensation is assured as long is these limits are not exceeded.

For the Loss of DHR accident sequence, local subcooling (Fig. 5.3) starts at over 55'C (100*F) and has decreased to about 17'C (30*F) by the end of the 15 h calculation period.

Subcooling would continue to slowly decrease af ter 15 h, and might be slightly below the 11*C (20'F) minimum subcooling requirement af ter 23.5 h, when the period of sustained, co n-tinuous SRV discharge ends due to the loss of remote-manual control of the SRVs;* however, it is likely that the pool would continue to completely condense the T quencher discharge throughout the first 23.5 h of the Loss of DER sequence.

Af ter 23.5 h, the drywell pressure is too high to permit remote-1 manual SRV actuation and flow through the SRVs ceases for several hours while the reactor vessel repressurizes. When a reactor vessel pressure equal to the opening setpoint of the lowest-set SRV is reached (at 27.7 h), T quencher discharge resumes but at a high rate of flow and in inter-mittent bursts instead of the essentially continuous low-flow discharge of the period before 23.5 h.

With T quencher discharge at high flow into an uncirculated and nearly saturated suppression pool, it is likely that the local subcooling would be well below 11*C (20*F) and might be lost en-tirely, allowing direct bubble-through of steam into the wetwell atmos-phere. Without any condensation of SRV discharge, it would take about 20 min. to pressurize the primary containment from its 0.61 MPa (74 psig) 4 pressure at 27.7 h to the 0.91 MPa (117 psig) primary containment f ailure pressure. Therefore, as a worst case, the containment f ailure would occur af ter 28 h instead of af ter 35 h.

i References for Section 5 5.1 B. J. Pa tter son, " Mark I Containment Program - Monticello T-Quencher Thermal Mixing Test Final Report," GE NEDO-24542 Class I, August 1979.

5.2 T. M. Su et al., "Suppre ssion Pool Temperature Limits for BWR Con-tainments," USNRC Report No. NUREG-0783, November 1981.

I

• As diccussed in Sect. 3.3.2, control air pressure must be at leart 0.17 MPa (25 psi) higher than drywell pressure to permit remote manual actuation of the SRVs.

1 n n,

--m-

,__,-,.--.--nn._,..---,,-,..nn,-__-r-,n..n

.,_.,,--.,,-.w

46 ORNL-DWG 83-4287 ETD 300 14 0 -

SRV DISCHARGE 280-INTER- *

- CONTINUOUS 13 0 -

MITTENT f

260-12 0 -

240-11 0 -

DEPRESSURIZATION OF RE AOTOR VESSEL 220-m COMPLETE G 100-b w

n w

m 200-T 90-Dk ti e

g w 18 0 -

80-a y

Q.

2 w

N 70-16 0 -

60-14 0 -

50-12 0 -

1 40-TEMPERATURES

,00-O = SURFACE I

V = WHOLE POOL A = BOTTOM 30-80 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 TlME (H)

Fig. 5.1.

Unmitigated Loss of DHR - suppression pool average tem-peratures.

e w-r v

47 ORNL-DWG 83-4288 ETD 300 l

SRV DISCHARGE INTER-

• 280-

• CONTINUOUS MITTENT 13 0 -

260-12 0 -

240-11 0 -

DEPRESSURIZATION OF REACTOR VESSEL 220' COMPLETE n

G 100-b w

u w

m 200-T 90-Dk W

x x

w 18 0 -

80-a y

Q 2

5 w

s N

70-16 0 -

d, t

60-14 0 -

l.

l d

~

50-I LOCAL TEMPE R ATURE INCREASES 12 0 -

i DURING SRV DISCHARGE, DE-CREASES WHEN FLOW STOPPED 40-r POOL TEMPERATURES 10 0 -

o = AVG. SURFACE I I v = BULK POOL a = LOCAL 30-80 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 TIME (H)

Fig. 5.2.

Unmitigated Loss of DHR - suppressson pon1 average tem-l peratures and local temperature in the Bay of SRV dischstge.

s e

48 ORNL-DWG 83-4283 ETD 11 0 -

200 SRV DISCHARGE INTER- -

, CONTINUOUS MITl ENT 10 0 -

180 _

90-16 0 -

80-

{

140 -

6 C

/

v 70-

$ 120-7 a

a O

60-O I

DEPRESSURIZATION OF REACTOR O

O 10 0 -

VESSE L COMPLETE m

m D

50-D cn to

-J

-J 80- ""

O 40-O O

O J

J 60-30-40-LOCAL SUBCOOLING DECRE ASES 20-WHEN SRV DISCHARGING,IN-CREASES WHEN DISCHARGE STOPPED 20-10 -

O-0 i

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 TIME (H)

Fig. 5.3.

Unmitigated Loss of DHR - Local Subcooling in the Bay of SRV discharge.

e h

n

'= -

i 49 6.

LOSS OF DHR WITH STUCK-OPEN RELIEF VALVE 6.1 Introduction This chapter considers the effect of a stuck open relief valve (SORV) on the Loss of DER accident sequence. The assumption of a SORV is in addition to the assumption of all of the other f ailures that must take place

• in order to cause a loss of DER (see Sect. 3.1). During the early part of the Loss of DER accident, numerous actuations of the SRVs are re-quired in order to control reactor vessel pressure.

For the analysis of this chapter, it is assumed that a single SRV, af ter opening when called upon, fails to close, either automatically [when the reactor vessel pres-sure has decreased to below the 7.38 MPa (1055 psis) closing pressure of the lowest set group of SRVs] or in response to operator manipulation of the remote-manual SRV controls in the main control room.

The calculations for this section were performed using the BWR-LACP code. As in Chap. 3, a uniform suppression pool temperature was assumed for one set of calculations, since the RHR pumps might be available to circulate the pool water and thereby minimize thermal stratification and the existance of local temperatures in excess of the bulk pool tempera-ture. Additional calculations were performed using the special suppres-sion pool model (used in Chap. 5 and described in Appendix C) which is able to predict local temperature and thermal stratification in the event that there is no forced pool circulation.

l 6.2 Summary and Conclusions For the case of suppression pool circulation and uniform pool ten-per a t ur e, an unmitigated Loss of DER sequence with a SORV would lead to eventual primary containment f ailure by overpressurization at 34 h.

This is close to the 35 h f ailure time estimated in Chap. 3 for the non-SORY case.t If a loss of vessel water injection were to result from contain-ment f ailure, the reactor vessel would be at low pressure at the time of core uncovery. A pressurized boil-off was predicted (Chap. 3) for the no n-SORV ca se.

A SORV does not have a great impact on overall system behavior during a Loss of DER sequence. The continuously open SRV depressurizes the re-actor vessel sooner and more rapidly than would the controlled depressur-ization [55'C/h (100*F/h)] that the operators are required by procedure to initiate when suppression pool temperature exceeds 49'C (120'F). As a I

• It is possible that the reactor shutdown following the SORY could be the Loss of DHR accident initiator.

tPrimary containment f ailure occurs earlier for the case with an SORV because repressurization of the reactor vessel does not occur. As dis-I cussed in Chap. 3, suppression pool level and temperature remain constant during the lengthy period of reactor vessel repressurization, thereby de-laying the f urther increase of primary containment pressure.

50 result, the pool heats up f aster during the first several hours. How ev er, the pool heatup is soon limited by the rate of decay heat steam production in the reactor vessel. Af ter 3 h, the pool temperature (Fig. 6.5) has reached 73*C (163*F), and this is no higher than the case without the SORV (Fig. 3.6).

The slightly lower reactor vessel pressure does not cause the turbine-

.ccident driven RCIC system to become unavailable at any time during the The RCIC system actuates intermittently during the first 6 h, sequence.

but the vessel pressure remains well above the 0.45 MPa (50 psig) setpoint f or isolation of the RCIC turbine steam supply line although the vessel pressure has decreased to below the 0.79 MPa (100 psig) setpoint for isolation of the HPCI turbine steam line at about the 6 h point. After 6 h, the CRD hydraulic system pumps provide all required vessel water inj ection.

• 6.3 Detailed Results In Results of the BWR-LACP calculations are shown in Figs. 6.1-6.9.

these results are very similar to those for the Loss of DHR most cases, sequence with no SORY (see Chap. 3) in which the reactor vessel is depres-surized by the operator early in the sequence. Therefore, the discussion below is limited to major points which are unique to the case of depres-surization by a SORV. A more complete de scription of the physical basis for, and operator actions which can af fect, the behavior of each system variable can be found in the corresponding paragraphs of Sect. 3.3.

6.3.1 Reactor vessel water level Figure 6.1 shows water level in the downcomer region of the reactor and vessel.1 Figure 6.2 shows the total rate of vessel inj ection flow, the total amount of water injected is shown in Fig. 6.3.

The results are very similar to those shown in Figs. 3.1, 3.2, and 3.3 for the case with-out a SORV. The vessel level is maintained in an acceptable range throughout the 34 h period prior to containment f ailure.

During the first 6 h, inj ection is provided by the CRD hydraulic system and the RCIC system

~

in combination. There is a single cycle of HPCI operation, which the operators initiate during the first minutes in order to rapidly restore vessel level to the normal operating range.

Af ter 6 h the inj ection re-alone.

If the RCIC quirements are fully met by the CRD hydraulic system,

• With an SORV, more coolant is lost from the reactor vessel during the early portion of the accident sequence and must be replaced by the operating inj ection systems. Without the SORY (but with an operator-controlled depressurization), all required reactor vessel inj ection can be as discussed in Chap. 3.

supplied by the CRD hydraulic system af ter 4 h, TThis is the water level indicated on the control room instruments.

To read the full range of level variation shown on Fig. 6.1, operators would have to consult the wide range instruments.

51

.ystem had been unavailable af ter 4 h, reliance solely on the CRD hydran-lic system inj ection would not have resulted in an unacceptably low vessel i

level. The single RCIC actuation between 4 and 6 h (Fig.

6.2) is based on the operators desire to maintain vessel water level as close as possi-ble to the normal value of 14.25 m (561 in.) above vessel zero.

6.3.2 Reactor vessel oressure Figure 6.4 shows reactor vessel pressure following the SORV with loss of DHR.

Pressure f alls very rapidly at first, due to the combined effect of the SORY and of the HPCI system which draws steam from the ves-l sol to run its turbine, and inj ects a large quantity of cold water into the vessel.

Af ter the HPCI system is shutdown, the pressure f alls less rapidly.

As pressure decreases, the mass flow of steam (at sonic velocity) through the fully open SORY also decreases.

The rate of depressurization becomes slower and slower, until pressure reaches a minimum of 0.69 MPa (85 psig) af ter about 6 h.

Pressure then increases slightly, le: els off, and begins to decrease very slowly until 19 h af ter accident initiation.

By this time the pressure in the primary containment including the wetwell, which receives the SRV discharge, has increased to the point at which the down-stream pressure is not low enough to maintain critical flow through the SORY. Therefore, throughout the remainder of the sequence, SRV flow ca-l pacity is impaired and the vessel pressure slowly increases to about 1.03 MPa (135 psis) at the estimated 34 h primary containment failure time.

6.3.3

_Sanoression nool tennerature and water level i

Suppression pool temperature (Fig. 6.5) and water level (Fig. 6.6) increase steadily throughout the SORY accident sequence because of the continuous discharge of steam through the fully open SRV, without any pool cooling. The results shown in Fiss 6.5 and 6.6. were calculated assuming a uniform pool temperature.

The RHR pumps might not be available to circulate the pool water to minimize local temperature and thermal stratification. Accordingly, the 4

l transient pool temperature distribution throughout the pool was also calculated using the detailed thermal-hydraulic suppression pool model described in Appendix C.

This model used as input the BWR-LACP results for SORY discharge flow and suppression pool pressure throughout the 34 h long sequence. The results (not shown) indicate that as expected, the temperature is higher in the bay of SRV discharge and that the pool tends to stratify, with the hotter fluid on top.

How ev er, due to the relatively l

low release rate of steam into the suppression pool during the early part of the accident se que nce, the local temperature differences are man 11.

At the end of 10 h, the average pool surf ace temperature is only about 3'C (5'F) above the whole pool bulk temperature, and the temperature of the water in the locality of the discharging T quencher is approximately equal to the whole pool bulk temperature. The effect of the higher surface ten-perature on primary containment pressure is assessed below.

i

52 6.3.4 Primary containment oressure and temperature Drywell pressure and atmosphere temperature are shown in Figs. 6.7 and 6.8 for the case with uniform pool temperature (as if the RHR pumps were circulating th3 pool water, but without pool cooling). As discussed in Sect. 3.3.4, the drywell coolers might trip af ter about I h, leading to an increase in drywell temperature (Fig. 6.8). As pool temperature l

increases, steam escapes from the surf ace of the pool, pressurizing the primary containment, eventually leading to drywell failure af ter 34 h.

If there were no forced pool circulation, higher surf ace temperatures would result from the thermal stratification effect and there would be a more rapid loss of steam to the primary containment atmosphere. With the 3*C (5'F) higher surf ace temperature reported in Sect. 6.3.3, the drywell failure pressure would be exceeded et 32 h instead of at 34 h.

This is 4 h later than for the corresponding case without the - SORV.

6.3.5 ECCS onmo not oositive suction head (NPSH)

Figure 6.9 shows the NPSH that would be available if a single RHR pump were operated (taking suction on the pool and discharging back to the pool) throughout the sequence. The NPSH at the pump suction is below the manuf acturers recommended minimum NPSH during the final one-third of the sequence. As discussed in Sect.

3.3.5, it is possible to operate below the recommended minimum; in addition, the available NPSH at the pump suc-tion can be increased by throttling the pump discharge to operate the pump at a lower flow. Therefore, RHR pump operation could be maintained (or begun) throughout the sequence.

I

53 t

ORNL-DWG 83-4235 ETD R-s-

~

S' 8_

s S-CRD INJECTION e.

REDUCED g-s s

d: L CRD INJECTION a

d RESTORED eg a

J d "'

INJECTION MODE m

a

=

=

0 d @-

RCIC CRD d

AND HYDRAULIC CRD SYSTEM S-h-

HYDRAULIC ONLY TOP OF ACTIVE FUEL SYSTEM

... - - - - - - - - -..... - - -. - - -. - - - - - - - - - - - - - - - - - - - - - - ~ ~ ~ - - - ~ ~ ~ ~ - - ~ ~ ~ - - ~ - - ~ ~ ~ ~ " " " - - " "

r, H-M-

BOTTOM OF ACTIVE FUEL 0 3 is s 6 do 6 s s s s icbo inbo do 13bo isbo isoo is'ao i7bo re'ao isbo acbo 2:00 TIME (MINI Fig. 6.1.

Unmittigated Loss of DHR with SORY - reactor vessel level.

l l

e

54 ORNL-DWG 83-4236 ETD M-K-

g.

A-R-

8-G M-g.

A g-E' I

05 5 n-r

.--SINGLE HPCI INJECTION CYCLE 0,R-y TERMINATED BY OPERATORS 52-a h e-d 8

h[- S.

~

g u

2 OPERATOR ADJUSTMENTS TO RCIC FLOW f

!~

USE OF RCIC m-l DISCONTINUED t,l~ PERIODIC ADJUSTMENTS TO CRD HYDR AULIC SYSTEM FLOW ~

n-i o-o 0 100 200 300 400 500 60" 700 800 900 1000 1100 1200 1300 1400 1500 1600 1700 1800 1900 2000 2100 TIME (MIN)

Fig. 6.2.

Unmitigated Loss of DHR with SORY - total rate of inj ec-tion flow.

ORNL-DWG 83-4237 ETD g,

~

l_

g_

~

l_

INITIAL VOLUME OF WATER IN CONDENSATE STORAGE TANK

~

3_

7 5

$ $~ h_

w 3

3 g_ 9 9-0 -

8 t3 @- Ng-H Eq 5

5 @-E@

_i Es

~

E B-l_

~

WATER USAGE MOST RAPID DURING THE EARLY PART OF THE DEPRESSURIZATION o-o 0 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600 1700 1800 1900 2000 2100 TIME (MIN)

Fig. 6.3.

Unmitigated Loss of DHR with SORY - total volume inj ected into reactor vessel.

I l

I

56 oRNL-DWG 83-4238 ETD e-s-

$~

E.- s a-E M

E M

E~

S *_ M COMBINED EFFECT OF SORV AND E

h g~

HPCI OPE RATION CAUSES RAPID m

3 DECREASE IN PRESSURE E

r S m-5-

r M

5

$ g-

• PRESSURE DECREASE LESS RAPID AFTER HPCI TRIPS SLOW REPRESSURIZATION BEGINS (CAUSED BY INCREASING Q-CONTAINMENT BACK PRESSURE)

~

m g_

STEAM VENTED CONTINUOUSLY THROUGH 1 SRV 0 150 250 3b0 4b0 5b0 6b0 7$0 8b0 9b0 10bD11b012b013b0l4bo15'0016b01/0018h019b020b02100 TIMC (MIN)

Fig. 6.4.

Unmitigated Loss of DER with SORY - reactor vessel pres-sure.

-n,-

,e

• t/

g//

3 1.0

'd E n D E EliEE i,i i'

EM

~

1.8 1.25 1.4 1.6 4

150mm 6"

4

+?h+

f I<$;f3 r

57 e

ORNL-DWG 83-4239 ETD 8,

8-E-

?~

C G 8.

~

u e R-ue s

5 E

c a. u bi~ e e

fd u

o o) $ ~

H o_

id~ 2 oa z2 z

28- $ g.

8 M~

u a.

E e

e R-

• m is 8-f R~

8

. 200 300 460 500 s$a 760 e60 960 10'0011bu12bD13b014'0015b016'001/0018b019b020b02100 0 100 TIME (MIN)

Fig. 6.5.

Unmitigated Loss of DER with SORY - suppression pool ten-perature.

w l

t I

i f

I

58 ORNL-DWG 83 4?40 ETD 8

e-l2 -

2-3-

5:

S E-O x

G d e- $ M-d

~

_a e- _2 J

o u-i, E

x - gn o

oa-

"g m-m z

E'. + -S, a

v e

5,,. 02 m

a:

h v>

~-

D-o 7-2 8

N 13bo is is 34 gg,g

=

UImitigated Loss of DHR with SORY - suppression Pool water level e

d

=

m-

59 5

i ORNL-DWG 83-4241 ETD S

E-p.

DRYWELL FAf LURE PRESSURE h-E s_ g-E Ui o_ g _

S *- u 0

5 y g.

DRYWELL DESIGN PRESSURE y e-y u

o.

a 8 g,. d-a r

N U

o-m g n_

9-e, -

~-

~

( DRYWELL COOLERS TRIPPEG ON HIGH DRYWELL PRESSURE (AND NOT RESTORED BY OPERATORS) o-o 0 100 200 300 400 500 600 700 800 900 1000 1100 1200 130J 1400 1500 1600 1700 1800 1900 2000 2100 i

TIME (MIN) l Fig. 6.7.

Unstitigated Loss of DER with SORY - drywell pressure.

i e

W l

60 J

ORNL-DWG 83-4242 ETO 8

w k'

S.

S E8 B.

e u'

u

?

x s

d c

c

$ $~

i 0-1 El r5

~

c:

U 9-u g' (DRYWELL DESIGN TEMPERATURE

$~ $

281 F)

E, a,

it-v9

.o h

h f."

~

NeA

- e s.

+

g

- g, r

"-J DRYWELL COOLERS TRIP ON p

HIGH DRYWELL PRESSURE P

1

=.

. 300 4.T 500 600 700 8C0 900 1000 1100 1200 1300 1400 1500 1600 1^00 1900 1300 2000 2100 0

100 200 TI'1E IMINI Fig. 6.8.

Unmitigated Loss of DBR with SORY - drywell temperature.

I e

A

...,,.x._..

,_m._

61 4

e ORNL-DWG 83-4243 ETD E

E, I

S-

^

Il 2d, #~

DRYWELL p ![~

COOLERS TRIPPED 8245 S

'hg.

E.a =E MANUFACTURERS a'

RECOMMENDED E

,F MINIMUM NPSH w..x 2

y-2

'1 i

O*

O i

r 1

5 i

t 5

i O 100 2JO 300 4C 500 600 700 800 900 1@ 11;;0 IE'O 13001431500160010019001900 2'M' 2100 TI":. (Mim Fig. 6.9.

Unmitigated Loss of DHR with SORY - RHR[nsp net positive suction head.

5 e

e

,,-,--,..r.

62 7.

STATIC OVERPRESSURIZATION CONTAINMENT FAILURE MEGANISMS AND PHENOMENA 7.1 Introduction The loss of decay heat removal accident sequence s de scribed in Chap-ters 3 through 6 produce a gradual heatup and pressurization of the pri-mary containment system.

If the accident is allowed to progress indefi-nitely, this pressurization will eventually result in loss of containment integrity, i.e. containment f ailure.

The purpose of this chapter is to briefly review static overpressurization containment f ailure mechanisms and the phenomena induced in a NARK I containment system by such a fail-ure.

Sections 7.2 and 7.3 present a brief review of the Browns Ferry MARK I containment system de sign.

Section 7.4 discusses static over pressure containment failure mechanisms, and Sect. 7.5 de scribe s thermodynamic and physical phenomena which might be expected to occur following containment failure.

Section 7.6 is a discussion of the long term core cooling re-

~

quirements and system capabilities of the Browns Ferry Unit i reactor dur-ing the Loss of DHR accident.

An evaluation of the possible impacts of post-containment f ailure phenomena on reactor vessel inj ection system cap-4 abilities is presented in Sect. 7.7.

Section 7.8 summarizes the phenome-nological discussions presented in this chapter and Sect. 7.9 de scribe s their significance in light of current Loss of DHR accident probabilistic risk assesmaent practice s.

l

~

7.2 BFNP Primary Containment Structural Desian*

s The Browns Ferry Reactors employ a NARK I pressure suppression con-tainment system which houses the reactor vessel and coolant recirculation loops.

The design consists of a drywell, constructed in the shape of an inverted light bulb, a toroidal pressure suppression chamber, which nor-s (one million gallons) of water, and a mally contains approximately 3785 m connecting vent system between the drywell and the pressure suppression pool (Fig. 7.1).

Pertinent primary containment design parameters are given in Table 7.1.

The drywell is a steel pressure vessel with a spherical lower portion 19.8 m (65 f t) in diameter and a cylindrical upper portion 11.7 m (38 f t, 6 in.) in diameter.

The overall height of the drywell is ~35 m (115 f t).

The drywell is designed for an internal pressure of 0.478 MPa (56 psig) coincident with a temperature of 411.5 K (281*F), plus the dead, liv e, and seismic loads imposed on the shell.

The thickness of the drywell wall varies f rom a minimum of 1.9 an (3/4 in. ) in the cylindrical section, to a maximum of 5.9 cm (2 5/16 in.) in the toriodal sphere / cylinder knuckle region.

The pressure suppression chamber is a steel pressure vessel of toroi-dal shape, located below and surrounding the drywell.

The centerline

• The majority of the information in this section is excerpted f rom Ref. 7.1.

I e

. ~ _,..

M 4

63 diameter of the torus is ~33.8 m (111 f t) and the cross-sectional diane-ter is 9.5 m (31 f t.).

It contains ~3823 as (135,000 cubic ft) of water at maximum pool level.

The thickness of the torus wall varies between 1.9 and 2.9 cm (3/4 and 1-1/8 in.).

The suppression chamber is designed to the same material and code requirements as the steci drywell vessel, and all attachments to the torus are by full penetration welds.

The drywell and suppression chamber are connected by a vent sy st em which conducts flow from the drywell into the suppression pool and dis-tributes this flow uniformily around the pool.

Eight circular vent pipes,

- each 2.06 m (6.75 ft) in diameter connect the drywell.to the suppression' ch ambe r.

Jet deflectors are provided in the drywell at the entrance to each vent pipe. These vents are connected to a 1.45-m (4-f t, 9-in.) di-aseter vent header of toroidal shape, which is contained within the air-space of the suppression chamber. Ninety-six downconer pipe s, each 0.61-m (24-in.) diameter, project downward into the suppression pool, terminating 1.22 m (4 ft) below the surface of the pool. Vacuum breakers discharge from the suppression chamber free space into the vent pipes to equalize the pressure between the drywell and suppression chamber.

The suppression chamber, which is located in a separate room in the reactor building base-ment, is accessable only through two normally closed 1.22-m (4-f t) diame-ter manhole entrances with double testable seals and bolted covers.

Several types of piping and electrical penetrations, as well as per-sonnel and equipment access hatches penetrate the primary containment.

The general design of the piping penetrations incorporates a penetration sleeve which passes from the reactor building, through the shield wall concrete, and proj ects into the gap region between the shield wall and the drywell liner.

Guard pipes and expansion bellows are incorporated where necessary to allow for movement and protection of process lines.

Pe r so n-nel and equipment hatches incorporate double, testable seals to ensure containment integrity.

7.3 Drywell Liner Gao Construction The BFNP drywell is surrounded by reinforced concrete for shielding purposes, but the steel drywell liner is in direct contact with this surrounding concrete only below elevation 548.79 f t (Fig. 7.2).7.8,7.8 Above this elevation, the gap between the drywell liner and the reactor building concrete is filled with a variety of materials.

Be tween eleva-tion 548.79 f t and 550.29 f t, the drywell liner is surrounded by a sand transistion zone which is designed to transmit seismic loads f rom the dry-well liner evenly to the concrete foundation. 7.4 This sand transition zone is drained,by eight, 10.2-cm (4-in. ) diame ter, sand filled pipes, which are spaced at equal intervals around the drywell. 7. s-7. 8 These i

drain pipes discharge onto the suppression chamber room floor.

Be tween l

elevation 550.29 f t and 566.0 f t the drywell liner is surrounded by a 6-cm

(

(2-3/8-in.) layer of fiberglass.

Above the 566.0-f t elevation, the dry-well liner is surrounded by a 5.7-cm (2-1/4-in. ) layer of polyester-based l

foam filler.

This foam has a maximum service temperature of 413.7 K (285'F),7.' and is designed to accomodate compression due to thermal ex-pansion of the drywell liner.

Between this foam filler and the concrete ym_,.

y-_._-

_.,___y.,

g.,..__,,

..___.,__,,___m,.,__

.,,y,,,,

h 64 is a fiber glass laminated concrete pouring form. Both the foam filler and the fiber glass form extend up to the 636.67-ft elevation.

Pressurization of the volume enclosed by the drywell liner, reactor building concrete, and drywell shield plugs is prevented by several leak-age paths which vent to various regions of the reactor building (Fig.

2 l

7.3).

A maj or gap vent path is provided by the annular spaces between embedded containment penetration sleeves and their associated penetration assemblies.

This vent area is available at all sleeves for piping pene-trations, the personnel air lock, and the two equipment access locks. v. s All embedded sleeves are a minimum of 15.2 cm (6 in.) larger in diameter th,an their associated penetrations.7.s A typical penetration sleeve con-figuration is shown in Fig. 7.4.

The annular gap betwoon the inside of the penetration sleeve and the outer surf ace of the penetration assembly extends into the drywell liner gap region, providing a direct pathway for flow from the drywell liner gap into the reactor building.

Over one hun-dred drywell penetrations of various sizes are scattered over the face of the drywell liner, affording a combined drywell liner gap vent flow area in excess of 9.29 m8 (100 ft*).

As shown in Fig. 7.3, the drywell gap can also vent to the suppres-sion chamber room via the annular gaps between the eight drywell vent pipes and their surrounding vent sleeves.

The total flow area for these 8 (100 ft8).

The suppression eight flow paths is approximately 9.29 m chamber room is connected to the HPCI,"RCIC, RHR, and core spray pump rooms via four (one for each room) open aanways (Fig. 7.5).

These ECC pump rooms are in turn connected to the remainder of the reactor building

=

via open stairwells.

The suppression chamber room also connects directly to the drywell personnel access room on the 565-f t level via the floor penetration sleeves of the RHR system shutdown supply and return lines.

The drywell personnel access room connects to the surrounding reactor building atmosphere via the access room valve operator roof openings at the 580.0-f t level.

These room openings provide a minimum flow ' area of 0.93 m2 (10 fts),s.se The annular gap between the drywell liner and the surrounding reactor building concrete shield is sealed where the removable drywell head j oins the liner (elevation 639 f t) by drywell-to-reactor-building bellows which are designed to accomodate the differential expansion be tween the drywell i

liner and the reactor building concrete during plant heatup and cooldown.

The bellows is a single piece stainless steel structure.7.11 In the event this bellows seal were breached, the liner gap would vent into the upper drywell head region below the drywell shield plugs.

These shield plugs are constructed in a three layer, six piece, circular configuration with an 11.56 m (37-f t,11-in. ) inner diameter and weigh between 67,100 and i

90,700 kg (74 and 100 tons) each.

Since the plugs are held in~ place only by their weight, they do not form a leakproof seal between the reactor building refueling floor and the drywell cavity. 7.18 I

j 7.4 Static Overoressure Containment Failure Mechanisms As previously stated, the loss of decay heat removal accident se-quences de scribed in Chapters 3 through 6 produce a slow heatup and i

a t

-e

,,m

---r.

y

--,---wv-m,-~,_.-w.-

n m-v.

m..,,

,,...-.c e

..,,,.-,,._..,.m,

,,.. ~., _, -.....

~

65 pressurization of the primary containment system, ultimately resulting in containment failure.

An early study of the maximum BWR containment cap-ability for such accidents was conducted as part of the Reactor Safety St udy (RSS). 5.18 The containment studied was that of the Peach Bottom plant, a BWR4/ MARK I reactor / containment system very similar to the Browns Ferry facility.

The design pressure of both the Peach Bottom and l

Browns Ferry containments is 0.49 MPa (56 psig). The conclusion of the i

RSS study was that a best estimate f ailure pressure for the subject con-tainment design is 1.21 1 0.17 MPa (175 1 25 psia). This pressure corre-sponds to a stress level in the base material midway between the yield and

, ultimate strength of the metal.

[If the ultimate strength of the struc-ture could be developed, the study concluded that the failure pressure a

would be 1.724 MPa (250 psia)]. The failure was predicted to occur in the upper half of the pressure suppression pool wall (Fig. 7.6, point A), al-though it was also stated that the toroidal knuckle between the drywell spherical and cylindrical sections (point B in Fig. 7.6) is a potential failure area.

No estimate of the size of the f ailure opening was made, but the f ailure was assumed to be of suf ficient size to rapidly depres-surize the containment system.

In a recent study conducted by the Ames Laboratory at Iowa State Uni-versity,v.as the f ailure pressure of the Browns Ferry containment was esti-mated to be 0.908 MPa (117 psig). Maximum circumferential member strain 4

was utilized as the f ailure criterion.

The study assumed uniform static 4

internal pressure loading, and only shell failure modes were considered.

The effects of penetrations, anchorage s, etc., were ignored.* A failure of indefinite size was predicted to occur at the toroidal knuckle inter-a face between the drywell spherical and cylindrical sections (point B, Fig.

7.6).

Neither of these two studies made any definitive statements regarding 1

the shape, size, or propagation rate of the containment failure opening.

It is probable that the f ailure would take the form of a circumferential ductile rip which would propagate at subsonic speeds 1/4 to 1/2 of the way around the drywell sphere / cylinder knuckle. 5. 24-7. as The ultimate size of the break would probably.be between 0.003 and 0.929 m8, i.e.,

greater than a f ew square inches but less than ten square feet. v.as As previously stated, neither of the two studies mentioned above in-corporated explicit treatment of containment penetrations and both studies assumed that the drywell liner is free to expand in a radial fashion.

The I

drywell liner could yield and deform significantly prior to shell failure.

The Ames researchers (Ref. 7.13) indica ted, however, that the maximum radial expansion of the drywell liner into the gap region would be lens than 3 cm (1.2 in.).

It is uncertain whether containment failure would occur due to the mechanism described in the two reports cited above, but the drywell liner can expand only 5 cm (2 in.) before contacting the sur-i rounding concrete.

It is possible that-localized strains near liner pene-trations would exceed shell strain values, perhaps causing seal and gasket leakage around intact penetration assemblies prior to shell failure.

(

• The study notes, how ev er, that construction codes require the con-tainment penetrations to be designed to more stringent standards than the liner itself.

I i

4 66 I

Failures of this type might result in slow containment depressurization transients, rather than the violent containment blowdown which might be expected to occur following a gross rupture of the drywell liner shell.

7.5 Post Containment Failure Phenomena Based on the containment f ailure pressures, locations, and sizes dis-cussed in Sects. 7.3 and 7.4, the post containment f ailure phenomena which might occur in a typical MARK I containment system will be briefly dis-cussed in this section.

For the purposes of this discussion, it has been assumed that the drywell liner f ails in the knuckle region, at a pressure

- of 0.908 MPa (117 psig), and that the pressure suppression chamber bound-ary remains intact following drywell liner f ailure.

The size of the fail-s (10 fts).

It has also been assumed are opening is assumed to be 0.929 a that at the time of drywell liner failure, the suppression pool is near its normal operating level and at the saturation temperature corresponding to 0.908 MPa (117 psig), i. e. 4498K (349'F).

Since it is probable that the drywell would begin leaking at some i

pressure below 0.908 MPa (rather than failing catastrophically at that pressure), the reader should regard the analysis and results presented in this section as a reasonable upper limit approximation to the forces in-volved in the drywell blowdown transient.

An additional conservatism is introduced by the high mass flows employed in this analysis.

These flow predictions are based on extremely simple models which do not account for the pressure losses induced by the complex flow path configurations which would be involved in an actual drywell liner rupture accident.

The immediate impact of a drywell liner failure under the conditions assumed above is a drop in drywell and suppression chamber pressure.

Since the suppression pool water is originally saturated, this pressure reduction results in flashing of the water in the drywell vent downconers and in the main body of the pressure suppression pool.

The resulting steam would enter the drywell via the eight 2.06-m (6.75-f t) diameter vent pipes, and leak from the containment via the drywell liner rupture.

The original nitrogen atmosphere of the containment would be swept out by the large drywell break flow expected in this scenario.

As will be described in Chapter 8, MARCH calculations for this accident indicate that over 453,500 kg (1,000,000 lba) of steam is generated and leaked from the con-tainment during the one hour period following containment f ailure.

Peak break flows immediately following liner rupture are pr'edicted to exceed 570 kg/s (75,000 lba/ min) or approximately 164 m8/s (347,000 fts/ min) of steam.Detailed analysis of the scenario described above is extremely dif-ficult due to uncertainties in the break size, flow topology, and flow path configurations available to material exiting the drywell break.

As previ'ously described in Section 7.3, the gap between the drywell liner and the reactor building concrete is filled with a polyester fosa material with a maximum design service temperature of 413.7 K (2858F). The drywell gas temperatures prior to liner f ailure would typically vary between 450 and 478 K (350 and 400*F).

It is reasonable to assume that the foam filler would lose structural integrity under these conditions, allowing 4

l i

. - ~..,

-,.,_e-.-,

.n..-,nn..,-_

nn-,._.n,.

n-

,=

67 the drywell leakage to flow into various reactor building floors, the re-actor building refueling floor, and the suppression chamber room via the flow paths described in Section 7.3.

The specific flow paths involved are particularly difficult to predict since ballooning of the drywell liner l

could effectively close flow paths above or below the liner knuckle re-gion. The configuration of the break opening is also uncertain.

Four possible cc.nfigurations are shown in Fig. 7.7.

The rapid flow of material through the break opening results in a thrust force, F, which acts on the drywell compartment.

This thrust force, can be calculated as?.18 F = IEJs+ (P, - P )A 1 (7.1)

where, If = break mass flow rate = 570 kg/s U

break flow velocity = Volumetric flow rate / break area

=

2 (164/0.929 = 180 m/s)

P break exhaust pressure

=

P* = Ambient pressure in drywell gap A8 = Break area = 0.929 m8 Depending upon the exact geometry of the break, the exhaust pressure, P, can assume any value between the internal drywell pressure, P'*"

e d

drywell gap pressure P.

It is probable, due to the proximity of the a

break to the surrounding vertical wall, that a stagnation pressure between P and P would develop at the break exit.

In any event, based on the d

g mass flows previously quoted, the thrust forces can be bounded by i

IEI, (570)(164)

F

=

=

= 100,000 N (22,500 lbf) 3, (1)0.929 L

l l

and IRIs F

=

+ (P - P )A, = 100,000 + (806,000) (0.929) g, d

g u

= 850,000 N (191,000 lbf) where g is a dimensional constant in English units.

This fo!ce would be directed radially inward around one-fourth to one half of the drywell knuckle perimeter.

The effects of this thrust force on the drywell structure are uncertain, but the exiting steam would impact the

68 adjacent ree"'or building wall, exerting a radial force, T,7.17 of pU *A NU, g

= 100,000 N (22,500 lbf)

T=

=

I 8

c c

on the concrete.

This force corresponds to a normal stress, S, of I

0.11 MPa (15.6 psi).

S = T/A

=

1 The shield wall concrete is designed for a minimum compressive strength of 20.7 MPa (3000 psi). v. se It is, therefore, improbable that steam jet in-pingement forces would result in yielding of the shield wall.

Further analysis is required to determine whether sustained exposure of the shield wall to such steam jet flows could induce wall failure due to ablation or i

disintegration of the concrete.

The impact of internal drywell blowdown forces and environments on drywell structures and equipment is also an area of concern. During the blowdown transiant the steam produced by suppression pool flashing flows up through each of the eight drywell vent pipes, impinging on the jet deflectors near the bottom of the drywell sphere.

The impingement force on the jet deflectors can be estimated as 1g U A, (570)(180)(0.929) 1 1

8 8 g, 8 g, A,

26.6 (8)

~

= 448 N (100 lbf) where A, = total vent pipe flow area = 26.6 m*.

It is unlikely that forces of this magnitude would result in detachment of the jet deflectors.

Degradation of drywell equipment operability due to harsh environ-mental conditions is probable after drywell failure.

The effects of equipment exposure to long term, high velocity steam flows such as those predicted to occur during the containment depressurization phase of this accident are exceedingly difficult to quantify. As will be described in Sect. 7.6, the continued operability of the primary system SRVs following containment depressurization is a topic of particular concern. Damage to SRV control air lines during the blowdown phase of the accident could re-sult in inability to regain remote manual operability of the valves foi-lowing drywell depressurization.

7.6 Lona-Term Core Coolina Reauirements and System Canabilities i

As a result of the relatively slow heatup of the pressure suppression pool, primary containment f ailure pressures are not achieved until some 30

_ _. -.,. -...., - - _,. -., - _.. - - _ _ ~. _

.._,.....~._1,._.-_

69 to 40 h af ter reactor scram.

At the time of containment failure, the re-actor vessel injection requirements are substantially reduced due to the low decay heat levels involved (less than 1% of full power).

Indeed, the analyses presented in Chapt. 3 indicates that at 34 h af ter scram a throt-tied uCED hydraulic pump flow of only 0.006 a /sec (100 gym) is sufficient s

to asintain a covered core, without any assistance from other reactor ves-sel water injection systems.

Table 7.2 is a summary of reactor vessel water injection systems which might be available for maintenance of vessel water level during emergency conditions.

The data in Table 7.2 is based on the BFNP Emer-gency Operating Instruction No. 417.18 and the BFNP FSAR.7.8' The HPCI, RCIC, core spray, RHR, and CRD hydraulic system pumps are located in the basement of the reactor building, in rooms which are ad-jacent and open to the suppression chamber room.

As noted in Table 7.2, the plant auxilliary boiler system provides a secondary steam source for the HPCI and RCIC pump turbines, although operation in this manner is pos-sible only after the installation of a piping spool piece which requires approximately four hours of labor.

However, it is clear that this capa-bility significantly increases the utility of the HPCI and RCIC systems during long term accident si tua tions.

It should be noted that the CRD hydraulic system pump automatically j

inj ect s approximately 0.0085 m8/s (170 gpm) of water into the reactor via the CRD mechanism assemblies when a reactor scram is in effect and the reactor vessel is depressurized. No operator action is necessary to initiate this flow.

Alternatively, the operator can manually realign the l

system to,inj ect 0.01 m /s (200 gpm) into the reactor vessel via a feed-s i

water line.7.21 The standby liquid control (SLC) system is a low flow, high pressure injection system which can be utilized under accident conditions to inject e

small amounts of domineralized water into the reactor vessel.

The maj or SLC system components are located in the northeast corner of the reactor building on the 639.0-ft floor level.

The condensate / condensate booster pump system is a low pressure makeup system which has the capacity to pump water through the idle tur-bine-driven feedwater pumps into the reactor vessel at vessel pressures I

up to approximately 2.9 MPa (400 psig). The major components of this I

system are located on the lower level of the turbine building, on floor levels 551 and 557 ft.

The condensate transfer pumps (located on the 565-ft level of the turbine building) are low pressure pumps, which could only be utilized following reactor vessel depressurization.

The RHR drain pumps are located in the suppression chamber room of i

the reactor building, adj acent to the RHR/LPCI pumps.

The RHR service water pumps are located in a reinforced concrete building ht the 1snd end of the river water intake channel.

Both of these systems have large pump-ing capacities, but the reactor vessel woul have to be depressurized prior to placing either system in operation.

In summary, it is clear that the profusion of BFNP emergency inj ec-tion systems and system operating modes provides significant assurance that reactor vessel injection flow would be available during a wide lL.

_ _ _ n _ _ _ _., _ __ _ _ _

70 spectrum of accident conditions. The use of these systems under abnormal conditions is prescribed by BFNP Emergency Operating Instruction No. 41.

7.7 Insact of Post Containment Failure Phenomena on Inlection Availability Having reviewed the BFNP containment system design, static overpres-i surization containment failure mechanisms and phenomena, and long term reactor core cooling requirements and capabilities, the possible impacts of containment f ailure phenomena on reactor vessel inj ection systems avail-ability will now be examined.

Following drywell liner failure, steam will begin dumping into the refueling floor, the pressure suppression chamber room, and various other reactor building floors via the flow paths discussed in Section 7.3.

As described in Section 7.3, the suppression chamber room is connected to the rooms containing the HPCI, RCIC, RRR, core spray, control rod drive hy-draulic and RHR drain pumps via open manways. The rate at which these rooms fill with steam is, of course, dependent upon the drywell rupture area and the flow paths involved.

It is reasonable to assume that the HPCI, RCIC, RHR, core spray, CRD, and RHR drain pump room atmospheres would eventually become filled with saturated steam [i.e.,100*C (212*F) and 100% relative humidity]. Due to the lower surface temperatures of the room walls and equipment, significant amounts of condensation would be expected.

Since the room coolers for these areas are not designed to function under such conditions, it is probable that the HPCI, RCIC, core spray, RHR, CRD, and RER drain systems would be rendered inoperable fol-lowing containment failure.*

A review of Table 7.2 indicates that there are four remaining systems which could inject water into the reactor vessel af ter the six systems noted above were rendered inoperable. Only one of these (the SLC system) is a high pressure system, and as described in Chap. 9, the-pumping ca-pacity of this system is insuf ficient to maintain the reactor vessel level above the top of the core at the time containment failure occurs.

If the reactor can be depressurized following containment failure, the pumping capacity of any one of the three remaining low-pressure inj ection systems listed in Table 7.2 would be suf ficient to cool the reactor core - pro-vided the containment f ailure phenomena do not result in a LOCA. This is a particularly interesting consideration, since it is dif ficult to envi-sion a mechanism by which all three of these systems would be rendered i

inoperable unless the vessel feedwater, recirculation, head spray, and core spray lines were severed as a result of containment failure phe-nomena. Following containment depressurization, remote manual operability of a single relief valve would be suf ficient to enable the operator to utilize the low pressure emergency inj ection systems discussed above.

Since containment pressure will drop following drywell liner rupture, such

)

l

• As discussed in Chapter 3, the HPCI and RCIC would be rendered in-operable before containment failure in this accident sequence because of high temperature in the suppression chamber room.

T J

h

-.--e--,-

y

-v---.

,,em.--m_-.,-.,-_%--

,mm-,,r.---v-,y.-,.. ~... -,,. _ _,,. -., _,.. -. _

-.--.~.-.-,._--.-_-,-._--s.-.--

-~-

71 remote SEV actuation should be possible (see Chap. 3) 11 the SRV actua-tors and control air systems are not damaged by the disruptive forces as-sociated with containment blowdown.

l-7.8 Loss of DHR Accident Containment Failure Phenomena - Summary I

An as r amption commonly employed in probabilistic risk assessments (PRAs) of the Loss of DHR accident is that post containment f ailure phe-namena result in the loss of all vessel inj ection capability. 7.88-'.84

. This is a particularly critical assumption for two reasons.

First, the 3

Loss of DHR sequences have commonly been held to be risk dominating se-quences in BWRs with Nark I containments, and second, the probability of inj ection loss following containment f ailure is rarely, if ever, incor-porated in the event sequence probability calculations.

Under these cir-constances, an examination of the validity and implications of this as-sumption is particulary important.

A thoughtful review of the discussions presented in Sect. 7.1 through 7.7 will reveal that there are actually six possible outcomes of the Loss of DHR accident containment failure event.

These six scenarios are listed in Table 7.3.

Scenarios 1 and 4 are not expected to result in core melting.

Scenario 1 is similar to the large-break LOCA design basis accident except that decay heat levels are significantly lower.

Scenario 4 would not result in core melting since, as previously discussed, less than 0.006 m /s (100 spm) of injection is necessary to cool the core at 8

j the time of containment failure.

The outcome of scenario 5 depends upon l

the amount of vessel injection available; this scenario is not expected to result in core molting if the reactor vessel can be depressurized and any injection system other than the SLC system is available following con-j tainment failure.

Scenarios 3 and 6 would definitely lead to core melt.

Scenario 6 is the common PRA assumption discussed above.

This scenario corresponds to a l

situation in which all of the inj ection systems in Table 7.2 are rendered inoperable due to harsh environmental conditions following containment i

rupture, or loss of high pressure injection capability due to harsh envi-i rommental conditions together with f ailure to regain remote control of any l

one SRV following drywell failure.

Since it seems unlikely that the RHR service water and condensate trains would be disabled due to environmental conditions, it appears that the probability of this scenario is dominated l

by the product of the probability of losing the HPCI, RCIC, and CRD by-draulic systems, and the probability of failure to regain remote control l

of a single SRV following containment rupture.

Scenario 3' is rarely, if ever, discussed in probabilistic risk as-sessment studies of the loss of decay heat removal capability sequence.

In this scenario, total loss of inj ection could occur due to both environ-mental conditions in the reactor building following drywell rupture, and severence of some reactor vessel injection lines during the drywell blow-down transient (i.e., a LOCA). Detailed analysis of this sequence is beyond the scope of thi.s report, how ev e r, it is clear that this scenario would result in a more severe accident than scenario 6, since (unlike 1

l

= _ _. - - -

l 72 scenario 6) the fission products released prior to reactor vessel melt-through would bypass the pressure suppression pool, directly entering the containment atmosphere.

7.9 Loss of DER Post Containment Failure Event Jesuence - Imolications In summary, it is evident that there are six possible scenarios for the loss of DHR event sequence following containment f ailure. TWo of the scenarios (1 and 4) would not lead to core molting, while two scenarios (3 and 6) would result in a severe accident.

The outcome of the two re-maining scenarios (2 and 5) is dependent on the amount of injection re-maining available.

Historically, probabilistic risk assessments of the Loss of DHR accident have ignored all except the sixth scenario described above. As discussed in Sect. 7.8, this might lead to non-conservative estimates of accident consequences since scenario 3 would involve direct release of fission products into the drywell atmosphere prior to f ailure of the reactor vessel bottom head (bypassing the suppression pool scrub-bing capability).

It appears that the probability of the traditional

" loss of inj ection following containment f ailure" PRA assanption (scenario

6) is dominated by the probability of failure of high pressure inj ection systems due to environmental concerns coupled with f ailure to regain con-trol of a single SRV.*

The total failure of all reactor vessel injection systems due to post containment f ailure environmental conditions alone seems quite unlikely due to the physical location of the systems.

The remaining mechanism by which all reactor vesel injection capabil-ity could be lost is by severance of reactor vessel injection piping during the disruptive blowdown of the drywell.

This accident sequence i

(scenario 3, Table 7.3), which could be induced by a violent.drywell de-pressurization transient following containment f ailure, has not been con- -

sidered in previous BVR probabilistic risk assessments.

Due to the sup-pression pool bypass phenomena previously described, this scenario would lead to the most severe fission product releases of the six scenarios identified in the analysis.

Future probabilistic risk assessments should consider and assign probabilities to each of the six scenarios listed in Table 7.3 rather than assume that scenario 6 is the only valid event sequence for the Loss of DHR accident after containment failure.

Finally, it does appear that the total probability of a loss of DHR induced core melt accident could be significantly reduced if the emergency operating instructions required that the operators vent the primary con-I tainment as necessary to preclude drywell f ailtre by over pressurization, thus reducing the probability of damage to the reactor vessel water in-jection lines and the SRV control air system.

• Possibly because of loss of the drywell control air system.

b 4

n--

-.s,-

,,.c a.-a e

,,.----,,---,-.--n---n--,n-

--,en

l 73 References for Chanter 1 7.1 Browns Ferry FSAR, Chap. 5.

7.2 TVA Browns Ferry Nuclear Plant Drawing #41N711-1.

7.3 - TVA Browns Ferry Nuclear Plant Drawing #41N720.

7.4 Browns Ferry FSAR, p. Q12.4-1/Q12.4-2.

7.5 Browns Ferry FSAR, p. Q12.2.2.11-1.

7.6 TVA Browns Ferry Nuclear Plant Drawing #41N720.

7.' 7 TVA Browns Ferry Nuclear Plant Drawing #47W481-11.

1 7.8 TVA Browns Ferry Nuclear Plant Drawing #47W482-4.

7.9 Browns Ferry FSAR, p. Q12.5-1/Q12.5-2.

7.10 Browns Ferry FS AR, Fig. 5.2-19.

I 7.11 Browns Ferry Nuclear Plant Hot License Training Program Manual, Vol.

4.

7.12 Reactor Saf ety Study, U.S. Nuclear Regulatory Commission, WASH-1400, 1

i 1975.

(

7.13 L. G. Greinann et al., Reliability Analysis of Steel Containment Strength, NUREG/CR-2442, June 1982.

7.14 Dick Cheverton, Private Communication, November 16, 1982.

i 7.15 L. G. Greinann, Private Communication, October 28, 1982.

7.16 M. Barrere, A. Jaumotte et al., Rocket Propulsion, Elsevier Publish-ing Co., 1960, p. 25.

7.17 Stephen Whitaker, Introduction to Fluid Mechanics, Prentice-Hall, Inc., 1968, p. 255, 7.18 Browns Ferry FSAR, Chap. 12, p. 12.2-5.

I!,

7.19 Browns Ferry Nuclear Plant Emergency Operating Instruction No. 41, November 7,1979.

7.20 Browns Ferry FSAR, Chap. 4.

7.21 W. A. Condon et al., SBLOCA Outside Containment at Browns Ferry Unit One - Accident Sequence Analysis, NUREG/CR-2672, Vol.1, ORNL/TM-8119/V1 (November 1982), Sect. E.3.

. _.. _ _ _ _ ~ _ _ - _

74 7.22 S. W. Hatch, P. Cybulskis, R. O. Wooton, Reactor Safety Study Methodlogy Applications Program: Grand Gulf #1 BWR Power Plant, NUREG/CR-1659/4, October 1981.

7.23 S. E. Mays, J. P. Poloski et al., Interim Reliability Evaluation Program: Analysis of the Browns Ferry Unit 1 Nuclear Plant, Main Report, July 1982.

7.24 Limerick Generating Station Probabilistic Risk Assessement, June l

1982.

i O

f t

l

--._w...,

-v_.,__.-

..-._.,___,__..w-veet w e w w-e-a~---w-w e-i ev w

-t-----t---rr+-

=-

-r-

-*-w--

a w w- =

ew e,__m-- m-e-=

--a~e

75

\\

Table 7.1.

Principal design parameters and characteristics of the BFNP primary containment Pressure suppression chamber Internal design pressure, psig 56 External design pressure, psig 2

.Drywell Internal design pressure, psig 56 External design pressure, psig 2

Drywell free volume, ft:

159,000 Pressure suppression chamber free volume (min.), ft8 119,000 Pressure suppression pool water volume (max.), ft:

135,000 Submergence of vent pipe below pressure suppression pool surface (normal), ft 4

Design temperature of drywell, 'F 281 Design temperature of pressure suppression chamber, 'F 281 O

O 1

-e,--n

,w n,., - - - - - -

y -,,, -. -..--<,.

.-.-n-.,.

__.~.,-_.,,_,-,.,,,,..-~n_..s.

76 Table 7.2.

Emergency reactor vessel (RPV) water injection capabilities Flow rate Shutoff head gg,,,

System Water source Injection point m /s spa MPa psid 8

HPCI.

RPV or CST Feedwater line 0.31 5000

>7.9

>1150 Auxiliary PSP boiler RCIC RPV or cst Feedwater line 0.04 600

>7.9

>1150 Auxiliary PSP bofLer a

Core spray N/A PSP Spray header 0.79 12500 2.4 342 CST RHR (LPCI)

N/A PSP Recirc loops 2.52 40000 2.3 331 Head spray GD N/A GT Via control rod 0.0103 170 10.3

~1500 drives Via Fgedwater 0.0124 200 10.3

~1500 line SLC N/A Domineralized SLC sparger 0.0035 56 9.7 1400 H0 3

Condensate N/A CST via Feedwater line 1.89 30000 2.9 415 Condense r Hotwells Conde nsa te N/A CST Core spray 0.06 1000 Unknown transfer header Head spray Recirc loops 4

RER drain N/A PSP Recirc loops 0.10 1600 0.4 65 CST Head spray RHR service N/A River Recirc loops 0.57 9000 1.1 162 water (Standby Head spray coolant supply system)

"See Sect. 7.6.

He rated head is 200 f t.

O e

--., -. ~ -

..c

77 e

Table 7.3.

Loss of DER accident post containment failure scenerlos Probable Scenerio outcome (1) Cont. failure + LOCA + all inj ection

= no melt (2) Cont. f ailure + LOCA + some inj ection

?

=

(3) Cont. f ailure + LOCA + no inj ection

= melt (4) Cont. f ailure + no LOCA + all inj ection = no melt (5) Cont. failure + no LOCA + some inj ection = no melt" b

(6) Cont. failure + no LOCA + no inj ection

= melt r

" Assumes some inj ection capability in addition to SLC.

Other than SLC.

l e

78 ORNL-DWG 81-8602A ETD f

bkgk(hh/ REACTOR VESSEL DRYWE LL

.g g

q MAIN STEAM LINE FEEDWATER LINE

/

i E IRCULATION

{.

VENT PIPE (i

1

\\f TORUS OR WETWELL

d 1,

p j/

DOWNCOMER W

\\

C.

3w.

CONCRETE FLOOR R SSION OF DRYWELL POOL Fig. 7.1.

BWR MARK I containment sy s t em.

e 1

79 ORNL-DWG 83-4244 ETD 5

A.

DRYWELL LINER E

.O E

~

2-1/4 in. POLYESTcR FOAM E

$ 2-3/8 in. FIBE RGLASS b

[

@ FIBERGLASS LAMINATE FORM 5

9

[

] SAND 2

..%'o

• CONCRETE E

• I$

' 9:p..

b..

l 2

E o

~

566.0 ft

• • CONCRETE '

SHIELD 4

550.3 ft 3 (;;lS.U.S.f!i i?

%E'.i k

.I,b;b[b.h}k,{,h,':5.' !*.! 5

,n'b p[

Y N ;.,., ; ~5I y;

'~

w Fig. 7.2.

BFNP drywell liner - concrete shield wall gap construc-tion.

l i

l l

~,

80 ORNL-DWG 83-4245 ETD TO REFUELING FLOOR.

jg i

L TO VARIOUS REACTOR BUILDING FLOORS j

f TO SUPPRESSION POOL CHAMBER ROOM Fig. 7.3.

Drywell liner - concrete shield wall gar venting paths.

t f

4 e

J

81 e

ORNL-DWG 83-4246 ETD h DRYWELL LINER GAP /VENTPATH f

DRYWELL INTERIOR PENETRATION SLEEVE

/

PENETRATION ASSEMBLY a

lllllll/////////////

l

/

Fig. 7.4.

Drywell penetration sleeve configuration.

4

w 82 ORNL-OWG 83-4247 ETD

.sr. e-

,o a e-I i

O JF w g

W / m,,,,,,,,,

so a,

4rese A,ess.r l

z,. _s

,e v.,ren

).

sr **.

d#N ' #*# 8'"## #

h

. Fy f ee. h# *"#3

__&__~-'

,._[.L.a

.i._.*

I 59 f ene O ee

,,,,_m s.e.,th

.~

.aa,,n-1

.f.,. g_

N rNs m

s4

. ~~~' %.c

~

< A'

'\\

c g

i i.l Y /

',*".""k' l'

b.

l h.

c.-

a..,.-

.-r

.(.,,u, ro. -- - -.

l l

y a.,,.

)

-o V

\\.

\\

.i N.

g.

c-Mot 9,,'8

' s.

_ z.. ', y s,g;yp2 p

r

%g

~,-

~ ~ ~ ' ' ' '

d

rias, M--

/ gU M

a

- E=r v-y-D 9 3'_.

Mr

\\,Q

• M*'

]QK-

%rg a

g

,c-...,

/

l x % c.,,

../

Na--~ ~ -

i 5

/

.c t

l

/

N

=

\\

/

3

.,.. n ara,.

ar e

.-e

,a,.-.

)

l r.

w,,, i b

41

}ql I

.\\

vc s

" "*~' ' u,l - ;'g/,[C'""

,'}1

?i 1 m

aNl eEe N.

f

• * " * ~ '4 f

5

..yl

)

g r

r t

,4 " f4GE:r. ' '

n__y 41[,

.,jk

. v2 -

.A g e

,c _,y!

s~

L

. - a.m.,a y

Fig. 7.5.

Unit I and Unit 2 pressure suppression chamber rooms and ECCS pump rooms (Elevation 519 f t.).

83 ORNL-DWG 82 -19311 N

BLOWOUT BLOWOUT PANELS PANELS REFUELING BAY

=

1 u

iii O

/

\\f I

RZACTOR CONTROL VESSEL ROOM i

.,,.. f".

4* 4

/..

p:..

By :::.

d ::

b:.

i!!!g iii :

A

'~

!!!)

i;iS

%... [

...f.

.::C::

.:C:: -

,;c..

i::

C::::.

.::C:: --

~ :.:.:.C.:::

.. f C.": /-

• C::;

. :.,:.... ]

E

.1 3.

~ :::

Ij[3[ 5$*!!

U:

....,'"=I* ~

C[: '

j' I

00'

.22: --? ) --

C ::

? """ ~~ -:'2 ,,,) ! f / ~,O, O [C{"b:*...::C C. ' -C":: " **""'-{::C::. CC:. :C:C::: T - -/

"R.

"Z:

C:

/ -~:- .::C::

?: p ::5

. 0.:: ,~~ ...: :.:0

37.'

r-l:.: '-?: t??: e , {:-- lj:: ' C_:: C.: -~:- i_ a Fig. 7.6. Static over pressurization containment failure locations.

i ORNL-DWG 83-4248 ETD / / t / r D 6-m's/ / / / / / t I / D / /# U v# / 6 / / / / h I/ / / Fig. 7.7. Drywell liner rupture configurations. 9 x

i 85 1 8. ANALYSIS OF CONTAINMENT FAILURE AND POST CONTAINMENT FAILURE LOSS OF DER EVENT SBQUENCE 1 8.1. Introduction Events preceeding containment failure in the uniform pool heating Loss of DER sequence were described in Chap. 3. This chapter will de-scribe the containment f ailure event and post containment f ailure event sequence for this accident (Scenario 6, Table 7.3). This scenario, whi:h is commonly analyzed in probabilistic risk assessments, involve s a con-tainment failure event followed by loss of all reactor vessel water injec-tion capability. The primary containment is assaned to f ail in the dry-well at the juncture of the cylindrical and spherical portions of the liner with a failure area of 0.929 m8 (10 ft8). The primary system is assumed to maintain its integrity and a pressurized boiloff of the water in the reactor vessel follows containment f ailure. All analyses described in this section are based on evaluations per-formed with the MARCHE.1 code as installed and modifiede.s at ORNL. The initial conditions for the analysis were derived f rom the BWR-LACP results discussed in Chap. 3. Table 8.1 is a summary of the initial conditions incorporated in the MARCH analysis described here. 8.2 Loss of DER with Loss of Inloction Followina Containment Failure Table 8.2 presents a summary of the accident event timing for this scenario as predicted by MARCH. The MARCH results for this accident are shown in Figs 8.1 through 8.13. Containment f ailure is predicted to occur 351/4 h af ter scram, as the drywell pressure (Fig.,8.9) reaches 0.908 MPa'(117 psig). All water flow into the reactor vessel is assumed to cease at that time. Core uncovery'is not predicted to occur until almost 2-1/2 h (147 min) af ter loss of injection (Fig. 8.1). This rela-tively slow core uncovery process is due to the low decay heat levels in-volved and the large water inventory in the reactor vessel at the time of loss of inj ection. The only mechanism for water loss from the reactor vessel prior to vessel head failure is via the reactor vessel relief valves, which are cycling f requently during this period (Figs. 8.2-8.3). The large relief valve flow variations shown in Fig. 8.3 are due to errors in MARCH's re-lief valve model, which represents the valves as orfices with flow rates dependent upon downstream (suppression pool) pressure. In actuality the BFNP SRVs are critical flow devices that are designed to maintain constant flows independent of containment pressure. This MARCH modeling error ac-counts for the SRV flow spike (Fig. 8.3) and rapid primary system depres-surization (Fig. 8.2) predicted to occur af ter containment f ailure. Re-actor water temperature (Fig. 8.4) is predicted to hold very near the pri-mary system saturation temperature throughout the period prior to vessel failure. j i - - _, _,, - - ~,.

86 Following core uncovery at 37.7 h (2262 min) af ter inception of the a accident sequence, maximum fuel temperatures are predicted to rise rapidly to 1533 K (2300*F) (Fig. 8.5) and the Zircaloy fuel cladding is predicted to oxidize rapidly (Fig. 8.6). Hydrogen leakage rates into the suppres- + sion pool peak at ~0.25 kg/s (33 lbs/m) (Fig. C.7). The energy from the Zircaloy oxidation reaction increases the fuel heatup rate (Fig. 8.5) resulting in initiation of fuel melting (Fig. 8.8) at 2321 min, almost 31/2 h af ter loss of injection. The length of time between loss of inj ec-tion and inception of core melting in this accident is closely related to the low decay heat levels involved and the intensity of the Zr-H 0 reac-3 tion, since the heat generation from this reaction can easily equal or exceed the decay heat levels present 30 to 40 h af ter reactor shutdown. Once initiated, core melting continues at a moderate rate, until 75% of the core is molten (Fig. 8.8), at 40.2 h (2413 min) into the accident. At this time the core is allowed to slump onto the core plate, based upon a MARCH user input option. Figures 8.9 through 8.13 are plots of the drywell pressure, leak rate, temperature, hydrogen solar fraction and liner temperature through-out the post containment phase of the sequence. Following drywell f ailure at 2115 min, containment pressure is predicted to drop rapidly as the con-2 (10 ft2) (assumed) opening in tainment depressurizes through the 0.929 m the drywell liner. Drywell pressure drops to 0.4 MPa (55 psia) within 10 min and to atmospheric pressure within 64 min. During this period the dry-well volumetric leak rate is predicted to hold rather steady at approxi-mately 15S m /s (335,000 ft8/ min). This flow rate corresponds to a mass 8 velocity of 177 m/s (558 ft/s) through the break. Following drywell fail-ure, the temperature of the drywell atmosphere drops substantially, due to depressurization and flashing of the wetwell through the downconers at-tached to the vent header. Following core uncovery at 2262 min, the tem-perature of the drywell atmosphere increases substantially due to gas and steam influx from the suppression pool which, in turn, is receiving the hot gas and steam from the uncovered core via the reactor vessel relief valves. References for Chanter 8 8.1 R. O. Wooten and H. I. Avci, MARCH User's Manual, NUREG/CR-1711. 8.2 W. O. Co ndon, S. R. Greene, R. M. Harrington, S. A. Ho d ge, "SBLOCA Outside Containment at Browns Ferry Unit One-Accident Sequence Analysis," NUREG/CR-2672, November 1982. 4 1 e

87 Table 8.1. Initial conditions for LDER sequence Time from scram, h 34 RPV water icvel, in. 562 RPV pressure, psia 1072 RPV water temperature, 'F 554 RPV water inj ection flow, gym 102 Drywell pressure, psia 125 Drywell temperature, 'F

• 31 Drywell relative humidity, %

26.5 Drywell liner temperature, 'F 362 Assumed drywell failure size, ft2 10 Table 8.2. Loss of DER accident event timing Time Time f** f*8" Event scram LI (min) (min) Containment failure 2115 Loss of RPV injection 2115 0 Core uncovers 2262 147 Core melting begins 2321 206 Core slump 2413 298 l RPV bottom head failure 2504 389 " Time f rom Loss of Inj ection. l. ~, - -,. - -,, - - _. -. _

88 ORNL-DWG 83-4249 ETD o lQ ~ d M O CONTAINMENT FAILURE (CRD g /

• INJECTION CEASES)

M NN N A A No m m w. 8. 8-M N RPV WATER INVENTORY g O* LOSS DUE TO SRV mo ACTUATION 4 TAF AT 364 in. a6 A .3 N W >o $~ b~ N N BAF AT 216 in. E<* g N o C.q. o g. N* gaa - M m A CL. A A A A O-y $_ g,o g_ 22' 0.0 2300.0 2400.0 2500.0 0 2000.0 2100.0 TIME - MINUTES Fig. 8.1. Reactor vessel water level. (Time is counted from reactor scram). 9 s -w iga --n -4

89 a ORNL-DWG 83--4250 ETD b* E g 4 CONTAINMENT FAILURE k.-. O - ^9

a. g -

co N <C b w x w D M, D4 v g-Co 8 - [0

• m a,

e E* PRESSURE HELD CONSTANT BY SRV ACTUATION fn N p p. = i en Cn b e 2 42o k N % o k l 22' 0.0 2$00.0 2M.0 2500.0 2000.0 2100.0 0 TIME - MINUTES Fig. 8.2. Reactor vessel pressure. l e l 1 l ~

90 t ORNL-DWG 83-4251 ETD 9 o k-d 0-x s N I,,, f-N- CORE SLUMP % 5

• a s

s E m m $q 4-CONTAINMENT FAILURE $ $~ $ "~ E s.- o. E 2 2 O o u. E" l" y S- < o-wr n a: 4 W E o m 4 O w M w a "I

• 2 e s

~ 6 d~ 6 m 8.. 8 C 22' 0.0 2300.0 2400.0 2500.0 0 2000.0 2100.0 TIME (min) Fig. 8.3. SRV steam flow. t 4 a

91 e e ORNL-DWG 83 -4252 ETD $~ @ r jw CONTAINMENT {h- {h-FAILURE a; a: b e g-e s-0: a: N M E-Ed k 8 g 5-g n-B B >= S h~ b~ D D 2: 5 c: 8 CL. g-

4. g.

I8n 2000.0 21b0.0 2200.0 23b0.0 24b0.0 2500.0 TIME - MINUTES l Fig. 8.4. Reactor vessel water temperature. i __-.,..m

l 92 ORNL-DWG 83-4253 ETD m o 9 8_ m o ~ ~ ^ m D d do 2h E $- E4 g Ng E-E- Ed M C4 a: O O O O e o DN_ DE_ 2 2 e.* w X X M 2 9 o o_ N c-5 a_ a 2000.0 2100D 2200D 2300.0 2400.0 2500.0 TIME - MINUTES Fig. 8.5. Maximum fuel temperature. O g b .n - - - - - - -,, ~.

93 e e ORNL-DWG 83-4254 ETD 8-8 o-a Q f m a b $- b 5-M M M M Q Q <n <n a o-3 o-O O Z Z O O 5 5 ON ON 4 o-4 o-MELTING BEGINS N N AT 2321 min o-o- CORE UNCOVERS AT 2262 min h o o l o_ o. 22' 0.0 2300.0 24'00.0 2500.0 l 2000.0 2100.0 0 TIME - MINUTES Fig. 8.6. Total fraction of fuel cladding reacted. b I . -.. ~

a 94 ORNL-DWG 83--4255 ETD o N N Q n 2 i 'cn m Eo ~ ~ a = e z m e m 23 o E n-

• 6-

n. -

m' 2o 2 E k' m 8 o S-C S-m W d O<d 0 k h 5 2w z O 6 C 0 I J I e g V e o 6 6 2000.0 2100.0 2200.0 2300.0 2400.0 2500.0 TIME (min) Fig. 8.7. Hydrogen leakage from the reactor vessel. b o ~-

i 95 ORNL-DWG 83-4256 ETD g g = n G G 8-8- 6 6 Q Q CORE SLUMP b o_ 46_ AT 2413 min W w m 2 N W Q: 8 Q: 8 O Q. O c_ Ue U$ Z Z O O 5 e-5 r-O co oe < M- < 8-Q: o C4 o N N b b f a-a- CORE MELTING BEGINS AT 2321 min 6. 6 2000.0 2100.0 2200.0 23'00.0 24'00.0 2500.0 TIME - MINUTES I i Fig. 8.8. Fraction of core melted. 6 I i l

A%>. 5N3M 6O <. L%Mmmp%W vDl-3 3 S Q%>*N4J. POhA %%WMmD%N v%m 2 go No 2000 Y 0 F ig 25 0 4 0 8 9 AB AD T TL TR I O 2 M2 1W 2Y 1W D 2 7D 1 r E4' 9 E O 5 9 y 0 mW mL 6 L w - 0 i i nN e nF l C A M O I l L M U I P p N2 L R r E 3 E e U 6' T s T0 E O s E 0 R u S N r .e L D 24 W 8' G 0 0 8 3 n 4 2 5 2 7 6 E 00 T 0 D 4 8

W* OGO cWL3C O 2 OmZ n D 2 h 81 2 9 4 1 G 5 7 0 5 6 1 81 3 ) 5 ) 3 s x / 1 3 n m i m ( / E 3 T ft A ( R E K T 3 A 2 A 9 R 5 E 8i L 1 K 2 1 C A I E R L T E C I M R U T L E O 9 M 8 V 2 U 6 9i L T 7 O 1 N E V e M T N N I E A M TN N e O 4 AI C 6 T 4 9t N 8 3 O C 0 0 _mL 0I 0 o g$ 86 56 N$o6 N O t d 5 3_.b 1g* . O. ,4<#W W 4SgenMmO we*W

• n***,

'T .,i il1

i ~ NAB N2%NxkD v2c g, o lM 2N n _k ~ ~ p NAB N%%NxkD ve b g, o ko Hh o ~ ~ 2000 F 0 ig FC AO 2 IN 1 8 2 r L T 0 U A 1 0 RI 1 EN M i E T N I D T r M22 9 y E4' 8 w 0 e - 0 l l M t I e Y m N2 C p U3 O 6 R e T0 E O r 0 E U R a t S N t N u C L r O V .e 2 E D 4 R W 8 S G 0 0 83 J\\ 4 2 e 5 m 2 9 6 E 0 T 0 0 D w n e

DRYWELL-MOLE FRACTION OF HYDROGEN ,0.00000 0.08333 0.186s7 025000 0.33333 o.4,1687 o.50000 DRYWELL: MOLE FRACTION OF HYDROGEN .o.00000 0.08,333 0.16,667 0.25, 000 0.33, 333 o.41667 0.50000 s

• y a

e y tp a = b e E I8 o l'3 C e o a l o. n o. 2 o C - Ne 8 = m i g m i N O N N g-o o o ) b P A L m h ITI 8 e o i

100 ORNL-DWG 83--4261 ETD N 9o ] I ^ mb F' Q ~ ~ a ~ M N Q C O O 7. V. g [ h I A % g-M M N N H H 1 J J$ .1o .$ @ - 5 %, _/ ~ H H CORE '4 UNCOVERS x x Zg

7. o Q o_

Q g_ CONTAINMENT E-* Q E-n FAILURE 7. Z O O U U D 8 J a 2000.0 2No.0 2240.0 2$0.0 24N.0 2600.0 TIME - MINUTES Fig. 8.13. Drywell liner temperature. 4

101 9. IMPLICATIONS OF RESULTS The first purpose of this chapter is to provide a discussion of the present state of readiness at the Browns Ferry Nuclear Plant to cope with an' accident sequence involving a long-tern loss of decay heat removal (DHR). This' accident involves an improbable combination of inability to use the main condenser, condensate, and feedwater systems for decay heat removal with an inability to use the residual heat removal (RHR) system in either the pressure suppression pool cooling mode or the shutdown cooling mode.* The unavailability of these RHR system operational modes might be due to complete system failure, or the f ailure might be confined to an inability to use the residual heat removal service water (RHRSW) system to remove heat f rom the heat exchangers in the RHR system. In the latter case, the RHR system could be used to circulate the pressure suppression pool water to promote mixing and avoid thermal stratification. Both cases of loss of RHR system decay heat removal function have been analyzed in J this report. The available instrumentation, the level of operator train-ing, the existing emergency operating instructions, and the overall system design at the Browns Ferry Nuclear Plant Unit 1 are discussed in Sects. 9.1 through 9.3 from the standpoint of adequacy in the event of a long-tern Loss of DER accident sequence. The second purpose of this chapter is to provide a discussion of the impact of this detailed SASA analysis of the Loss of DER accident sequence on the conclusions of the recently completed Interim Reliability Evalua-tion Program (IREP) report,*.2 which identifies this accident sequence as a maj or contributor to the overall risk attendant to the operation of the Browns Ferry Nuclear Plant. This discussion is provided in Sect. 9.4. 9.1 Instrumentation All control room and other plant instrumentation normally available af ter a reactor scram would be available for operator use during a long-tern Loss of DER accident sequence even if a prolonged loss of offsite power were involved. The primary containment parameters measured by the available instruments and displayed in the control room include the pres-sure and temperature of the drywell atmosphere, the temperature and level l of the water in the pressure suppression pool, and the pressure in the l we tw ell. The range of indication and the associated alarms for each of these parameters are provided in Table 9.1. As discussed in Sect. 7.4, the best-estimate pressure currently avail-able for f ailure of the Browns Ferry Mark I primary containment is 0.908 MPa (117 psig). Therefore, during the latter stage s of a loss of DHR ac-cident sequence which the operators permitted to go to completion (i.e., containment f ailure by overpressurization), the drywell and wetwell pres-sure instruments would be off-scale high. The pressure suppression pool

• These RHR system operational mode s are de scribed in Appendix A.

l The accident also implies inability to use the RHR system unit crosstie capability de scribed in subsection 2.5.5.

102 level instruments would be off scale-high af ter about 700 min (11.7 h) into the accident sequence' as the suppression pool continued to swell in re-sponse to heating and the absorption of the SRV discharge. On the other hand, the existing drywell and pressure suppression pool temperature in-dication would remain onscale throughout almost all of the accident se-quence. Since the wetwell atmosphere would be virtually 100% steam during the 1atter part of the accident sequence, 'the pressure in the primary contain-ment could be inferred from the pressure suppression pool temperature and the saturation tables. 9.2 Operator Preparedness The Browns Ferry training simulator does have the capability to model the portion of a Loss of DER accident sequence before drywell failure. However, a complete run-through of this slowly-developing accident se-quence would require about 35 h of simulator time and it is doubtful that the simulator computer could continue to develop a realistic model of the plant response for such a long period of computation. For these reasons, a Loss of DER sequence is not included in operator training although the importance of pressure suppression pool cooling and the methods for ac-complishing this function are stressed. Af ter reactor scram or manual shutdown under accident conditions when the main condenser is not available, the pressure suppression pool serves as the heat sink for the. decay heat generated by the reactor core. Thus it is important that operators understand the temperature-time response of 4 the suppression pool when all of the reactor decay heat must be absorbed + therein. If the suppression pool cooling does not function, the operator should appreciate the enormous capacity of the pool for energy storage. This has safety implications, because the operators might be reluctant to 4 use the pool to best advantage under accident conditions if they do not understand its potential. For example, in a Loss of DHR accident sequence the operators might be reluctant to depressurize the reactor vessel when required because of an unjustified assumption that this action might lead to a large increase in pool temperature and consequently in primary containment pressure. In f act, over the long term the only significant difference in the amount of energy stored in the pressure suppression pool with or without-a reactor vessel depressurization is the difference in the sensible heat of the water mass in the reactor vessel associated with the saturation tempera-tures at the pressures before and after depressurization. Similarly, the existence of a stuch-open relief valve does not cause the pressure sup-pression pool to be heated much more rapidly; in the long term, the energy storage within the j ool is limited to the decay heat generation within the core in either case, It is recommented that methods to promote better operator understand-ing of the capabilities and response of the pressure suppression pool when pool cooling is not tvailable be included in operator training. It is also recoatended that Emergency Operating Instruction (EDI) 41 be upgraded to indicati the amount of reactor vessel inj ection available -~ -e- -,n-- ,,--.y-, ->~,,-.-~,,n , ~,,, ,,---,,,,-m._<,en- ,,-em .-m,-

? 103 under the various operating modes of the CRD hydraulic system and as a function of reactor vessel pressure. It is particularly important that operators, recognize that the cooling water inj ection by the CRD hydraulic system 'is significantly increased when a scram is in effect. The f act that the raw cooling water system is needed for CRD pump cooling should also be taught. The portion of EDI 41 dealing with the injection capa-bility of the standby liquid control (SLC) system should also be modified to indicate the correct injection rate under the consideration that only one SLC pump can be operated at a time. The importance of running the RHR system in the suppression pool cooling mode even if cooling water to the heat exchangers is not available should be stressed. The action significantly reduces the thermal strati-fication in the pool. Finally, it is recommended that the appropriate E01 stipulate that if very high containment pressures are encountered in a Loss of DER acci-doni sequence in which the core remains covered and there has been no re-lease of fission products into the primary containment system, then the primary containment should be vented by whatever means necessary to pre-clude containment f ailure by overpressurization. In such a case, there is no need to risk the chance that violent disruption of the drywell might cause rupture of piping systems connected to the reactor vessel. 9.3 System Desian As described in Appendix A, pressure suppression pool cooling for Browns Ferry Unit 1 is provided by four.EER pumps and heat exchangers arranged in two separate loops. Each RHR pump is powered from a separate shutdown board; in the event of a loss of offsite power, each shutdown board is ' supplied f rom a different diesel generator. It is shown in Sect. 4.2 that adequate suppression pool cooling is provided if any one RHR pump and its associated heat exchanger are in operation in the suppression pool l cooling mode. Successful heat exchanger operation requires that RHR service water (RHRSW) be supplied to the tube side for cooling of the suppression pool water which is on the shell side of the heat exchangers. RHRSW is pumped from the Tennessee River to each of the RHR heat exchangers by pumps lo-cated on the intake structure at the river and dedicated for this purpose. Each unit I heat exchanger is served by a different RHRSW pump, with a backup pump available if required. The arrangement of diesel generator power supplies for use in the case of a loss of offsite power is shown in Table 9.2. l Should all of the RHR pumps and/or heat exchangers on Unit I be un-l available, a crosstie arrangement permits the use of the A or C RHR pumps and heat exchangers on Unit 2 to circulate and cool the Unit 1 pressure suppression pool water. The crosstie network is designed for a flow of 0.315 a /s (5000 spa) which would permit operation of one Unit 2 heat ex-s changer at about 91 percent of its heat transfer capability at full flow [0.630 m8/s (10,000 spm)].8.2 The results of this study do not indicate that an improvement in the de sign of the RHR system or RHRSW system at Browns Ferry is required from

104 the standpoint of readiness to cope with the initiation of a loss of DER accident sequence involving the RHR pumps and heat exchangers of Unit 1. This conclusion is based on the results presented in Chap. 3 which show that about 21.5 h would be available before the containment pressure in-creased to its design value of 0.487 MPa (56 psig) and a total of 35 h would be available before the estimated drywell failure pressure of 0.910 MPa (117 psig)'.s is reached. This, combined with the redundancy of the systems involved, would seem to ensure a very high probability that at least one of the six* available RHR pumps and heat exchangers could be brought to bear on the Unit 1 pressure suppression pool as necessary to prevent containment f ailure. A design consideration first identified in the SASA study of Station Blackout at Browns Ferry'.* also has direct application to the loss of DER accident sequence. Provision is made for an automatic shif t of the high-pressure coolant inj ection (HPCI) booster pump suction from the condensate storage tank to the pressure suppression pool on high sensed suppression pool level. Once this shift occurs, the pump suction cannot be trans-ferred back to the condensate storage tank. Because the HPCI turbino lubricating oil is cooled by the water being pumped and the pressure sup-pression pool temperature is elevated in many accident sequences, this autamatic shif t can cause f ailure of the HPCI system by overheating of the lubricating oil. The normal pressure suppression pool level is between -2 and -6 in. indica ted and the autaentic HPCI pump suction shif t will occur if the water level increase s to +7 in. This would occur be tween 2 and 4 h af ter the inception of the accident, when the suppression pool temperature had increased to about 344 K (160*F). The pool temperature would continue to increase after the shift. Since the turbine oil cooler is designed for a maFimum inlet water temperature of 333 K (140*F), the oil would become overheated leading to a possible f ailure of the turbine bearings. An emple amount of relatively cool water would remain available in the condensate storage tank at the time the HPCI pump suction was automa-tically shif ted to the pressure suppression pool. High suppression pool level at +7 in. requires the addition of between 257 and 375 m8 (68,000 and 99,000 gal) whereas the normal condensate storage tank volume is about 1370 m8 (362,000 gal). Thus water transferred fram the condensate storage tank into the reactor vessel and from there to the suppression pool as steam via the vessel relief valves would produce a pool level of +7 in. long before the condensate storage tank was emptied. The threst to the HPCI system i(entified here is not unique to Loss of DER sequences; it would also exist in other accident sequences because high suppression pool temperature is caused by the pool heating attendant to the condensation of steam in the pool, which is also the source of the increased water level. It should be noted that separate provision is made for an automatic shif t of the HPCI pump suction if the normal source of condensate storage tank water becomes exhausted. Thus it appears that the automatic high 4

• This includes the two available via the crosstie connection to Unit 2.

l

I 105 pool water level shif t must have been straight-forwardly based on a con-corn for the effect. of high water level in the pressure suppression pool. The basis is not given in the plant Technical Specifications and there is no corresponding high-level shif t for the Reactor Core Isolation Cooling (RCIC) system, whose operation can also lead to high suppression pool level. A survey of plant suppression pool drawings does not reveal why a suppre ssion pool level of +7 in. should be of concern. It is recommended that the desirability of an antamatic shif t of the HPCI pump suction on high sensed suppression pool level without the oppor-tunity for reversal by the operator be reconsidered. It is a conclusion of the authors of this study that the Loss of DER acc ide nt sequence is probably not a dominant contributor to the overall core melt probability at Browns Ferry when all available mitigation f ac-tors

• are taken into consideration (see the discussion in the following Sect. 9.4).

However, should subsequent PRA studies confirm that loss of DER is a relatively high-risk threat that must be guarded against, then (1) The control air system pressure should be increased to permit remoto-manual operation of the SRVs at high containment pressures. (2) The pressure suppression pool water level instrumentation should be temperature-compensated and the range of control room readout should be expanded, and (3) the range of control room readout of drywell and wetwell pressure should be expanded. 9.4 Reconsideration of IREP Study Conclusions i A probabilistic risk assessment (PRA) of Browns Ferry Unit I was re-cently completed as part of the overall effort of the Nuclear Regulatory Commission Interim Reliability Evaluation Program (IREP).8.1 One of the specific goals of the study was to identify the dominant contributors to core molt. The accident ser,uences so identified are listed in Table 9.3. The discussion can be simplified by combining certain of the accident sequences. As indicated in Table 9.3, the fif th sequence is identical to the first and the eighth sequence is identical to the third except that the HPCI system is used for vessel inj ection instead of the RCIC system in sequences Nos. 5 and 8. Since it makes no difference to the progression of these Loss of DER accident sequences whether the RCIC or HPCI systems are used for vessel level control during the periods when the reactor ves-sel is pressurized,t the eight dominant sequences identified by the IREP study can be regrouped into six sequences as listed in Table 9.4. This is done by co,mbining sequences 1 and 5 and combining sequences 3 and 8. If all of the sequences listed in Table 9.4 that involve loss of DER are com-bined,t we have an initial frequency of 1.8 x 10-s and a final core melt l

• Including the potential for effective operator action over a 30 h int erv al.

l iThere alght be a difference if it were the HPCI system that was un-l available in either sequence, since it has a higher turbine exhaust pres-sure trip [1'.138 MPa (150 psig)] than does the RCIC system [0.276 MPa (25 psig)]. tThese are sequence s 1, 3, 4, and 6. l

106 frequency of 1.4 x 10-4 events per reactor year for BWR 4 -- Mark I con-tainment plants identical to Browns Ferry Unit 1. This is a high pre-dicted probability of corc melt as a result of an inability to cool the pressure suppression pool or establish reactor vessel shutdown cooling. Indeed, the following excerpt from the executive summary to the IREP study illustrates the importance of these findings: "Six of the eight dominant sequences identified involve failure of the torus cooling and shutdown cooling modes of the i RHR system. These sequences account for ~73% of the sum of the dominant sequence f requencies. Therefore, no significant reduction in core melt f requency can be achieved without reduc-ing the unavailability of the RHR system or providing an alter- - nate means of long-term decay heat removal. Thus, the RHR sys-tem is the most risk-critical system at BF1." It is one purpose of this report to make recommendation as to whether these findings of the IREP study should be reconsidered based upon the results of this detailed SASA study of the Loss of DHR accident sequence. The SASA approach involves a much more detailed analysis than was possible in the particular PRA methodology employed in the IREP study. The find-ings of the SASA analysis that are believed to have a major bearing on an evaluation of the conclusions of the IREP analysis are discussed in the following sections: 9.4.1 Operation of the CRD system hydraulic ouso The operation of the CRD system hydraulic paap was neglected in the IREP study. This pump takes suction on the condensate storage tank and serves to inj ect 0.004 m8/s (60 spm) of control rod drive mechanism cool-ing flow into the reactor vessel under normal operating conditions. Fol-lowing a scram, the vessel inj ection flow increases to over 0.006 m /s 8 (100 gym)* until the scram is reset, when inj ection flow is again reduced to 0.004 m /s (60 spm).8.8 8 In the Loss of DHR accident sequence, a scram signal is continuously present af ter the drywell pressure reaches 0.115 MPs (2 psig). Thus ves-sel inj ection by the CRD hydraulic system would be available at an in-creased rate throughout most of the accident sequence.t Af ter 4 h, this inj ection rate would be suf ficient to maintain reactor vessel water level without the aid of any higher capacity system; in fact, the operator would have to throttle the flow or cycle the CRD hydraulic pumps on and off to prevent overfilling the vessel. The condensate storage tank can easily be refilled f rom a variety of source s,- if nece ssary, but the volume of water normally maintained in this tank would be more than suf ficient to maintain vessel level until the time of containment f ailure as shown in Fig. 3.3. 'Ihe flow depends on reactor vessel pressure, ranging from 113 spa with the vessel fully pressurized to 182 spa with the vessel fully depres-surized. i tThe drywell pressure reaches 0.115 MPa (2 psig) about I h af ter in-ception of the accident se que nce. This is a scram trip setpoint. 1 -s + -, v-.- -.n -,n -,-n------e-a.-,,., .nn,.--_-------,--n~

107 i Since the CRD hydraulic pump is capable of supplying all required reactor vessel inj ection af ter the first 4 h of a Loss of DHR accident sequence and because the pump takes suction on a water source that is not affected by the heatsp of the pressure suppression pool, it eeems that = the effect of its existence must be considered in any analysis of a prob-ability of core melt by way of a Loss of. DHR accident sequence. Operation 4 of the CRD hydraulic pump prevents core melt during the pool heatup phase e of. Loss of DER sequence because it is independent of the status of the pressure suppression pool. The effect of the CRD hydraulic system inj ection should be f actored into the PRA by assigning a probability that this inj ection would not be available during a Loss of DHR accident sequence. For a transient-induced sequence or a transient-induced sequence with SORY (sequences 1 and 4 of Table 9.4), this probability woald be very low since no operator action is required and there is no common mode f ailure mechanima for the CRD hydrau-lic pump. Thus consideration of the CRD pump operation should signifi-cantly affect the calculated f requency of these sequence s. In the case of the Loss of DER sequences whose initiating events in-clade a loss of off site power (sequences 3 and 6 of Table 9.4), the reduc-j tion in core melt f requency due to consideration of CRD hydraulic pump a operation would not be as significant because only the spare CRD hydraulic pump 1B is supplied f rom an electrical bus (shutdown board A) powered by a diesel generator. Thus operator action would be necessary to restore CRD hydraulic pump operation following a LOSP including the starting of a raw cooling water pump for CRD pump motor cooling. Nevertheless, the very long period of time available for such action should be taken into con-sideration.* 9.4.2 Effect of containment backeressure The IEEP study assumes loss of inj ection capability to the reactor vessel as soon as the pressure suppression pool bulk temperature reaches 355 K (180*F) because of loss of net positive suction head (NPSH) to the ECCS pumps which take suction on the suppression pool. This assumption leads to early core uncovery (about 5 h) with the containment intact and is unrealistic for the following reasons: 1. Containment backpressure acts to maintain the NPSH above the minimum recommended for operation by the RHR pump manuf acturer for over 9 h if the drywell coolers remain operating and for over 22 h if the drywell coolers are not restarted by the operator af ter tripping on a combina-tion of low reactor vessel pressure and high drywell pressure at about 2 h after inception of the accident (see Fig. 3.12). As discussed in Sect. 3.3.5, in plant testing at Browns Ferry has shown that the RHR pumps are capable of operation with a NPSH signi,ficantly lower than

• For example, the RCIC system would remain available for inj ection l

from the condensate storage tank until the turbine trip setpoint of 0.276 MPa (25 psig) containment pressure is reached about 13 h into the acci-dent.

108 the manuf acturer's recommended minimum. A very low pumped flow [less than 0.009 m8/s (150 spm)] is sufficient to keep the core covered af ter 4 h (see Fig. 3.2). 2. The RHR pumps and the core spray pumps can be manually shif ted by the operator at any time

• to take suction on the condensate storage tank instead of the pressure suppression pool. This action would maintain operation of these systems even if sufficient NPSH did not exist for suction on the suppression pool. The RCIC system normally takes suc-tion on the condensate storage tank and would remain so aligned unless shifted by the operator.t For these reasons, it is not reasonable to assume that all inj ection to the reactor vessel is lost when the suppression pool temperature reaches 355 K (180*F), even if the contribution of the CRD hydraulic pump (Sect.

9.4.1) continue s to be neglected. In any event, the erobability that there is no containment backpressure and the erobability that the operator would not shif t the suction of the RHR or core spray pumps to the conden-sate storage tank should be included in the analysis. This inclusion would reduce the frequencies of all of the Loss of DER accident sequence s (Nos.1, 3, 4, and 6) listed in Table 9.4. 9.4.3 Content of condensate storane tank s (135,000 gal) in the The IREP study assumed a volume ;of only 511 a condensate storage tank at the inception of the accident; this corresponds to the volume guaranteed to the ECCS systems and the CRD hydraulic pumps by the existence of a standpipe within the tank which feeds all other sys-tem s.t The assumption used in the IREP study is unrealistic because the Browns Ferry operating instructions for the condensate system require the operator to keep the condensate storage tank nearly full. The median vol-une for the al.'.owable operating band is 1370 as (362,000 gal). As shown in Fig. 3.3, this amount of water is more than sufficient to maintain j vessel level up to the time of containment f ailure by overpressurization. Even if the initial volume were significantly lower, the condensate stor-age tank can be easily refilled from several sources, including a fire-truck from a nearby town if necessary during the very long Loss of DBR accident se que nce. The matter is important because at Browns Ferry all ECCS systems can take suction on a condensate storage tank 5 and therefore succe ssful inj ection doe s not depend on the status of the pressure sup-pression pool as long as water remains in the condensate storage tank. s Thus the probability that the condensate storage tank is at only 500 a

• There are no interlocks on the pump suction valves.

iThe HPCI pump is also normally aligned to the condensate storage tank. However, the HPCI booster pump suction is automatically and irre-versibly shif ted to the suppression pool when the sensed suppression pool level reaches +7 in. iPrimarily makeup to the main condenser hotwell. sit should be noted that the condensate storage tanks can be cross- . connected. -.,------------e -,.--,-.-,---m.----..# ,-- - ~ - - -. -3.n*,-,.m-,o---,,,-.-e.,--c,-.-,--.--w. .ww-n--r---,----m-,---

l 109 (135,000 gal) at the inception of the accident and the urobability that the tank is not refilled if necessary during the accident should be f ac-tored into the analysis. 9.4.4 RER system minimum flow bvosss valves The RHR system is equipped with minimum flow valves which open to per-mit a limited pumped flow to the pressure suppression pool when the main pump discharge paths are closed. This is for protection of the pumps, and the minimum flow bypass valves are automatically interlocked to close when a flow is established in the main discharge paths. The IREP study assumed a complete f ailure of the RHR system function if the minimum flow bypass valves f ailed to close. This is unrealistic because 90% of the flow to the main RHR discharge path can be maintained with the minimum flow bypass valves open. Recognition that the RHR system can perform its function during a loss of DER accident even with the minimum flow bypass valves f ailed open would reduce the frequencies of the loss of DHR sequences listed in Table 9.4 by a f actor of 22.* 9.4.5 Use of the SBCS system to control crimary containment pressure The IREP study did not consider use of the standby coolant supply system (SBCS) t for control of primary containment pressure. If all else failed, this system could be used to inj ect river water into the drywell or wetwell spray headers as a means to reduce the pressure in the primary containment and thereby avoid containment f ailure. Since the shutoff head of the RHRSW pumps is about 1.22 MPa (162 psig) and the elevation differ-ential between the river and the suppression pool is insignificant, this method of containment spray could be used at any time during the loss of DHR accident sequence. I l i 9.4.6 Reauirements for the EECW system The emergency equipment cooling water (EECW) system provide s cooling water under emergency conditions to the diesel generators and other safety systems. For the IREP study, it is assumed that three of the four EECW pumps are necessary for the performance of function unless the operator takes action to manually eliminate the less-essential loads. This assump-tion has a significant effect on the initial frequencies for the sequences listed in Table 9.4 that involve a loss of offsite power. In fact, the elimination of less-essential loads is automatic,t reducing the require-ment for EECW performance to two out of four pumps. Since the EREP study 'This according to the sensitivity study given in the IREP report. tThe SBCS is described in Sect. A.S.2 of Appendix A. tIhe service air compressor and RBCCW heat exchanger loads are auto-natically eliminated when water pressure in the EECW headers f alls below a preset value. I

110 1 _ did grant credit for manual action for load reduction, correction of this assumption abould not significantly affect the final frequencies listed in Table 9.4. . 9.4.7 Other considerations Other dif ficulties with the conclusion of the IREP study that loss of DER sequences are dominant at Browns Ferry include the study procedure that no credit is ever taken for recovery of the power conversion system (PCS) as a heat sink and that the potential for suppression pool cooling via the provided cross-connection to other units is ignored. These con-siderations have direct application to the loss of DHR accident sequence s and should not be neglected, especially af ter these sequences have been tentatively identified as dominant. The IREP study also did not consider use of the standby liquid con-trol (SLC) system as an alternate method for high pressure water makeup to the reactor vessel. The primary purpose of this system is inj ection of the neutron-absorbing sodium pentoborate in the unlikely event that this became necessary due to f ailure of control rod insertion. How ev er, it is possible to shif t the suction of the inj ection pumps from the SLC tank, which contains a solution of sodium pentaborate, to a storage tank con-8 (375,000 gal.) of domineralized water. taining approximately 1,419 m There are two positive-displacement SLC pumps at each Browns Ferry unit, each capable of inj ecting approximately 0.004 m /s (56 gpa) of water 8 into the reactor vessel. How ev er, the pump control circuitry is provided with interlocks to ensure that only one pump is operated at a time. The procedure for using one of the SLC pumps to inj ect demineralized water i into the reactor vessel for high pressure makeup under emergency condi-I tions is contained in Browns Ferry Emergency Operating Instruction (E0I) No. 41.

• The IREP study amission of consideration of the use of the SLC system as an alternate method for reactor vessel inj ection is probably not sig-nificant. First, the system is not needed: adequate vessel inj ection to replace the water boiled to steam in the reactor vessel and transferred to the suppression pool through an SRV can be provided by occasional supnen-tation of the CRD hydraulic system flow by the RCIC system during the first 4 h af ter a scram, even if the operator acts to depressurize the re-actor vessel during this period.

Successful operation of these systems does not depend on the status of the pressure suppression pool, and opera-tion of the CRD hydraulic system alone is sufficient to keep the core covered af ter 4 h following reactor scram.t Secondly, the SLC inj ection capacity of 0.004 m /s (56 gpm) is simply 8 not enough to maintain a constant reactor vessel water level even at the time predicted for containment f ailure 35 h af ter scram and inception of the loss of DHR accident. It might be postulated that credit should be

• It should be noted that this E0I erroneously claims that 0.008 m /s 8

(112 spm) can be inj ected by this method. As noted above, only one 0.004 a /s (56 spa) pump can be operated at a time. s tAssuming the reactor vessel has been depressurized, as per proce-dure. .. - -..,. -.... -. -. ~.- _ -.

111 taken in the IREP study for use of the SLC inj ection capability in conj unc-tion with another system whose normal capacity is degraded. How ev er, the needed inj ection rate is a very man 11 fraction of the normal inj ection rate of all other systems.* If these systems operate at all, they should provide the necessary inj ection [less than 0.011 m /s (170 spm) after 8 4 h]. t 9.4.8 Recommendations 1 There is concern that the IREP study assumptions concerning the fac-tors discussed 'in Sects. 9.4.1 through 9.4.6 might have caused the loss of DER accident sequences to unrealistically appear to constitute the majority of the dominant sequences for core melt at Browns Ferry Unit 1. Accordingly, it is recommended that the conclusions of the IREP study with regard to the Loss of DER accident sequences be reconsidered, with the information discussed in this section and summarized below: 1. The CRD hydraulic pump is capable of supplying all needed vessel injection af ter the first 4 h of a Loss of DER accident sequence. For a transient-induced accident sequence without loss of off site power (LOSP), this requires no operator action. If LOSP is involved, then the operator must start the standby CRD pump, which is powered from a diesel generator bus. 2. The Browns Ferry operating instructions for the condensate system require that the condensate storage tank for each unit be maintained nearly full, i.e., between 1419 and 1325 as (375,000 and 350,000 gal). As shown on Fig. 3.3,' this quantity of water is more than sufficient to maintain reactor vessel level up to the time of calculated containment f ailure by overpressurization (about 35 h). Furthermore, in the unlikely event that the condensate storage tank were at a significantly lower level at the inception of the accident, this tank can be easily refilled from several source s. 3. With the effect of containment backpressure considered in the cal-culations, there would be sufficient NPSH to permit RHR pump operation with suction on the pressure suppression pool at any time during the acci-i dent seque nce. It might be necessary to operate at reduced flow during ) the latter part of the sequence, but,the reduced flow would be sufficient to satisfy the requirements for reactor vessel injection or suppression pool cooling (if the RER heat exchangers were restored to service). 4. The RHR pumps and the core spray pumps can be manually shif tad by the operator at any time to take suction on the condensate storage tank instead of the pressure suppression pool. 5. The RHR pumps can perform their reactor vessel inj ection or pres-sure suppression pool cooling function even with their minimum flow bypass valves f ailed open. {

• In English units, the normal capacities are: HPCI (5,000 spa), RCIC (600 gpa), each of 4 RHR-LPCI mode s (10,000 gpm), each of 2 core spray mode s (3,125 gpe). All of these systems can take suction on the conden-sate storage tank.

4 ...,a, ,,--.,e,. .-m

d 112 6. The cooling loads supplied by the EEG system are automatically reduced if necessary so that the essential cooling loads (including the diesel generatcrs) can be carried by any two of the four available EECW pumps. 7. The SBCS could be used to inject river water into the drywell or wetwell spray headers at any time during the accident sequence. 8. With over 30 h available for remedial action, there is a signifi-cant probability that the PCS could be recovered bef ore containment f ail-ure. 9. Unit 1 suppression pool cooling can be accomplished by certain Unit 2 RHR pumps and heat exchangers, via cross-connecting piping. References for Chanter 9 9.1 S. E. Mays et al., "Interin Reliability Evaluation Program: Analysis of the Browns Ferry, Unit 1, Nuclear Plant," NUREG/CR-2802, BGG-2199, July 1982. 9.2 Browns Ferry Nuclear Plant, Final Safety Analysis Report, Revision 67, Section 4.8.6.4, Tenne ssee Valley Authority. 9.3 L. G. Griemann et al., " Reliability Analysis of Steel Containment Strength," NUREG/CR-2442, June 1982. 9.4 D. H. Cook et al., " Station Blackout at Browns Ferry Unit One - Acci-dont Sequence Analysis," NUREG/CR-2182, ORNL/TM-455/V2 (November 1981). 9.5 S. A. Hodge et al., "SBLOCA Outside Containacnt at Browns Ferry Unit One - Accident Sequence Analysis," NUREG/CR-2672, ORNL/TM-8119/V1, November 1982, Sect. E.3. b ~. .... - -, - - -, ~,,,

113 Table 9.1. Control room indication aint alarms of primary containment variables important to analysis and control of a Loss of DHR accident sequence Variable Range or setpoint Drywell pressure Indication, MPs (psia) &-0.55-(0-80) Alarms, MPa (psig) 0.112 (1.6) 0.113 (1.65) 0.114 (1.75) 0.115 (2.00) Drywell atmosphere temperature Iniication, K (*F) 0-477 (0-400) Alarms, E (*F) 336 (145) Netwell pressure f Indication, MPa (psia) 0-55 (0-80) Alarms, MPs (psis) 0.115 (2) Pressure suppression pool temperature Indicasion, L (*F) 0-477 (0-400) Alarm, I (*F) 308 (95) Pressure suppression pool level" Indication, m (in.)' -0.64 - +0.64 ( +25) Alarm, m (in.) less than -0.15 (-4) e more than +0.15 (+6) "In st rument zero is 4.6 m (15.2 f t) above the bottom of the wetwell torus. A water level of zero indicates that the wetwell is ~1/2 filled with water. l Table 9.2. Arrangement of emergency diesel generator power supplies to the' pumps associated with sach init I heat exchanger ' s s s Unit I heat er[ hanger A B C D Diesel assigned to RHR pump '[ C B D Diesel assigne~d to RHRSW pump s Primary RRRSW. pump g I^'C, B D Eackup RITRSW pump A 3C B 3D q s Note: ' Diesel's denoted A, B, 'C, s.nd D are shared s between Units 1 and 2. Die s'i s ' denoted \\ 3 A, 3 B, 3 C, and 3D are provided for Osi t 3. Diesels A and 3A, B and 3B, e tc.', can be run in perallel. s w } s l s A % >n .n - - - ~ ~ - - - -. - - > ~ < - - - - - = - - - - - = ~ ~ ~ ~ - - - * ~ - = '

• 114 Table 9.3.

Browna Ferry Unit I dominant accident sequences as identified by the IREP study Relative a Frequency IREP nomenclature and de scription g g Initial Final g,,q,,,,p. T R,R : Transient, PCS unavailable, 1.3 x 10-8 9.7 x 10-s 1 U A loss of DER 2 T B: AM S, PCS unavailable 5.1 x 10-s 3,1 x go-s g 3 TRR: LOSP, loss of DBR 1.5 x 10-8 2.4 x 10-s pBA 4 HRB A: ransient, , I ss 1. x 10-s 9,3 x gow DER i 5 T GR R : Same as No. I except RCIC 5.5 x 20-8 4.1 x 10-* g BA unavailable so HPCI used instead 6 T BN: AMS, PCS available, no 3.7 x 10-8 3.7 x 10-8 A recirculation pump trip 7; T IR R : LOSP, SORV, loss of DER 8.3 x 10-s 1.6 x 10-* p BA 8 T 0R R : Same as No.' 3 except RCIC 6.2 x 10-s 1.2 x 10-8 p BA unavailable so HPCI used instead

1. The initial frequency pertains to the probability that the sequence will be initiated. The final frequency takes into account the potential for recovery befo:e the sequence proceeds to core melt. Units are events per reactor year.

Table 9.4. Dominant accident sequences at Browns Ferry considering HPCI and RCIC system use to be equivalent 4 Relative p,q,,,,7 I IREP nomenclature and description II"'I Initial Final f reque ncy 1 TRR

• • ~

gBA U BA available, loss of DER 2 Tf: AU S, PCS unavailable 5.1 x 10-s 5.1 z'10-8 ss of 1. x 10-8 2.9 x 10-s 3 TtR B A. 1, pBA P 4 TRR: Transient, SORV, loss of 1.2 x 10-8 9.3 x 10-* gBA DER 5-T BN: AUS, PCS available, no re-3.7 x 10-* 3.7 x 10-8 g circulation pump trip 6 T KR R : LOSP, 50RV, loss of DHR 8.3 x 10-s 1.6 x 10-8 p B 4

1. Re initial frequency pertains to the probability that the sequence will be initiated. The final frequency takes into account the potential r

for recovery before the sequence proceeds to core melt. Units are events per reactor year. y - - - - - - - - - - ,n , - - ~ - - -m-

115 10. Om4CLUSIONS The major conclusions of this study are itemized in Sect.10.1 of this chapter. A reference follows each of the conclusions, indicating the location within this report where the subj ect is discussed in detail. A brief discussion of the uncertainties pertinent to the analyses is pro-vided in Sect. 10.2. I 10.1 Itemized Conclusions 1. If the pressure suppression pool water is circulated and mixed by the operation of at least one RHR prep during the Loss of DHR accident seque nce, then the. assumption of uniform suppression pool heatup is j usti-fled for calcul.tional purposes (Sects. 2.1 and 3.1). 2. The normal arpply of water in the condensate storage tank is suf-ficient to maintain a normal water level in the reactor vessel throughout the accident sequence during the pericd before prirary containment f ailure by overpressurization (Sects. 3.2 and 3.3, Fig. 3.3). 3. Without. cooling, the suppression pool temperature would increase to 49'C (120'F) af ter 1 h, requiring the operator to begin a controlled manual depressurization of the reactor vessel (Sect. 3.3.3 and Fig. 3.6). 4. Normal reactor vessel water level can be maintained by continuous operation of one CRD hydraulic pump sugnented by periodic operation of the RCIC pomp during the first 4 h of the accident seque nce. Af ter 4 h, the reactor vessel is depressurized and the CRD pump operation alone is suffi-cient to maintain vessel level (Sect. 3.3.1 and Fig. 3.1). 5. No operator action is required to establish and maintain the CRD hydraulic system inj ection rates assumed in this study unless the Loss of DER ~ initiating event includes a LOSP. If a LOSP occurs, then operator 4 i action is required to restore and maintain continuous CRD pump operation (Sect. 9.4.1). 6. The continued availability of the HPCI and RCIC systems would be threatened during the accident sequence by the following automatic control actions: Time (h) Event ~3 HPCI pump suction shifted to overheated pressure suppression pool (Sect. 9.3). ~13 HPCI and RCIC turbine steam supply line s i so-lated because of high torus room temperature i (Sect. 3.3.6). ~14 Containment pressure exceeds RCIC turbine ex-haust high pressure trip setpoint (Table 3.1). 7. Remote-manual SRV operability would be lost af ter aboat 24 h and the reactor vessel would begin a slow process of repressurization. The l reactor vessel would be fully pressurized at the time of containment f ail-ure (Sect. 3.3.2, Fig. 3.4).

116 \\ 8. Containment f ailure pressure would be reached about 35 h af ter l the inception of the accident sequence assuming that the drywell coolers operate for only the first 2 h (Sect. 3.3.4). 9. Sufficient NPSH can be maintained to permit RHR pump operation with suction on the pressure suppression pool throughout the accident so-quence (Sect. 3.3.5). 2 10. If the operator takes action to restore drywell cooler operation af ter the 2 h point, this can have a significant effect on the accident sequence timing. The containment f ailure would be delayed about 2.5 h and l the RHR pump flow would have to be throttled to permit continued operation ~ after about the 14 h point (Sect s. 3.3.5 and 3.3.6). 11. Operation of any one RER heat exchanger in the pressure suppres-sion pool cooling mode would prevent the pool water from reaching exces-sive temperature and would therefore preclude containment f ailure (Sect. 4.2 and Fig. 4.1). 12. The 5.1 cm (2 in.) vent lines from the drywell and wetwell are not large enough to prevent primary containment pressure from increasing l to the f ailure point during an extended Loss of DER accident sequence (Sect. 4.3). 13. For a Loss of DBR accident sequence in which there is no forced circulation of the pressure suppression pool, primary containment f ailure by overpressurization is estimated to occur at the 28 h point. This is seven hours earlier than for the case where assumption of a uniform pool temperature is justified (Chap. 5). 14. The occurence of a SORY would not have a major effect on overall system behavior during a Loss of DER accident sequence. If mixing and uniform suppression pool heatup is assumed, the primary containment fail-uro is advanced just 1 h, to the 34 h point. If there is no forced sup-pression pool circulation then the existence of the SORV delays the con-tainment f ailure by about 4 h, to the 32 h point (Chap. 6). 15. It is very unlikely that the blowdown forces associated with primary containment f ailure would result in degradation of the concrete shield wall surrounding the drywell (Sect. 7.5). 16. Loss of reactor vessel inj ection as a result of over pressuriza-tion f ailure of the containment in a Loss of DER accident sequence might be caused either by (A) loss of all of the vessel inj ection systems 10-cated within the reactor building as a result of the harsh environmental conditions there af ter containment f ailure coupled with an inability to depressurize the reactor vessel to permit use of the low pressure pumps in the turbine building or by (B) loss of the vessel inj ection systems within the reactor building and the occurrence of piping breaks sufficient to render all re'maining vessel inj ection systems ineffective. Case (B) has not been considered in previous PRAs (Sects. 7.7 and 7.8). 17. There are several possible consequences of a f ailure of the con-tainment by overpressurization. There might or might not be a LOCA caused by the disruptive blowdown of the drywell. Sufficient reactor vessel in-jection capability to maintain the core covered might or might not remain. Thus it is unrealistic to assign a 100% probability to the scenario in which all vessel inj ection is lost but no piping breaks occur. The actual probability would be the sum of (A) the product of the probability of the loss of all reactor building inj ection systems because of harsh environ-mental conditions and the probability of failure to regain remote-manual 4 ., - -... -. - ~ ..-m ,-.,e..,, ,,yy ,__.._,,_,,,.,,.mg -._m..-.,_,.-,,-,.-,--_w,..,mm..mw ,__.,nm,%,,e...y---

117 control of even one SRV af ter containment f ailure and (B) the product of j the probability of failure of some injection systems because of harsh on-vironmental conditions and the probability of piping breaks sufficient to render the remaining inj ection systems ineffective (Sects. 7.7 and 7.8). 4 l 18. Lack of consideration of Case (B) of conclusion 16 above can lead to nonconservatism in PRA analyses since the piping breaks could per-mit early release of the volatile fission products into an already-failed e drywell (i.e., the fission product scrubbing function of the pressure sup-pression pool would be bypassed) (Sects. 7.7 and 7.8). 19. If all reactor vessel injection is assaned lost at the time the primary containment fails and it is assumed that there is no LOCA, then core uncovery will occur af ter boiloff of the volume of water above the core at the time of drywell failure. MARCH runs indicate that the core would uncover about 21/2 h af ter loss of inj ection and that core melting would begin about I h later (Chap. 8). 20. During the latter stages of a Loss of DER accident sequence, the drywell and wetwell pressure indication and the pressure suppression pool level indication would be off-scale high. The drywell and pressure sup-pression pool temperature indication would remain onscale and the primary containment pressure could be inferred from the suppression pool tempera-ture and the steam tables (Sect. 9.1). 21. The automatic shif t of the HPCI pump suction from the condensate storage tank to the pressere suppression pool on high sensed suppression pool level should be removed or modified to permit the operator to return j the suction to the condensate stcrage tank if necessary (Sect. 9.3). 22. The conclusions of the IREP study regarding the probability of core melt at Browns Ferry unit 1 as a consequence of Loss of DER should be reconsidered, based upon the better accident sequence definition provided by the detailed analysis presented in this report (Sect. 9.4). 10.2 Uncertainties in the Analysis The calculation of accident sequence events before containment fail-are was performed using the ORNL-developed BWR.-LACP code which incorpor-l ates reactor vessel, primary containment, and secondary containment models specific to Browns Ferry Unit 1. The BWR-LACP code was also used in two previous SASA studies; additions made to the code for the Loss of DER ac-cident sequence calculations are described in Appendix B of this report. Code results for a Station Blackout accident sequence have been compared to results calculated for the same sequence by the Browns Ferry simulator and RELAP4 Mod 7.1*

• 1 Code results for a small-break LOCA with condensate l

booster pump injection have been compared with results calculated for the same sequence by RELAP5 Mod 1.1' 2 Agreement was good in all cases. It should be noted that primary system calculations for the portion of a severe accident sequence before core uncovery are much simpler for a BWR than for a PWR. The MSIVs are shut during a severe accident sequence and the reactor vessel is isolated. In general, the recirculation pumps are tripped and core flow is solely due to natural convection circuits within the reactor vessel itself. Therefore, for other than large-break LOCA or ATWS studies, sophisticated primary system analysis codes such as RI! LAP, RETRAN, or TRAC are not necessary; fundamental modeling of the l ,. ~. _. _._..,.-._-..,,____-,.__._,-~,_..___m___.,,_,,__m_

118 processes within the reactor vessel in a relatively simple code such as BTR-LACP is suf ficient. On the other hand, the interaction between the reactor vessel and its relatively maali primary containment is very important to the determina-tion of the sequence of events for a BWR severe accident. In this regard, BWR-LACP is efficient because it combines primary system, primary contain-ment, and secondary containment analyses in one code. There is no need to convert the output of one code into the input for another with the atten-dant oppostunity for error. BWR-LACP is specific to the Browns Ferry MK I containment system and is therefore a straightforward application of basic thermohydraulic and heat transfer theory. The uncertainty in the results presented in Chapters 3 through 6 caused by modeling inaccuracies is be-lieved to be negligible. Uncertainties do exist in the input parameters supplied to the BWR-LACP code for the study of the Loss of DER accident sequence before core uncovery. These include: 1. The primsry sy stem events during the very brict' period (~1 min) af ter scram and MSIV closure when multiple SRVs are open and the feedwater turbines are cotsting down can not be modeled by BWR-LACP. Normal reactor vessel indicated water level is 561 in, above vessel zero and the BWR-LACP calculaticus are begun at time 20 s with a water level equivalent to an indicated 500 in. in consideration of the ef fect of level shrink upon MSIV closure. This assumption is based upon the accident stadf es presented in Chap. 14 of the Browns Ferry FSAE and upon the level indications at the Browns Ferry simulator when scram and MSIV closure are simuisted. Since reactor vessel water level is subsequently controlled, the uncertainty in the brief period just after accident initiation is not believed to be in-portant. 2. It has been assumed in this study that the only coolant loss from the reactor vessel is through the SRVs to the T quenchers in the pressure suppression pool or as a driving force to the RCIC or HPCI turbines.* In fact, there would be a slight leakage (less than 25 spm) into the drywell, i and a slight leakage through the MSIVs into the main condensers. The amount of leakage is uncertain, and has been neglected in the analysis. 3. As the pressure suppression pool is heated, evaporation from the water surf ace tends to increase the volume of steam in the primary con-l tainment atmosphere and consequently the pressure. Leakage f rom the pri-l mary containment has been modeled as equivalent to that measured during the most recent containment integrated leak rate tests [ conducted at 0.274 MPa (25 psig)], adjusted for different containment pressures.t It is entirely possible that the, leakage paths f rom the primary con-tainment would both enlarge and become more numerous as the internal pres-l sure increased above design pressure and approached the f ailure level. For this study, it has been assumed that no new or enlarged leakage paths develop. This approach is conservative and produces the earliest cata-strophic failure of the drywell. If enough additional drywell leakage paths did develop to permit the escape of suf ficient steam to maintain the primary containment pressure constant at some level below failure pres-sure, then reactor vessel injection capability would not be threatened and a Severe Accident with core uncovery would not occur. I

• Which also exhaust to the pressure suppression pool.

tSee Sect. B.2 of Ref. 10.3. g e -, w-,vr-----ww- ,w,-- ww .rnnnre--wc--=a-e- -,---,:,---p-,-w,--n ,-,,,e.--e,~.,er-- a,---we.-wr,-,-o,-w--- ,, -,-m-,m-- --wm,--------e- - -ow w----wom

119 4. As discussed in Chap. 3, operation of the drywell coolers has a significant effect upon the results of this study. It is uncertain what I action the operators might take to restore the drywell coolers -once they i have been automatically tripped as a consequence of a LOSP and a core spray initiation signal. Both the case of no restoration and the case of Lamediate restoration of drywell cooling have been included in the analy- ~* ses. 5. The rate of CED hydraulic system inj ection into the reactor ves-sel is not known with certainty. The system employs centrifugal pumps and l with a scram in effect, the inj ection is primarily leakage past the gra-l ' phitar seals in the CRD mechanism assemblies. The Browns Ferry FSAR esti-l mate s the inj ection to be 182 gym with the reactor vessel depressurized and 113 gym with the reactor vessel pressurized for one pump in operation j with the normal system lineup for the case of a scram in effect. A raxi-mum inj ection of 170 gym was used for the analysis. ( As shown on Figs. 3.2 and 3.4, the reactor vessel is depressurized during most of the early part of the accident se qne nce. During the first i four hours before the vessel is depressurized the operator is able to con-trol vessel level by running the RCIC system in conj unction with the CRD hydraulic system and if the actual available CRD hydraulic system inj ec-l tion varied from that assumed in the analysis then the operator could ad-just for the difference by running the RCIC pump for longer (or shorter) periods. In the latter part of the accident sequence the vessel is re-pressurized, but by this time the decay heat is low enough so that the op-erator throttles the flow to less than 113 spm. In summary, since the l-rate of CED hydraulic system inj ection is an operator-controlled variable, I uncertainty in the maximum available flow at various times during the se-quence doe s not significantly affect the analysis. 6. The containment f ailure pressure and the f ailure location are uncertain. A value of 0.910 MPs (117 psig) for f ailure at the cylinder-sphere intersection in the drywell was assumed for this study, based on the information presented in Ref.10.4. 7. The calculations of Chapts. 3, 5, and 6 assumed that pressure fluctuations in the pressure suppression pool during SRV discharge would be mild, and that (except as noted in Sect. 5.3.3) 100% of the discharge, would be condensed providing that the pool remained as much as 1.11*C (2*F) subcooled. The assumption of mild pressure fluctuations during SRV j discharge is realistic because the T quencher was specifically designed and extensively tested to e1Luinate the violent condensation oscillations possible with the formerly employed rans-head (the T quencher has many very small perforations over a long, 30-cm dian, submerged pipe, whereas the rans-head has only two 25-cm diam discharge stubs). The uncertainty to be examined here concerns the amount of subcooling required for complete condensation of SRV discharge. The range of experi-mentally proven quencher performance as a function of pool local subcool-ing (difference between saturation temperature defined at quencher depth l and pool temperature) is summarized on Fig. 9 of NUREG-0783. These data can be used to define the minimum subcooling required for 100% condensa-tion: 11.1*C (20*F) for mass fluxes below 206 kg/s/m8 (42 lb/s/f t*), cor-responding to SRV discharge when reactor vessel pressure is low, and 22.2*C (40*F) for mass fluxes exceeding 460 kg/s/m (94 lb/s/f,t8), corre-2 sponding to SRV discharge when reactor vessel pressure is high. t I _ _ _ _ _ _ _,.,.. _ _ _ _ _ _. _ ~ _... _ _ _ _. _ _ _ _. _, _ _ _

120 In the loss of DHR sequences, a subcooling requirement can be viewed as a requirement that saturation temperature at total suppression pool pressure must exceed the pool water temperature by the specified margin. If primary containment pressure is too low there will be incomplete con-densation and pressure will increase until subcooling approaches that re-quired fer complete condensation. At the primary containment failure pressure of 0.908 MPs (117 psig), the saturatio'n temperature is 176*C (348'F). The suppression pool temperature at containment f ailure would be approximately equal to 176*C (348'F) less the assumed pool subcooling re-quirement. As summarized on Table 10.1, the ef fect on containment failure time of the stricter pool subcooling requirements specified in the pre-vious paragraph would be to shorten the reported times to 28 h. 8. The containment failure mechanism assumed in this report is fail-ure by overpressure at 0.908 MPa (111 psig). Since the drywell atmosphere temperature exceeds the 138.3*C (281*F) design temperature of the drywell before the time of overpressure failure, ft is appropriate *o consider in this section the possibility of containment failure due to excessive ten-perature. Dryvell components, including electrical penetrations, must be able to withstand temperatures in excess of design temperaturs fcr limited periods. If a f ailure temperature of If 2.8'C (325'F) is assnued for the drywell ele:trical penetrations, then the time to containannt failure would be 26 h about 2 h shorter than the shortest overpressure failure time considered in this report.

• 9.

There is uncertainty regarding ths sequence of events during the first few hours of the accident. The results presented in Chap. 3 were obtained with the assumption that the operator would not open the 5.1 cm (2 in.) vents from the drywell and wetwell and that the CED hydraulic sys-tem inj ection would increase when a high drywell pressure scram occurred at 0.115 MPa (2 psig). The wetwell-to-drywell differential pressure com-pressor, which maintains the drywell pressure about 0.008 MPa (1.1 psi) higher than the wetwell pressure,* was modeled for automatic actuation. In their review of this report, the Tennessee Valley Authority (TVA) reviewers pointed out that the setpoint for the high drywell pressure scram had been increased to 0.119 MPa (2.5 psis) and that the differential pressure compressor is operated in the manual mode so that operator action is required. They also suggested that the operator would open the 5.1 cm (2 in.) vents when the primary containment pressure reached 0.115 MPs (2 psig) in an effort to forestall the high drywell pressure scram. To determine the ef fect of the uncertainties revealed by the TVA com-monts, an additional BWR-LACP calculation was performed in which it was 4 assumed that: (a) The differential pressure compressor was not operated so that i the flow from the wetwell airspace to the drywell was only j through the vacuum breakers, which are open only when the wet-well pressure exceeds the drywell pressure by at least 0.003 MPa (0.5 psi).

• The purpose is to keep the downconers in the pressure suppression pool almost totally free of water. The compressor and associated piping are shown on Fig. 4.2.

i l

-121 (b) The 5.1 cm (2 in.) vents were opened when the primary contain-ment pressure reached 0.115 MPs (2 psig). (c) The high drywell pressure scram setpoint was 0.119 MPa (2.5 psig). The calculation indicates that the high drywell pressure scram would 1 occur at time 3.8 h under the new assumptions as opposed to time 1 h as reported in Chap. 3. Although this delays the increase in CRD hydraulic pump inj ection, the RCIC system remains available fo. inj ection throughout this period so there is no effect on the ability to control reactor vessel l 1evel. The opening of the primary containment vents delays the time of primary containment f ailure from 35 h as reported in Chap. 3 to 40.5 h. 10. The Reactor Dater Clean-up (RWCU) system operates continuously during normal reactor operation, removing impurities from the primary coolant, and also removing a small quattity of heat (about 4.5 MW if the reactor is at f ull pressure and temperature). Although it is possible that tLe RWCU system could continue to operate during the Loss of DER so-quence, the analyse s of Chapters 3, 5, ar.3 6 tock no credit for heat re-moval by the RWCU system. When the reactor coolant sytten is depressur-ized and therefore at a lower temperature, tbo rate of heat removal by the RWCU oystem is only about 2.5 MW. Iloweve r, this is 10% of the rate of beat generation by decay heat 12 h af ter reactor scram. It is estimated that if the RWCU system wore operated continuously throughout the 35 h j Loss of DhR sequence, the ultimate containment failure would be delayed by about 10% (i.e. 4 h). The primary reason for assuming no heat removal by the RWCU system is that one can envision circumstances in which it would not be operated in a Loss of DHR accident sequence. In order for the RWCU system to remove heat from the primary coolant system, both the Reactor Building Closed Cooling Water (RBCCW) system and either the Raw Service Water (RSW) system l or the Emergency Equipment Cooling Water system must be operational. The RBCCW system supplies cooling water to the RWCU non-regenerative heat ex-changer via its non-essential loop, which is automatically isolated in the event of a loss of offsite power. After a successful start of the station diesels, the RBCCW non-essential loop isolation valve can be opened by the operators; however, the plant operators might decide to allow the valve to l remain closed in compliance with the non-essential category to which this cooling load has been designated. The MARCH code was used for calculation of the results presented in Chap. 8 of this report. The general limitations of this code with res-pect to LWR accident analysis with emphasis on PWR applications have been discussed elsewhere.18. s However, the specific limitations with regard to l BWR analysis are even more confining. Many of the difficulties with re-spect to the use of MARCH for BWR accident analysis are discussed in Ap-pendix B of Ref.10.6. Some of these difficulties have been bypassed in the present study by initiating the MARCH analysis at a time just preceed-ing containment failure. However, other MARCH code limitations do have an effect on the current study. These include: 1. The core model does not represent the Zircaloy channel boxes and the control rods present in a BWR core. 2. Reactor vessel pressure control by the SRVs is not correctly rep-l resented in that a continuous release of steam from the reactor vessel is l l l I._._-_

122 modeled instead of the actual periodic blowdowns followed by relatively quiescent periods within the reactor vessel between relief valve lif ts. Both of these limitations have a significant ef fect on the calculation of the metal-water reaction rate and consequently, on the core heatup siter partial uncovery. Since ;the BWR-LACP code has been used to establish the approximate time of containment failure for this study, the chief function of the MARCH code has been to provide a basis for examination of _ the possible effects of containment failure on reactor vessel injection capabilities and to, establish the timing of the events after containment failure for the case in which all reactor vessel injection is assumed lost. As dis-cussed in Chap. 8, the MARCH code predicts core uncovery 2-1/2 h af ter containment f ailure and loss of injection followed I h later by core melt-ing. With the code's present limitations, these can be considered to be no more than reasonable approximations to ths actual timing. A fissica product transport analysis is underway as a follow-on tc the accidsnt sequence analysit presented in this report. Ongoing modifi-cations to the MARCH code sro intended to remove the more significart limitations of this code in time fer an improved version to he avc11able for, support of the fission prcduct trarsport analysis. S s ,,,.n.

---

r-.-.

-, ~, - - - .,n ,,,. _., -m..r +--.~ -

123 References for Chanter 10 10.1 R. M. Harrington, " Comparison of Station Blackout Calculations on BWR-LACP, TVA Browns Ferry Simulator, and RELAP IV Mod 7," letter report to NRC SASA Program Technical Monitor, August, 1981. 10.2 W. C. Jouse and R. R. Schultz, "A RELAPS Analysis of a Break in the Scram Discharge Volume at the Browns Ferry Unit One Plant," EGG-NTAP-5993, August,1982 (Also published as Appendix G to NUREG/CR-2672, ORNL/TM-8119/V1). 10.3 S. A. Hodge et al., " Station Blackout at Browns Ferry Unit One - Iodine and Noble Gas Distribution and Release," NUREG/CR-2182 Vol. 2, ORNL/NUREG/TX-4$5/V2, August 1982. 19.4 L. G. Greitana et al., "Raliabilf ty An,1ytes of Steel Ccutainnent St rength," MIEEG/CE-2442 (June 1982), i 10.5 J. B. Rivard et al., "1uscrim Technical Assensuent of the MARG Code," NUR3G/Gr2285, SAND 81-1672.R3. November 1981. 10.6 S. R. Greene et al., "S310CA Outsida Contaimnent at Browns Ferry Unit Cne - Accident Sequence Analysis," NUREG/CR-2672, V<>1.1. ORNL/TM-8119/V1, November 1932. e i 9 e-y --g 3 -+ y p

---

ty-wrr-

-w--ew-'-*--w r v' -m-7 y --m--- --w,-W--

124 Table 10.1. Effect of stricter suppression pool subcooling requirement on time to containment failure" Containment Nominal assumed Reported f ailure time 3,,,,,, g, ) , containment for stricter subcooling discussed requirement failure time subcooling g,,,, g,, 'C (*F) (h) requirement (h) 3 1.11 (2) 35 28 5 1.11 (2) 28 28 22.2 (40) 6 1.11 (2) 32 28 "The strictor reycirnreuss are: 11.1

• C (20*F) foz 1"w mass fluz and 22.2*C (40*F) for h1 h mass fluz.

5 F 4 e e e 4 --.--..-,- n., -, -,,., -,, - - -. ,----..,c, ....,..n---

125 Appendix A. DESCRIPTION OF THE BROWNS PERRY UNIT ONE RESIDUAL HEAT REMOVAL SYSTEM A.1 Introduction The residual heat removal (RHR) system at Browns Ferry Unit 1 com-prises four pumps and heat exchangers arranged in two basic piping loops. 2 The maj or piping associated with components B and D is shown in Fig. A.1; the piping for loops A and C is similar. The RER system valves on this figure are shown in their normal positions during reactor operation. The three basic operating modes for the RHR system are: 1. Low pressure coolant injection (LPCI), 2. Primary containment cooling, and 3. Reactor vessel shutdown coolir.g. It is the purpose of this apperdix to describe the operatien of the EHR system in ecch of its modes in sufficient detail tc provide the background necessary to an understanding of the natorial discussed f u the main body of this report. Discussions of the design and areavailableinmuchgreaterdetailelseshere.gpgrgtjonofthissystem i The LPCI operational mode is an ECCS mode provided f or ese ir. the event of a design basis accident. As such, i t would not be utilized in a loss of DER accident sequence Taless the accident initiator included a loss of coolant accident (LOCA). The LPCI mode of the RHR system is de-scribed in Sect. A.2. Thc primary containment cooling mode of the RHR system is utilized for pressure suppression pool cooling as well as for drywell or wetwell spray. This operational mode is discussed in Sect. A.3. The reactor vessel shutdown cooling mode is used af ter reactor shut-down and vessel depressurization for long-term decay heat removal. Sec-tion A.4 provides a brief description of this operating mode, which is not available by definition in a Loss of DHR accident sequence. Other features of the RHR system include the capability to pump con-densate storage tank water or to channel river water through some of the piping. These and other special system features are discussed in Sect. I A.S. A.2 LPCI Operational Mode l l Referring to Fig. A.1, in the LPCI operational mode RHR pumps B and D take suction on the pressure suppression pool ring header through valves 74-24 and 74-35, respectively. Each punp's discharge is routed through the associated heat exchanger and thence into a common 0.61 m (24 in.) line leading through valves 74-66 and 74-67 and check valve 74-68 into the reactor vessel via the piping on the discharge side of recirculation pump A. The arrangement for RHR pumps A and C is similar. I i l y-- m.,-r -,,y-,we---e, y----. r,---w + ,+r---~s rr ee e mw w e-

---

=--=er-

+ - - * - - - *- * **- - - =-+ - - - - * - - - - -- - - ' ' - * - -

• 126 i

The LPCI operational mode is autanatically initiated if the water level in the reactor vessel drops to 9.77 m (384.5 in.)* above the bottom of the vessel or if high drywell pressure [0.115 MPa (2 psig)] exists in conjunction with a low reactor vessel pressure [less than 3.20 MPa (450 psig)]. The RER pumps start immediately upon an automatic initiation signal, and valves 74-66 and 74-67 automatically open when reactor vessel pressure is less than 3.20 MPs (450 psig). The pumps are protected by a minimum flow line to the pressure suppression pool (not shown in Fig. 2 A.1). The purpose of the LPCI operational mode is to restore and maintain the reactor vessel coolant inventory af ter a loss of coolant accident. Water pumped into the reactor vessel would spill from the piping break into the drywell and flow from there back into the suppression pool; this establishes a closed cycle f or the flow. Cooling water for the secondary side of the RPR heat exchangers is provided by the RHR service water system. (11e portion of this system f i that providea cooling water to the D heat exchantor is represented by the dashed times in the lower right corner of Fis. A.1). However, service water flow to the RER heat exchangers is not required immadiagely af ter a [ LOCA and there is no provision for autoratic actuation of the service i watar flor to the RHE heat exchanterr. l I The valve configuration shown in Fip. A.1 supports the LPCI op2xa-tional mode and is the normal lineur of the RHR system. F2rtiermore, all l 2HR system motor-oporated valves will automatically realign to the con-figuration shown in Fig. A.1 if they should be in another alignmert when a LPCI initiation signals is sensed.t As previously discussed, valves 74-66 and 74-67 will open only when the reactor vessel pressure is less than 3.20 MPs (450 psig). i With a rated flow of 0.631 m /s (10,000 gym) per pump, the LPCI mode 8 ) provides a means to rapidly recover the core following a design basis accident. Over the long term, however, the decay heat would have to be removed f rom the closed cycle. This can be accomplished by continuing to operate one loop in the LPCI mode while switching the othe,r loop to the 1 suppression pool cooling mode, which is described in Sect. A.3. 1 1 I A.3 Primary Containment Coolina Operational Mode In the suppression pool cooling mode, the RHR pumps take suction on the suppression pool ring header as in the LPCI mode, but valves 74-66 and 74-67 are shut and the pump discharge passes into the suppression pool through valves 74-71 and 74-73. The RER flow through the heat exchangers is cooled by river water circulated by the RHR service water (RHRSW) pumps. As shown in the lower right corner of Fig. A.1, RHRSW pump D2 i normally serves RHR heat exchanger D with return to the river through 'This is about 0.61 m (2 ft) above the top of the active fuel in the core. Since the high drywell temperature will cause a decrease in the water density in the reference leg of the level instruments, the actual water level would be 10.34 m (407 in.). i TThe one exception if the case where the RHR system is aligned for shutdown cooling. See the discussion in Sect. A.4. y ,---.v._.,...-,.--w.-,-,_.--,m_.,,--,.-.wmy, ,.,....ym. m.,,... s.m -. - - ~----- m

4 127 valve 23-52. (To avoid unnecessary clutter, the RHRSW piping for RHR heat exchanger B is not shown.) Manual action by the operator is necessary in order to place either loop of the RHR system in the pressure suppression pool cooling mode. If a LPCI initiation signal has occurred, an interlock prevents closing valve 74-66 natil at least five minutes has elapsed. Valves 74-71 and 74-73 can not be opened anless the reactor core is at least 2/3 covered with water. The operator must also manually initiate the flow of RER service water to the RHR heat exchangers. However, the RER system can be operated in the suppression pool cooling alignment to circulate and mix the suppression pool water even if RER service water is not available for cooling. One modification of the primary containment cooling mode involves di-version of about 5% of the pressure suppression pool cooling flow through valve 74-72 to a single ring header

• located in the uppermost portion of the airspace in the wetwell ab,ve the pressure suppressio'n pool. As in the case of valves 74-71 and 74-73, valve 74-72 is interlocked closed I

unless the reactor core is at least 2/3 covered. In addition, valves 74-71 and 74-72 are intsrlooked so that both can not be simultaneous 1, i open unless a LPCT initiation signal is present. Sicco e LPCI initiatior signal autaasticsily realigns all rotor-operated valves on the disc' arge a j side of the RHR pumps to the configuratica shown in Fig. A.1, an operator l desiring to spray the werwall airspace must first g61n control of the RH2 system valves by piscing a control roas selector switch is "asanal". If necessary, the esquiremest tha t iho reactor core be 2/3 covered can be i over-ridden by a key!ock bypass switch. A second modification of ths primary containment cooling mode in-volves diversion of some or all of the pressure suppression pool cooling flow to a drywell spray headert via valves 74-74 and 74-75. These two valves can be opened simultaneously only if a LPCI initiation signal is I pre se nt, the reactor core is at least 2/3 covered, and drywell pressure is at least 0.108 MPa (1 psig).i The operator gains control of the valves by placing a selector switch in the " manual" position; if necessary, the requirement that the reactor core be at least 2/3 covered can be over-l ridden by a keylock bypass switch. A.4 Reactor Vessel Shutdown Coolina Mode i i During a normal reactor shutdown and cooldown, the reactor vessel is j depressurized by steaming to the main condenser at a cooldown rate of less than 37.8 K/h (100*F/h) until the reactor vessel pressure is less than 1.03 MPa (135 psig). Then suppression pool suction valves 74-24 and 74-35 are shut 5 and shutdown cooling suction valves 74-47 and 74-48 are opened. I

• This ring header is served by both RHR loops.

tThere are two drywell spray headers, each fed from one of the RHR loops. I IIhis protects against vacuan-induced internal collapse of the pri-mary containment. $ Assuming the portion of the RHR system shown in Fig. A.1 is to be l used. A lineup ~similar to that described in this subsection could be accomplished using pumps A and C. l \\ . - -... -- ------ ----- - - -- - - - --.~-.....,..--~.. - -..- - --

128 The flow induced by one RHR pump is sufficient for the purposes of shutdown cooling, and the operator opens valve 74-25 if RER pump B is so-1ected.* RER service water flow is established to the selected heat ex-changer. The selected pump is started and valve 74-66 is throttled as necessary to maintain the desired reactor vessel cooldown rate. j If a LPCI initiation signal should occur when the RHR system is oper-ating in the shutdown cooling mode, the operating RER pump will trip but ths suction valves will not automatically realign. Thus, it would be nec-essary for the operator to shut valves 74-47 and 74-48, reopen valves 74-25 and 74-36, and start the desired RER pump or pumps if LPCI inj ection were desired. In this connection, it should be recalled that the RHR sys-ten would not be operated in the ' shutdown cooling mode unless the reactor vessel is at low pressure so that the probability of a large-break LOCA is remote. A.5 Spoph]. RER '.v a tes Fe s tang 1 I A.5.1 Inle m on of k d epa.elsAta s to r_ Lee t ank wa tn The sup;rossi ca pool section valves "s4-24 and 74-35 can be shut at sny time and valves 74-34 and 74-45 can La opened to permit RRR pump suo-tion to be taken on tLe condensate storage tank. This operator action I wonid be desirablo 11 the RH2 puaps were to be operated during the latter stage s of a loss of DER sccident sequence when the suppression pool il temperature./pressurn profile signt not supply the required not positive suction head to the RER ptsps. A.S.2 Standby coolant sunniv l The Unit 1 RER loop comprising the B and D pumps and heat exchangers includes provision for the passage of river water into the reactor vessel or into the containment spray headers by virtue of a connection to the RER service water system. With reference to Fig. A.1, it can be seen that with valves 23-57 and 74-101 open and valve 23-52 shut, river water can be inj ected into the RHR system downstream of the heat exchangers and flow l from there into whatever RHR system discharge path is available.t The RHRSW pumps have a rated flow of 0.205 a /s (3250 sys) and a shutoff head s of 1.22 MPs (162 psig). Thus the standby coolant supply feature can sup-ply reactor core coverage and cooling if the reactor vessel is depressur-ized o'r containment sprays at any time before containment f ailure by over-I pre s suriz ation.

• or valve 74-36 if RHR pump D is selected.

+ l tFor Unit 1, the only cross-connection between the RHRSW and the RHR system is that shown on Fig. A.1 (i.e., there is no similar arrangement in the piping loop containing pumps A and C.) l i i i m__ ,___.m_.--m__--,,7 ...._-._-,,,_y., ,_-,w..-... y .._,..m., .- w ...~,,. - - _,. --- :

129 A.5.3 Head sorav 4 At the latter stage of reactor vessel shutdown cooling, it becomes desirable to completely fill the reactor vessel with water in preparation for refueling. An RHR system connection to a spray nozzle in the reactor vessel upper head is provided for this purpose (not shown in Fig. A.1). The spray acts to condense the steam in the reactor vessel apper head and thereby facilitates flooding of the reactor vessel. A.5.4 Fuel storane cool coolina Connections not shown on Fig. A.1 can be used to augment fuel storage pool cooling. This might be necessary if all of the fuel in the reactor vessel had to be unloaded under emergency conditions. A,5.5 _Cre a s-c onne c t ion: to unit 2 Provision har neon made et the Brownr Ferry Nuclear Plant fer cross-i cor.nections between the individual kHE systems of each of the three units. The parrose is to maintein a long-torm resctor ve.sse) rad pressure sur-pression pool cooling capsbility which does not dspend on the integrity of I the primary containmoet or the operability of the RHH systen associaird with a partictier uni t. The entire cross-connection network is 111ns-trated in Fig. 4.8-1 of 1:af. A.2. The suctions of RER pumps 2 or D in Unit 1 can be crost-connected with the sucticus of RHR pumps A or C in Unit 2. The common discharge line from heat exchangers B and D in Unit I can be cross-connected with the common discharge from heat exchangers A and C in Unit 2 through valves 74-101 and 74-100 shown in Fig. A.1.

• Thus water fram the reactor vessel or pressure suppression pool of Unit 1 can be circulated through the A or C heat exchangers in Unit 2 and returned to the source in Unit 1 if the Unit 1 RHR pumps or heat exchangers are not functional.

Siwilarly, the B l or D RHR pumps on Unit I can be used to circulate Unit 2 pressure suppres-sion pool or reactor vessel water through the B or D heat exchangers on Unit 1. l An arrangement similar to that described above for Units 1 and 2 exists between the B and D RER pumps and heat exchangers of Unit 2 and the l A and C RHR pumps and heat exchangers of Unit 3. The piping in the crosstie network is sized for a minimum flow of 0.315 m /s (5000 spa) whereas under normal conditions full heat exchanger 8 primary side flow is 0.630 a /s (10000 spa). With the lower flow, the s i j heat transfer capabgljty of a RHR heat exchanger is about 91% of the cap-acity at full flow. The cross-connection piping can also be used for not transfer of water between units by leaving the RHR pump suction valves of one unit in normal alignment while opening the heat exchanger cross-connection to the adjacent unit. In this manner, pressure suppression pool water from the first unit, which has been cooled by the first units' heat exchangers, can

• For sireplicity, these cross-connections have not been shown on Fig.

A.1. i ~.

130 be discharged into the second units' suppression pool, or used to flood the reactor vessel or spray the drywell or wetwell airspace of the second unit. References for Annendix A A.1 Systems Manual - Boiling Water Reactors, Inspection and Enforcent:nt Training Center, U.S. Nuclear Regulatory Commission, Sect. 10.6. A.2 Browns Ferry Final Safety Analysis Report, Sect. 4.8. I J s e i e 9 e n..- -..n, ,,.,,,.-m-,. _. - -,.,. ~., -, - -.,.......,..,., - -,

131 ORNL-DWG 8 2 -19 310 e MACTOR WSSEL () tr 74__75 74-74 O [ 450 PStG 7**M \\ 24* 18" 24" EL EF T CLEAN RADWASTE db74-66 [74-67 A 74-49 [_ REC.RCULAfl W vy4 73 y I A C' 74-t41 ,' f r*4-43 74-72 i N 4 24" 74 73 f-p. [7447 ~~ $UPPRES$10N POOL RING E ADER RHR w 0 pm gat 74-24 24* [ W EXCMANGER f -+ B -D< U w o o 74-25 74-34 RMR 16" FROW CCHOENSATE STORAGE TANK h ,' 0" A ADO C 1r74-36 k 74 -45 l db d Rm KAT ( W EXCHANGER e l74-101 p< D 74-35 l l Rm = o p _ _ _, _ _,...,1 . 23-s7 74-ioO 2

I233,

. r + EECwS 02 @ oi @ 03 @ = RHR SYSTEM l PUMPS B AND D l RIVER RIVER Fig. A.1. Unit 1 RHR system. n

133 Appendix B NODIFICATIONS 10 BWR-LACP FOR THIS STUDY The BWR-LACP code was used to perform the calculations discussed in Chaps. 3, 4, 5, and 6. Modifications to the code that were necessary for the Loss of DHR accident sequences are described in this appendix. The BWR-LACP calculations for Chaps. 5 and 6 also used a special pool model, described in Appendix C, which is able to calculate the thermal stratifi-cation effect which occurs when the pool is heated by the SRVs without RHR flow to mix the pogly ghg BWR-LACP code has been described in previous ORNL SASA reports. B.1 RHR Hest Exchanaers The RHR system accomplishes pool cooling by circulating water from i the suppression pool, through the shell side of the RER heat sachangers, and back to the pool. The RHE Service Water (13RSW) is pumped in an open cycls f rom the river thro sgh the tube side and back-to the river. Tae rate of heat exchange is calculated by utilizing the ef fectivo-noss formulation: Q=ECmin (Th,in 6,in) -T

where, Q = rate of heat transfer from pool water to river water, E = heat exchanger effectiveness, C,g, = the maaller of the tube side and the shell side mass flow times specific heat products, T',in = inlet temperature of the hotter fluid as it enters the heat h

exchanger, and T,,g,= inlet tempe:ature of the colder fluid (i.e., the river water). Each of the RHR heat exchangers has two tube passes and one shell pass. The formula for ef fectiveness of this arrangement is:B.3 r (1+e*)/ (1-e*)] E = 2/[(1+C ) + 1+C j r l l

where, C, = C,g, (defined above) divided by C,,,

(the larger of the shell side and tube side mass flow rate

• specific heat products, x = UA( $f 1+ C,) /C,g,,

2 U = overall heat transfer coefficient, based on the total heat transfer area, A.

134 The total heat transfer coefficient, U, is dominated by the thermal resis-ances;gf tho tube metal and by the shell-side and tube-side fouling allow-tance

• 4 therefore, the effectiveness is not sensitive to fluid tempera-tures and there is no need for an iterative solution.

B.2 Puno Net Positive Suction Head (NPSH) In general, pump NPSH can be expressed as H = (P - P )/W + Vs/23 npsh s y t'

where, P, u static pressure at the centerline of the pump suction inlet pips vapor pressure of the fluid being pumped i

P u y W = waight density of fluid being pumped V = vclocity of fluid in the pump suction inlet g = acceleration of gravity 1 For the care of the ECC5 pumps taking succion on the suppretsion pool, the 7 formula, above, is equivalent to the following: H = AH + (Z -Z ) - H. nysh c np1 ps 4 e

where, AH, = a combination of terms which is the same for all pumps which take suction on the pool po)/12

= (P - P )/W + (L -L tspa y p P = t tal Pressure of the wetwell atmosphere ts a = measured level of the pool (in. from instrument zero) P L, = normal measured level of pool (i. e., 4 in, below instrument zero) Z = normal elevation of the pool surf ace (i.e., elevation when "E pool level is 4 in. below instrument zero) Z = elevation at centerline of pump suction k = head loss be tween pool and pump suction inlet i For each of the pumps that can take suction on the suppression pool, the specific fonsula for NPSH (f t) is based on information provided by the TVA for Browns Ferry Unit 1: rhr/10,000.0)2 H-NPSH = AH + 14.5 - (N + 0.9) (B rhr c rhri H-NPSH = AH + 14.8 - 1.03(N + 0.83)(B /3125.0) cs c cs1 cs t I _. _ _ _ _ _, 1

135 H-NPSH = AH, + 12.0 - 6.5 (Bhyci /5,000) hpci rcic/600)2 H-NPSH = AH + 12.0 - 6.5(B reic c where N = number of RHR pumps running per loop (there are two pumps in rhri each loop and they share a length of suction piping), B = flow Per RHR pump (expressed in spa) rhr N,,7 = number of core spray pumps in each loop (there are two core spray pumps in each loop and they share a length of suction P Ping) i j B,, = flow per core Spray pump (spm) 3 = HPCI pump flow (spm) 3p,g L,,,, = RCIC pump flow (gpm) I B.3 RIvwell Ccoltts The drywell coolers transf er heat f rom the drywell atmosphere tc the reactor building closed croling water (RBCCS) system. TSc NBCCW system pusps water through the inside of the heat exchtnger tube s; the cooler's i blowers draw the drywell air (i.e. nitrog:n and water vapor) actors the exterior of the tune s. Air-side hest transfer is enhanced by the many closely spaced parallel copper sheets which are attached to the outside of the tube s. The calculation of heat transfer within the drywell coolers is simi-lar to that for the RHR heat exchangers in that the effectiveness formula-tion is used; however, an iterative solution is employed beause the heat transfer properties of the air-side are radically altered when the water vapor fraction becomes significant (af ter 8 h, or more, into the loss of l DER accident sequence). The equation for counterflow heat exchanger effectiveness is:B.3 E = (1 e *)/(1 - C e *) ~ ~ r where C =Cmin/C r max C = maaller of the air-side mass flow

• specific heat and RBCCW-side mass flow
• specific heat products C,, = larger of the air and RBCCW-side mass flow
• specific heat products x = (1 - C,)UA/C,g, UA = overall air-to-water heat transfer coef ficient
• ef fective heat' transfer area l

I

136 The iteration proceeds as follows: 4 1. A value of air-side C, is assumed (mass flow

• specific heat),

2. E is calculated, along with the resulting heat transfer rate, airside AT, and air-side exit temperature, 3. The air-side condensation rate is then calculated f rom the results of j step 2; combined with the known saturation pressure of steam as a function of temperature, - 4. The condensation rate is then used to calculate the total heat trans-for rate (latent + sensible), and a corrected value for C,: Q C' = AT a If C'~C,, the iterstion i s terminated. If not, then steps 1-4 are 5. repeated until convergence is achieved, i Major assumptions of the drywell cooler model f aciude : 1. constant volametric flow naintained by the drywell cooler blorers, 2. constant EBCCR system inlet temperature and flow, 3. constant everall air-to-3BCCW heat transf er coef fielent

• ef fective heet tra:1sf er area.

j B.4 Torus Room lemocratures In an unmitigated loss of DER accident the surf ace temperature of the uninsulated torus can exceed 149'C (300*F). The greatest heat loss (at this temperature) from the surface of the torus is by radiant heat trans-for directly to the ~0.9 m (3 f t) thick concrete walls (A'*l

• s ev**a*luated i

and ceiling) of the torus room. This heat transfer rate a using an assumed emissivity of 0.9 for both torus and concrete surf aces: [ (A /A,)(1/e,- 1)] Q,= SA (T - T,)/[1/e + g g t t where j Q, = total radiant heat transfer rate S = Stef an-Boltzman constant A = surface area of torus g A, = surf ace area of concrete Tt " surface emFerature of the torus T, = surf ace temperature of the concrete e = surf ace emissivity of the torus e, = surf ace emissivity of the concrete l

137 The rate of natural circulation convective heat transfer between the torus and the torus room air is as follows: Q = A h (T -T) c t t a h = 5.3(10)-s (T - T )*.88 t a where the units on h are Btu /(ft8s*F) A similar expression is employed for calculation of heat transfer be tween torus room air and torus room concrete surfaces. A differential energy equation is solved for each of the following temperatures: torus surface, torus room air, and torus room concrete. Th6 ~0.9 m (3 f t) concrete walls are divided into five parallel, slab-geometry, ragions to insure accurate computation of the temperature dis-tribution within the concrete. The surf a:n slab is 2.54 cm (1 in.) thick, the adj acert simb 5.08 cm (2 in.) thick, and so on, with stab thickness increasing with penetration into the concrete. A typical concrets dif-fereatial energy balance is: 1 l dT /dt = [2 K,/(DX C,)][(T _1 - T ) / (DX _3 + DX ) g g p g g g g - (T -T )/(DXg + DXg )] g g where T = temperature of the i-th concrete slab g K = Con'cret* thermal Conductivity DX = thickness of i-th concrete slab g C = specific heat of concrete l pc The expression for natural circulation of air from the reactor build-ing basement (i.e. the corner rooms), into the torus room, and out the top ~ of the torus room is based on a discussion in Sect. 5.2.6.3 of the Browns Fe rry FS AR. In adapting the circulation rate given by the FSAR, it was assumed that the rate of natural circulation is proportional to the square i root of the density difference (i.e. density of reactor building basement air outside the torus room vs. the density of the air inside the torus room) and, therefore, also proportional to the square root of the tempera-i ture difference. The basement air cutside the torus room is assaned to ( remain close to 32*C (90*F) throughout the loss of DER sequence. B.5 Containment Leakane and Containment Vent Flow I Two types of primary containment leakage are considered in this re-port: normal leakage (leakage through penetration seals, etc. ) and inten-tional venting (leakage through the drywell and wetwell vent lines). All I I l _ _ _ _ _ _ _.~..__. _,._.,_ - -.______,-_- __.____,_.___

133 the BWR-LACP calculations discussed in this report assumed that the pri-mary containment leaks at a rate consistent with that measured at the Browns Ferry plant. Only the calculation reported in Sects. 4.3.1 and 10.2 assumed that the 5.08 cm (2 in.) vent lines were open. As a practical matter, normal primary containment leakage is not cap-able' of retarding containment pressurization during the loss of DER acci-dent sequence. The Browns Ferry Technical Specifications would allow a + leakage of up to 2% of total primary containment volume p9r day at pres-sures up to the design pressure. Testing conducted to date at Browns j Ferry.has shown the actual leak rate to be less than a tanth of the allow-able leakage. In BWR-LACP, normal leakage is modeled r.s a constant 0.2% per day. Since the density increases with pressurization in the loss of i DER se quence, the leakage mass flow increases in proportion to the in-creasing pressure. Discharge through the 5.08 cm (2 in.) drywell and wetwell vsnts can have a significant effect en the timing of containment failure during an ] 4 otherwise umaitigated Loss of DER tequence (sea Sect. 4.3.1). The dryws11 and we rwell vent lines both discharge in to a common 5.08 cm (2 in.) vent l line which in turn discharges to the 45 7 cm (18 in.) Standby Gas Treat-l i ment system reactor turiding ventilation ducts; therefore, the total flow rate is limited when sonic flow occurs in the coamon section of dis-charge pipe. The ficw reristance afforded by ths discharge piping, valves, and fittings must be considered, even when there is sonic flow. Pe Le 4-13 of Ref. 6 presents two different calculatlocs of the flos rate j of dry sa :ntated steam f rom 1.17 VPa (1"O yslat through 9.15 m (30 f t) of i 2-in, Schedulo 4( rips to r.tnospi tric pres sure: both ptadict a dischar:o rete cf 1.45 kg/s (3.2 lb/s). The example in Ref. 6 incicdes t1e flow recistance of not only 9.15 m (;U ft) of pfping but also a fully open i globe valve and a standard 90 elbow, as well as entrance and exit losses. For this analysis it was assumed that these losses are representative of the losses in the piping a'ctually installed at Browns Ferry Unit 1. The BWR-LACP calculation of the 5.08 cm (2 in.) vent line flow was based on the reference condition of 1.45 kg/s (3.2 lb/s) of dry steam at 1.1 MPa (170 psia). The mass flow at other conditions was calculated by multiplying the reference flow by ratios to correct for the changed condi-tions: W = (P/P,)('dl71I) (VT,/T)(W ) i

where, W = flow at any pressure or temperature or nitrogen / steam composi-tion j

1 W, = flow at reference conditions P = primary containment pressure P = reference primary containment pressure M = mole-fraction-averaged molecular weight l T, = reference primary containment temperature T = primary containment temperature 1 i

i 139 This relationship is based on the variation of sonic velocity of ident gases with respect to upstream pressure and temperature and with respect to molecular weight. B.6 Evasoration from Surface of Sunoression Pool i s Evaporation of water vapor from the surf ace of the suppression pool to the wetwell atmosphere is the dominating mechanism for. pressurization of the primary containment during the Loss of DER sequences analyzed in this report. As the pool is slowly heated, evaporation causes the con-tainment pressure to rise f ast enough such that the suppression pool re-mains subcooled and there is no direct bubble-through of SKV discharge f ram the T quenchers to the wetwell atmosphere. The wetwell atmosphere is initially a mixture of nitrogen and a man 11 amount of we te.7 vapor, and remains a binary mixture throughout most of the Loss of DER sequences. If the wetwell atmosphere we re puze water vapor, then it would be corr 2ct to assume that the partial pressure of water l vapor in the wetws11 atmosphere is identical to the naturation pressure l evaluated at the temperature of the water at the surf ace of the pool. Since the wetwell atmosphere is a binary mixture, this assumption is not correct becante the rate of transfer of water vapor from the pool is limited by dif fusion and convection through the air. The relationship used to calculate the rate of evaporation from the pool is based on the heat transfer / mass transfer analog discussed in Chap. f l 13 of kef. 5: i W = (12.3 A R )(P -Pva)h/P pa vs n i

where, l

W = cvaporation rate (Ib/s) 12.3 = constant of proportionality A = surf ace area of pool (f t*) p R = Mole-fraction-weighted (i. e. between nitrogen and water vapor) perfect gas constant P, = vapor pressure (psia) of water, evaluated at the pool surf ace y temperature P, = partial pressure of water vapor in wetwell atmosphere P, = average nitrogen pressure in the convective boundary layer (psia) h = coefficient of natural circulation heat transfer between the surface of the pool and the pool atmosphere [ Btu /(s 'F ft8)] = 5.85 x 10-8 (T - T )e.ss l P a T = pool surface temperature (*F) P T, = wetwell atmosphere temperature (*F) I

1 40 B.7 Flow Rate of Vessel Water Inloction by the CRD Hydraulic System The CRD hydraulic system is described in detail in Appendix E of Ref. 2. The computations presented in this report make the assanption that the vessel injection flow rate before a reactor trip

• is 3.8 f/s (60 sys) and that af ter a reactor trip the vessel injection increases to 10.7 f/s (170 3pe).

The CRD vessel inj ection before a reactor trip is maintained at a con-stant flow rate by an autanatic flow control valvo. Part of the total normal inj ection flow goes to each of the 185 CRDs. By maintaining a cold flow into the reactor vessel the normal CRD flow prevents hot reactor coolant from casing into contact with the CRD seals. After a reactor trip, the 185 scram. inlet valves (connected to the j charging header which is upstrean of the flow control valvo, but down-stream from the flow measutenent orifice) cren and divert flow into the charging header. This causes measured flow to increase. The flow control valve closes in an attempt to hold measured flow constant, thereby divert-ing all the flow into the charging header. The maximum flow through the charging Londor is limited by a fixed flow-restricting orifice. This is necessary because the discharge head of the 00D hydraulic pumps is 9.07 i NPa (1300 psig) or more, and the pressure downstream of the scram inlet velves can vary fran normal reactor pressure all the way down to atmos-phoric pressure. The marimum flow permitted by the flow restricting crifice is 10.7 f/s (170 spe) and this flow cccurs when reactor vessel pressuie is icw. Since the reactor vessel is depressurized to (1 MPa (130 psig) throughout most of the loss of DER sequence, e constant flow rate of 170 spa was assumed for the post /scran vessel injection flow rate.

• i.e.,

during normal power operation, or following a reactor trip when the scram has been reset. O i 1 w.- -em, - ~.,,- ~.-,,--------- ---.c,--<-n--_--.,-e .-,-.,---.-.,--,-.-n,

141 References for Annendix B B.1. S. A. Hodge et al., " Station Blackout at Browns Ferry Unit One - Accident Sequence Analysis," NURBG/CR-2182, ORNL/NUREG/TM-455/V1, November 1981. B.2. S. A. Hodge et al., "SBLOCA Outside Containment at Browns Ferry Unit One - Accident Sequence Analysis," NURBG/CR-2672, ORNL/TM-8119/V1, November 1982. B.3. A. J. Chapman, " Heat Transfer," Second Edition, New York, Tho'Mac-millan Company,1967. B.4. S. Clark, " Instruction Manual for TVA Browns Ferry Residual Heat Re-moval Exchangers BF-1-10-2, BF-2-10-2, and BF-3-10-2," Ger. oral Elec-tric Co., San Jose, CA, August, 1970. B.5. Frank Kreith, " Principles of Heat Tr ansf er. " Second Edition, Scrantor., PA, Internatienal Textbook Company,1965. B.6. Crane Co. Engineering Division, " Flow of Fluids Throagh Valves, Fit-tings, and Pipe," Technical Paper No. 410 (Fourteenth printing-2 1974), Ths Crans Comp 6ny, Chicago, IL. f I e ? ^j P 9 ,.... ~

I 5 i ^ x 3143 hppendizC s MODELING FOR LOCALIZED HEATING OF' j PRESSURE SUPPRESSION POOLS 4 C.1 Introduction s Nost of the pressure. suppression pool (PSP) modeling that is done treats the pool as a single, well-mixed node. This is an accurate model if the energy is added to the pool, at many locations or if the mass fluxes are large enough to ensure thorough mixing. In a situation where these re-quirements are not neit(particularly during severe accidents) the single node model is inadequate. -For those situations, a model that produces more detailed information'about local temperatures is needed. The overall goal of the PSP modelin's is to write a camputer program that will fulfill this need, i.e., to produce local PSP temperatures as a 4 function of time. Ideally, thi coda would be able to handle 3-D toroidal transport with a free surf ace and a non-extraneous condensation source. The only practical way to model such a problem is to make such approxima-tions as are necessary to keep computer costs 1me, while preserving enough physics to make the solution reasonably accurate. The PSP model must be capabic of following the pool local tempera-tures f rom an initial, well-mixed condition through time to a point when = the drywell fails due to ov erpressure. ' This involves modeling SRV steam flow from those at full reactor pressure (~ 200 lbe/s) to those at the low end of the decay heat curve (~ 25 lbs/1); It also involves modeling a SRV that is either stuck open or held open,as well as a SRV that operates in-termittently at high mass fluxes.' With the above ranges of size and scope applicable to PSP modeling outlined, the remainder of this appendix describes the phenomena of in-terest, a computer model of the phenomena, and some details about plume transport analysis. This Appendix is written as a brief overview of the PSP modeling effort. A detailed report of the model and result s is being prepared for publication. x C.2 Phenomenoioav I There are many phenomena that exist in PSP dynamics, but the pre-sent work deals only with those that apply to SRV discharie through a T quencher. A T quencher is shown in Fig. C.1. The T quencher was designed to discharge toward the walls of the torus instead of circumferentially around the torus. This produces turbu-l lent mixing in Bay D (the discharge bay) but very little turbulent mixing with the adj acent bays. As a result, local temperatures

• in the bay of discharge can become high.

,e

• Local temperature here is defined as in Ref. 2:

the average of l temperatures at points directly above and below the T quencher. In the computer model, local temperature is defined as the tenperature associated with a given node.

144 The magnitude of the local temperature increase will be determined by -the Bay D recirculation, the whole pool recirculation, and the pool ther-mal stratification. The concept of a Bay D recirculation flow is shown schematically in Fig. C. 2. For high steam flow rates, the momentum of the steam jets is large enough that a turbulent hot water j et impinges on the torus wall and turns upward, accelerated by its buoyancy. When it reaches the surface, it turns downward and moves back to the T quencher. This case is shown on the right half of Fig. C.2. For low steam flows, the buoyancy is dominant and the hot water created by the condensation forms a plume that turns upward before reach-ing the torus wall, rises to the surf ace, and spreads out. The low steam flow case is shown on the lef t half of Fig. C. 2. In the high steam flow case, Bay D is well mixed. In the low steam flow case, Bay D is not well mixed. The entire range of T quancher flows f alls between these two extremes. For each steam flow through the quench-er, the Bay D mixing is determined by whether the dynamics are dominated by the plume behavior or the j et behavior. The concept of whole pool recirculation can be understood on the basis of continuity. As hot water moves upward in Bay D, it forces part of the water that is there to move circumferentially out across the sur-face of the pool. Continuity implies that there must be a cold water in-flow to Bay D from the lower layers to make up for the outflow. A whole pool recirculation flow is thus started that consists of hot water moving up to the surface in Bay D, around and across the pool, together with 4 colder water that is moving down and back toward the T quencher. It is this whole pool c Experi-mental evidence *jrculation that keeps the discharge bay cool. indicates that the magnitude of this recirculation flow is quite large. Correct modeling of this phenomenon is essential to deter-mination of local temperatures in Bay D. The thermal stratification phenomenon is very simple to understand. Hot water is buoyant and tends to distribute evenly in temperature across the surface of the pool. The cold water tends to distribute in layers underneath the hot water. However, the thermal stratification phenomenon is very difficult to model rigorously because of the' lack of basic physi-cal understanding of the turbulent mixing processes. An attempt is made to model the thermal s* ratification in the model based on a very simple kinematic treatment. j C.3 Pressure Sunoression Pool Model The pressure suppression pool model is an N-layer, lumped parameter model for hot water transport in the torus. The user can input N, the number of layers. The computer program is designed to follow the PSP in time from a well-mixed initial condition up to near the local saturation temperature. Stable condensation is always assumed to occur if the local fluid is subcooled. The objective of the code is to model the thermal mixing that occurs between the discharge bay and the rest of the torus, e and, in the process, to produce local temperatures as a function of time.

l 145 The code currently models one SEV discharging through a T quencher to the pool. The SEV steam flow and enthalpy as a function of time are as-samed to be known input data. SRV discharge line dynamics are not mod-eled. The pool is coupled to the wetwell airspace through the airspace total pressure and through the pool-to-airspace evaporation rate (both of l which are also assumed to be known input data). As shown in Fig. C.3, the pool is divided into N stratified layers vertically (N=4 is shown), and 18 nodes in the e direction (around the torus). There is one 6 - node for each bay of the PSP, except for the discharge bay and the bay located 180* from the T quencher. The se two bays have two 0 - nodes each. l The code uses a quasistatic approach in modeling the PSP. At the be-ginning of each time step the local pool temperatures are used to calcu-late the condensation at the T quencher. The condensed steam and the cold water feed flow that was necessary to produce the condensation are trans-ported upward from the T quenchor within the pressure suppression pool using a steady state plume transport analysis. The steady state plumes (four are calculated for each T quencher) are assumed to be formed in-stantly, and to exist over the entire timestep. The flow conditions at the end of the plume transport (the plume entrainment and outflow) are then used as effective sources for calculating the new overall pool ten-peratures. Once new temperatures throughout. the pool are known, a new local temperature can be used to calculate the condensation source at the next timestep. Thus, feedback from the entire pool trancport enters into the local condensation calculation. Figure C.4 is a schematic showing the steps in the pool calculation. In order to perform a lumped parameter fluid flow calculation, as-sumptions about the flow field must be made. In the current model, there are two different flow fields. The first flow field models PSP flow from node to node when the T quencher is discharging. The second flow field models PSP dynamics af ter the T quencher stops discharging. The first flow field is shown in Fig. C. S. Basically, the pool is treated as a very large convection cell. The energy source in Bay D con-sists of the flow and. temperature output of the T quencher / plume transport calculation. Inflow to Bay D is the cold water feed flow, Nc, which is evenly divided between the node containing the T quencher and the nodes below. Outflow from the discharge bay. consists of the cold water feed and the quencher steam flow, M Away from the discharge bay, the pool is g. assumed to move down unifor,mly (based on continuity) to accomodate the input flow to the top layer. In the discharge bay, an internal circula-i tion is modeled by putting the entrainment flow rates from the plume ( transport module back into the Bay D nodes above the quencher. The second flow field is shown in Fig. C. 6. This flow field is de-i signed to produce thermal stratification following SEV closure. An equal ac3 opposite mass flow between each cell and its neighbor is produced that is proportional to the square root of the density difference between the cells. This approach to numerically producing the thermal stratification is based on a simple buoyancy calculation and the Taylor instability me-a chanism. A general energy balancs is written for each cell that permits flow into and out of each face. Each node has an arbitrary source term that can be used to model phenomena such as evaporation from the surf ace and

1 1 46 entrainment from the plume. In addition, each node has a storage term to allow the mass of the node to change over the timestep. This feature is used to model the moving pool level in the surf ace nodes. At each time step, the system is represented in standard state vari-able formulation: + dT - = AT + S, dt where T = vector of unknown temperatures, A = the system matrix, S = source vector. This equation is then solved to determine the temperatures at the new timestep. s C.4 Plume Transoort Analysis The plume transport analysis assumes that four steady state plumes on each side of the T quencher arms are set up in Bay D at the beginning of the time step. The four different plumes correspond to the four zones of holes on the T quencher (there are 16 plumes in all). Each plume trans-ports an appropriate fraction of M (the cold water feed to the quencher) and I, (the quencher steam flow) ertically through the stratified layers above the T quencher. Entrainment occurs from these surrounding layers and is proportional to the average velocity of the plume. Figure C.7 shows a schematic view of the plume above a discharging T quencher, along with some of the more important variables. The unknowns in the problem are r(z), w(z), and T(z): the plume ra-dius, velocity, and temperature, respectively, as a function of z. Known from the previous pool calculation is the average temperature [T,, (z)] of the nodes surrounding the plume. The 1-dimensional steady state equations for conversion of mass, momentum, and energy for the plume are: d -- ( pwr 8 ) = 2 ap,rw, (1) d s (2)

{- (pw r8) = r83 (p, p) d (3)

--{ (pwr:T) = p,awrT,, w,, r,, ,-e ---.-.,----..-,__-,-------.--__-...e

147 where p = p[T(z)], the plume density p, = p,[T,(z)], the layer density a = the entrainment coefficient T, = T,,g ( z ), the stratified layer temperatures. The above equations are re-cast in the form: x 'pwr8 2 pw8rs (4) x= r s x,, . pwr 8T. dx

= f(x)

(5) Equation (5) is solved explicitly by marching in maall space steps from the initial condition near the quencher to the top of the plume. At each space step, continuity and conservation of energy are enforced by iterat-ing on the entrained mass flow rate and plume temperature. Figure C.8 shows the plume solution method. The initial condition for the plume is found by calculating the amount of cold water (at the quencher node temperature) necessary to bring l M to equilibrium at the saturation temperature. The initial plume ten-st perature and density are assumed to be those of saturated water at the local pressure. A guess :Ls made for r,, the initial plume radius, and the initial plume velocity is calculated based on continuity: 5+5 c st W (z) = (6) 3 p, The preliminary results are found to be relatively insensitive to r,. The plume transport analysis is designed to calculate the Bay D mix-ing that occurs when a T quencher is discharging. The analysis focuses on the plume dynamics in Bay D instead of the j et dynamics. It is expected that the results will be more applicable to medium and low SEV steam flows

• than to high SEV steam flows.

However, some success at high steam flows has been found by adjusting the entrainment coefficient. Adding a calculation of the mixing within the bay of discharge due to the jet dy-namics would improve the accuracy at high steam flow rates and is planned for future work.

• "high" SEV flow occurs when the reactor vessel is fully pressurized l

[ pressure ~7.59 MPa (1100 psia)] and is about 100 kg/s (220 lb/s). The " medium" and " low" discharge rates are experienced af ter depressurization of the reactor vessel has been initiated.

148 References for Annendix C C.1. B. J. Patterson, " Mark I Containment Program Monticello T-Quencher Thermal Mixing Test Final Report," NEDO-24542 (August 1979). C. 2. T. M. Sn, " Suppression Pool Temperature Limits for BWR Contain-mont s," NUREG-0783 (November 1981). 9 1 Y b .,.-n- --v

bW D / g TE 6 e 7 96 e 2 8 G WD -L N R O + 9 2) M3-R1 s M A E 0 H H 8 T EU H D O A RR HG CA BR SE NE CI H O H ) F N hcS C nM NN 1 EE UE e A RE C E S IN Q( 21 R EU TO L T AF E G PO f R ES e L E R A OD A l HS E m H I C S o N D O o I Z r V f R ( S B E N r E P O e I Z h P C 0 c 8 E n E 0 N L 8 O e U H Z D u E q D CR N E S T-E O H 2C C 1 U z S xD c 0E l h n. 1 R ac 0 1 ipy M2) T 1 R1 A E RR 1 EU HG / C CIF N EE UES g Q( i o F L q A C L I A R T T N S E O 22 Mz MiR YS O H S e E T L U OO Hs A -r g i

150 ORNL-DWG 82-6977 ETO LOW STEAM _ _ HIGH STEAM FLOW RATE ' ' FLOW R ATE L ?ft' ~ Y( f { \\ / I / / y W T-OUENCHER ET TORUS WALL Fig. C.2. Bay D recirenlation. i e S i l l

151 ORNL-DWG 82-6918 ETD e e i e e T-OUENCHER 6-180 0=0 0 = 180' / / Fig. C.3. Pool model geometry. l l 8 l~ l

152 ORNL-DWG 82-6979 ETD INPUT POOL

GEOMETRY, STEAM FLOW INITIALIZE, SET-UP NODES I

TRANSPORT PLUMES I CALCULATE EVAPORATION I SET-UP CELL VOLUMES I SET-UP AND SOLVE THE SYSTEM CALCULATE NEW POOL LEVEL, CELL VOLUMES CALCULATE OUTPUT INFORMATION - TIME STEP Fig. C.4. Calculational steps in the code. 6 n.. , -., -,,. -.., - - - -... ~.. - -. - -., - - - -, - - - - - -

%e e 0 D 8 T 1 E 0 e 8 0 96 28 GWD -L ll N I} R O Y A d B l E e i G R f A H w C o i} S l I f D M e O g R r F a Y h A c W I} s i A d WO r L e F h N c W n O e D u M q T-R O FI N U S. C g i F lI S E M U nf = t' i' L P 0 = R o E H C e N E UQ T . e I

wu* i f 0 08 1 D T E 0 1 89 6 28 G WD -L N R O .g n ig r a hcs A i S N d S O A I t T M o A S n E L C T LF s I ISEI i OCT A P N P R V E O T R E S DW S S NT E n AE C LB e U h AWD UOO w QLR d EFP l e i f r wo = lF 6 C g i F s 0 R 0 E H C N E UQ T i I l

155 i 4 ORNL-DWG 82-6982 ETD PLUME ENTRAINMENT 2 IS FED BACK INTO it BAY D POOL SURFACE w_-: _ ~ _ _ - _ = PLUME OUTFLOW TO SURROUNDING T out (3) NODE,THEN TO REST OF POOL ENTRAINMENT 3 ii T

• III out 12t ENTRAINMENT 2 i

T (z) p iz) T,,,,,, o ENTRAINMENT 1 w( z ) '. r l tikkx OUENCHER STEAM FLOW, ( COLD WATER FLOW IS FED INTO BOTTOM OF PLUME I Fig. C.7. Plume transport problem. i I t l l

156 ORNL-DWG 82-6983 ETD INPUT Wo. To, ro l LOOPOVER ALL BAY D LEVELS ABOVE THE T-QUENCHER LOOP OVER ALL SPACE STEPS WITHIN THE LEVEL (r) CALCULATE X' l w T) CALCULATE X' + 02 il ,z + d2 CALCULATE w 'gi il CALCULATE ENTRAINMENT e ENFORCE: CONTINUITY, CONSERVATION OF ENERGY c NEXT LEVEL OUTPUT ENTRAINMENT, PLUME OUTFLOW Fig. C.8. Plame transport calculational steps. t 7 e-w,,,-,,....,--n,. _,_r r-,-,.. - -, w

• 57 Appendix D.

MARCH CODE INPUT LDHR ACC. T=34, CONT. INTACT, WITH CR0 FLOW ECHANGE TRST = 2600., PRST = 1C00.0, CPSTP = 1000., IS = 10, IFPM = 10, IFPV = 10, PEL = -1, FORP = - 1. 0, TMX = -1.0, TFX = -1.0, IHOTX = 10, HIMX = -1.0, HIOX = -1.0, IGASX 10, = W ALL X = - 1. 0, PFAIL = -1.0, ACBRK = -1.0, 10 = 0, I F IS H = - 1, NCT7 = 1, NCRST = 1, L ST7 = 0, IPLOT = 0, EE ND e EFPANAL TIMEON=9000.0, EE NO CNLMAR ITRAN=1, IBRK=0, ICBRK=1, ISPRA=1, IECC=2, l ICE =0, NPAIR=0, NINTER=18, IXPL=0, IBURN=0, TBURN=0., H2HI=0., H2LO=0., IPOTL=7, IPOEF=0, IPLOT=3, IU=0, ICKV=0, IFPSM=2, IFPSv=2, VOLC=278000.0, DTINIT=0.02,

158 TIME =2040.0. TAP =1.18E6, CEND CNLINTL CEND STEEL CONCRETE D WL I NE R DWFLCOR UPRXPE0 LORXPED WWLINER CNLSLAB N P a T.,2, NSLAB=5, DEN (1)=486.924,157.481. H C ! 1 ) = 0.113 7,0. 3107. TC(1)=25.001,0.881, NOD ( 1) = 1,6,17,2 5,35, I V L ( 1 ) = 1,1,1,1, 2, IVR(1) 1,1,1,1,2, = NN01(1)=5,11,9,9,5, NN02(1)=0,0,0,0,0, MATi(1)=1,2,2,2,1, MAT 2(1)=1,2,2,2,1, SAREA(1)=18684.0,1640.3,4130.0,1815.0,17050.0, HIF(1)=5*0.0, DT0X(1)=0.0,0.0,1.0,1.0,0.0, X(1)=C.00,0.01,0.03,0.05,0.09375, X(6)=0.,0.08,0.25,0.50,0.75,1.0,1.50,2.0,3.0,4.0,4.7326, X(17)=0.00,0.10,0.20,0.50,1.146,1.792,2.092,2.192,2.292, X(26)=0.0,.25,.75,1.25,1.75,2.25,2.75,3.25,3.49, X ( 35 )=0.0 0,0,0100,0. 020,0. 040,0. 062 5, TEMP (1)=34*362., TEMP (35)=58330., CEND CNLECC PUHIO=0., UHID=0., PACM0=0., ACM0=0., TMHH= 1. E 8, PHH=1.E6, PHLO=100.0, hHH1=5000., TMSIS=1.E8, PSIS=0.0, PSLO=0., WSISt=0., TPLH=1.E8, PLH=1.E6, PLLO=50.0, WLH1=600., NP = 3, 1.E8, 1.E8, 0., TMil) = STP(1) 1.E8, 1.E8, 2116., = 295., 289., 1300., P(1) = WEC(1) = -41429., -16346., 101.8, 0., 0., 0., PLO(1) = STPHH=1.E8,

159 STPSIS=1.E8, S T P LH = 1. E 8, RWSTM=3107859., ECCRC=.01, CSPRC=0., DTSU8=-100.0, WTCAV=1.0, 8 TLHI=C.0, TACM=0., TRWST=90.e STPECC=1.E8, CEND CNLECX LEND ENLCSX LEND ENLC00L JCCUL=0, CQR= 6. 4 8 7 5E 6, CWPR=210000., C T PR=15 0., ChSR=8587.7, CT S R= 10 5., TCCOL=1.E8, NCUUL=0, QRC00L=0., PC00L=0.0, P0FF=0.0, EEND CNLMACE e NCUB=2, NRPVl=2, NRPV2=1, NRPV3=0, ICECU8=-1, PO=125., FALL =0., HMAX=280.0, DT0=0.05, DTS=5000., DTPNT=2., IDRY=1, I W ET = 2, IBETA=0, WPCOL=1.0175200E7, TP00L=340., DCF=1000.0, VDRY=533.7, VTORUS=257700., WVMAX = 533106.1, PRE SS ( 1,2 ) =1. 32, PRESS (2,1)=0.5, WICE=0., TICE=0., ThTR=0., TWTR2=0.,

160 TSTM=0., DCFICE=0., NSMP=-2, NSMP2=2, WVMAKS=0., NCAV=1, VCAV=133.7, VFLR=400., FSPRA=1., IVENT=0, TVNT1=0., TVNT2=0., AVSRK=0., CVBRK=0., VC(1)=159000., VC(21=119C00., AR E A ( 1 ) = 1640. 3, AREA (2)=10980.0, H UM ( 1) = 0. 2 65,0. 95. T EMP0 ( 1 ) = 4 31.,340., INERT (1)=1,1, N=le NSill= 2, NC(1)= 1, NT(1)= -7, C1(1)= 131.7, C2(ll= .593, C3(1)= 10.0, C4(1)= 0., K T ( 1,2 ) = 1,K T ( 2,1) = l e STPSPR=1.EB, CE NO ENLBOIL NDTM = 100000, R1 = le R2= 10, NNT = 45686, NR = 44749, NDZ = 10, ISTR = 3, ISG = 0, ME LMOD = -1, IMWA = 1. ISTM = 0. IHC = 0, IHR = 1, N0ZDRP = 2, INZ = 100, FR = 0.0, FM = 0.0, MWCRNL = le I FP = 2, ISAT = 1, IGRIDI = 1, IGRID2 = 0, KRPS = 0, TRPS = 0.0,

l 161 AhSK = 0.0, TDK = 0.0, Yi = 0.0, YB = 0.0, DTK = 1000.0, ICON = 0 TMSG1 = 1.0E06, a TMSG2 = 1.0E06, TPM = 1.0, A8(1)=0., T8(1)=0., TMY8K=0.0, Y8RK2=1000., TMLEG(ll = 1.0E06, hDED = 0.0, TPUMP1 = 1.0E06, TPUMP2 = 1.0E06, OPUMP1 = 0 0, QPUMP2 = 0.0, TMVP1 = 1.0E06, TMUP2 = 1.0E06, hMUP1 = 0.0, WMUP2 = 0.0, HCBOT=18.03, VSHEAD=981.4, HSHEAD=3.72, ASTAND = 42.5, HSTAND = 6.81, e ASEP = 173.8, HSEP = 7.73, e TSCT(1) = 1.E6, TS8(1) = 0.20. TALF1 = 1.0E10 TALF2 = 1.0E10, Q2ERO = 1.1242E10, H= 12.289, H0 = 28.8, DC = 15.59, ACOR 108.74, = ATOT = 261.43, WATBH = 150516.9, D = 0.0424, t OF = 0.0 358, OH = 0.0459, CLA0 = 0.00472, XOO = 0.0, RHOCU = 81.48, HW = 150.Q, TG00 = 553.5, CSRV = 0.0, TMELT = 43S2.0, TFUS = 5381.0, l TFAIL = 2500., l FOROP=.75, i FCOL =.75, l l

162 FOCR = -0.50, OPART = 0.020833, DUO 2 = 0.0358, FZMCR = 0.05, F20CR = 0.08, FZdSt = 0.10, F12 = 0.445. WFE2 = 2d65.8, TFE00 = 553.50. WTRSG = 0.0, FULSG = 0.0, PS G = 0.0, PVSL = 1072., TCAV = 653.5, YLEG = 16.0, A8RK=0.0, Y8RK=1000., DT PNT 8 = 5.0, DTPN = -3.0 TPN = 1.0E06, VOLP = 21338.4, VOLS = 9684.7, TMAFW = 1.0E06, WAFW = 0.0, TAFW = 0.0, WCST = 0. 0, Fil)=0.462,1.19,1.18,1.04,1.05,1.06,1.09,1.23,1.06,0.416, P F ( 1 ) = 1.011,1. 0 87,1. 09 3,1. 09 5,1. 0 90,1. 0 94,1. 0 875,1.12 8, PF(9)=0.9665,0.408, 10*0.1, VF(1) = 3 553.50,553.50,553.50, TT ( 1) = CM(1) = 2865.8,9852.3,8712.0, AH(1) = 286.6,5085.2,31700.0, 0.5,0.5,1.0, 00(1) = 286.0,332.9,72.5, AR(1) = TT(4) = 553.50,553.50,553.50, 2331.0,5259.0,23593.0, CM(4) = 400.0,6866.0,687.0, AH(4) = 00(4) = 0.17,0.02,0.703, AR(4) = 0.0,-7.083,-12.79, NVALVE = 13, RATFLO = 838900.0, RATPRS = 1143.0, RATRHO = 2.6083, PSETI = 1105.0, TSETZ = 1.0E8, PSET2 = 0.0, W ATM AS=382589.2, CEND CNLDP TOP = 1.0E8, MIMV AL = le PSETOP = 0.0, CENO

163 CNLHEAD WZRC = 144381.7, hFEC = 26980.0, WUO2 = 351439.9, a WGRID = 66750.0, WHEAD = 207500.0, TMLT = 4135.0, DBH = 20.915, THICK = 0.7031, COND = 6.39, El = 0.0, E2 = 0.0, FDPEN = 0.0, CEND GNLHOT IHOT = 100, MWR = 1, DP = 0.25, CCN = 6.39, FLRMC = 3360.0, WTR = 0.0, TPOOLH = 100.0, NSTOP = 200, CEND ENLINTR CAYC = 0.01524, CPC = 1.30, DENS C = 2.3 75, TIC = 494.8, FC1 = 0.455, I FC2 = 0.070, FC3 = 0.388, FC4 = 0.048, RBR = 0.135, R0 = 322.6, R= 6000.0, DT = 0.5, TF= 1.0E06, TPRIN = 300.0, DPRIN = 300.0, H IM = 0.20, HIO = 0.09, F10 PEN = 0.50, NEPS = 2. TEPS(1) = 0.0,1.0E07, EPSI( L) = 0.5,0.5, IWRC = 1, IGAS = 1, 2F = 1000.0, WALL =.001, l TAUL = 0.5, TAUS = 5.0, 'D EEND

165 Appendix E: ACR@lYMS AND SYMBOLS ADS Automatic Depressurization System ANS American Nuclear Society ANSI American National Standards Institute BAF Bottom of Active Fuel BCL Battelle Columbus Laboratories BFNP Browns Ferry Nuclear Plant DPNP#1 Browns Ferry Nuclear Plant Unit One BWR Boiling Water Reactor CBP Condensate Booster Pump CILET Containment Integrated Leak Rate Test CP Condensate Pump CRD Control Rod Drive CS Core Spray System CST Condensate Storage Tank DF Decontamination Factor DER Decay Heat Removal DW Drywell r ECCS Emergency Core Cooling System EECW Emergency Equipment Cooling Water j EPA Electrical Penetration Assembly E0I Emergency Operating Instruction EPRI Electric Power Research Institute FSAR Final Safety Analysis Report GPM Gallons Per Minute HCU Hydraulic Control Unit HPCI High Pressure Coolant Inj ection ID Internal Diameter INEL Idaho National Engineering Laboratory INTER Core-concrete interaction subroutine of the MARCH code IREP Interim Reliability Evaluation Program kPA Kilopascal LACP Loss of AC Power l LDER Loss of Decay Heat Removal LPCI Low Pressure Coolant Injection Mode of the RHR System

166 LPECCS Low Pressure Emergency Core Cooling Systems LOCA Loss of Coolant Accident LOCA/DC Loss of Coolant Accident Outside Containment LOSP Loss of Offsite Power MARCH Meltdown Accident Response Char 6cteristics MPa Megapascal MSIV Main Steam Isolation Valve Mwd /te Megawatt Day per Tonne NW(e) Megawatt electrical MW(t) Megawatt thermal NPSH Net Positive Suction Head NRC Nuclear Regulatory Commission ORNL Oak Ridge National Laboratory Pa Pascal PCV Pressure Control Valve PCIS Primary Containment and Reactor Vessel Isolation Control System PCS Power Conversion System PRA Probabilistic Risk Assessment PSP Pressure Suppression Pool PV Pressure Vessel PWR Pressu-ized Water Reactor EBCCW Reactor Building Closed Cooling Water RCIC Reactor Core Isolation Cooling, System RES Office of Nuclear Regulatory Research RHR Residual Heat Removal System RHRSW Residual Heat Removal Service Water RPS Reactor Protection System RPV Reactor Pressure Vessel RWCU Reactor Water Cleanup System SASA Severe Accident Sequence Analysis 4 SBCS Standby Coolant Supply System SBGTS Standby Gas Treatment System SGT Standby Gas Treatment sy stem SBLOCA Small Break Loss of Coolant Accident

l 167 l SDV Scram Discharge Volume SI International System of Units (Systeme International) SLC Standby Liquid Control SNL Sandia National Laboratories e SRV Safety Relief Valve TAF Top of Active Fuel TIP Traveling Incore Probe TVA Tennessee Valley Authority WW We twell Zr Zirconium O r e b l l I l i _____.,__,..,_,-..,4. _, ~.., _ _.,,

169 NUREG/CR-2973 ORNL/TNH8532 Dist. Category RI, 1S Internal Distribution 1. S. J. Ball 18. A. L. Lotts 2. T. E. Cole 19. F. R. Mynatt 3. D. H. Cook 20. S. J. Niemezyk 4. W. B. Cottrell 21. L. J. Ott 5. W. G. Craddick 22. R. S. Stone 6. G. F. Flanagan 23. H. E. Treasell 7. S. R. Greene 24. R. P. Wichner 8. D. Griffith 25. A. L. Wright 9. R. M. Harrington 26. Patent Office 10-14. S. A. Hodge 27. Central Research Library 15. J. E. Jone s Jr. 28. Document Reference Section 16. T. S. Kress 29-30. Laboratory Records Department 17. R. A. Lorenz 31. Laboratory Records (RC) External Distribution n 32-33. Director, Division of Accident Evaluation, Nuclear Regulatory Commission, Washington, DC 20555 34-35. Chief, Severe Accident Assessment Branch, Nuclear Regulatory Commission, Washington, DC 20555 i 36. Office of Assistant Manager for Energy Research and Development, DOE, ORO, Oak Ridge, TN 37830 37-41. Director, Reactor Safety Ressarch Coordination Office, DOE, l Washington, DC 20555 42-43. L. D. Proctor, Tennessee Valley Authority, W10D199 C-K, 400 West Summit Hill, Knoxville, TN 37902 44. Wang Lau, Tennessee Valley Authority, W10C126 0-K, 400 West Summit Hill, Knoxville, TN 37902 45. R. F. Christie, Tennessee Valley Authority, W10C125 0-K, 400 West Summit Hill, Knoxville, TN 37902 46. J. A. Raulston, Tennessee Valley Authority, W10C126 0-K, 400 West Summit Hill, Knoxville, TN 37902 47. H. L. Jones, Tennessee Valley Authority, W10A17 C-K, 400 West Summit Hill, Knoxville, TN 37902 6 48. R. A. Bollinger, Tennessee Valley Authority,1530 Chestnut Street, Tower II, Chattanooga, TN 37401 49. Z. R. Rosztoczy, Research and Standards Coordination Branch, Of-fice of Nuclear Reactor Regulation, U.S. Nuclear Regulatory Com-mission, Washington, DC 20555 50. K. W. Holtzclaw, General Electric Company,175 Curtner Avenue, San Jose, California 95125 51-52. Technical Information Center, DOE, Oak Ridge, Tn 37830 53-587. Given distribution as shown under categories RI,1S (NTIS-10) l

NRC e om 336 U.S. NUCLEAR REIULATORY COMMIS$10N in su NUREG/CR-2973 BIBLIOGRAPHIC DATA SHEET ORNL/TM-8532

4. T6TLE AND SUBTtTLE (Add volume No., of approprontel

2. (Leave blmkl Loss of DilR Sequences at Browns Ferry Unit Che -

Accident Sequence Analysis

3. ReclPIENT'S ACCESSloN No.

7. AUTHOR (0)

D. R. Cook S. R. Greene

5. DATE REPORT COMPLETED R. M. liarrington S. A. Ilodge uOsrw l YEAR Anril 1983

9. PEHFoRMING ORGANIZATION NAME AND MAILING ADDRESS (tactuae 2,0 Codel DATE' REPORT ISSUED Oak Ridge National Laboratory l vgg "oNTH Oak Ridge, Tennessee 37830 May

6. (Leave blanki

8. (Leave blanki 12 SPONSORING oRGANIZ ATeoN N AME AND MAILING ADDRESS (include 2,0 Covel

10. PROJECT / TASK / WORK UNIT No.

Division of Accident Evaluation Office of Nuclear Regulatory Research 11 FIN No. U.S. Nuclear Regulatory Commission B0452 Washington, DC 20555

13. TYPE oF REPORT PE RIOD COVE RED (inclus/ve daMsl Topical NA

15. SUPPLEMENTARY NOTES 14 (Leave almkJ

16. ABSTR ACT (200 words or less)

This study describes the predicted response of Unit One at the Browns Ferry Nuclear Plant to a postulated loss of decay heat removal (DHR) capability following scram from full power with the power conversion system unavailable. In accident I sequences without DilR capability, the residual heat removal (RIIR) system functions of pressure suppression pool cooling and reactor vessel shutdown cooling are unavailable. Consequently, all decay heat energy is stored in the pressure suppression pool with a concomitant increase in pool temperature and primary containment pressure. With the assumption that DHR capability is not regained during the lengthy course of this accident sequence, the containment ultimately fails by overpressurization. Although unlikely, this catastrophic failure might lead to loss of the ability to inject cooling water into the reactor vessel, causing subsequent core uncovery and meltdown. The timing of these events and the effective mitigating actions that might be taken by the operator are discussed in this report.

17. KEY WoRDs AND DOCUMENT AN ALYSIS 17a DEsCRIPTORS BWR Loss of Decay lieat Removal b

l 17tt IDENTIFtE RS oPEN ENDE D TERMS 18 AV AILABILITY STATEMENT 19 SE CURITY CLASS (Th,s reporrl 21 No of PAGES Unclassified Unlimited 20 se CuRiTY C(ASS (Thes pagel 22 PRICE Unclassified s N FdC F ORM 335 #11 sti 9 .}}