ML20234C112

Severe Accident Sequence Analysis Program - Anticipated Transient Without Scram Simulations for Browns Ferry Nuclear Plant Unit 1

ML20234C112
Person / Time
Site:	Browns Ferry
Issue date:	05/31/1987
From:	Dallman R, Gottula R, Holcomb E, Jouse W, Wagoner S, Wheatley P EG&G IDAHO, INC.
To:	NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES)
References
CON-FIN-A-6354 EGG-2379, NUREG-CR-4165, NUDOCS 8707060336
Download: ML20234C112 (92)

Text

s

.A NUREG/CR-415 di EGG-2379 May 1987 Severe Accident Sequence R. Jack Dallman Richard C. Gottula Analysis Program-Anticipated Transient Edward E. Holcomb Without Scram Simulations for wayne c. Jouse Browns Ferry Nuclear Plant Unit 1 Steve R. Wagoner J Philip D. Wheatley '

I F O R M A L R E P O R T n EGnG w.uo Work performed under DOE Contract No. DE-AC07-76lD01570 ,

for the U.S. Nuclear  ?

Regulatory Commission 1 l

Aiut j

.:e f

.N = ;;; $ h h ,

T f..

'.4 .

y fI s t-I i

Idaho National j Engineering Laboratory Ma n a g e d b y t h e U. S. D e p a r t m e n t o f E n e r g y 1

8707060336 870531 PDR ADOCK 05000259 p PDR

U 4

• o, Available from Superintendent of Documents U.S. Government Printing Office Post Office Box 37082 Washington, D.C. 20013-7982 and National Technical Information Service Springfield, VA 22161 1

NOTICE This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United Sates Government nor any agency thereof, not any of their employees, makes any warranty, expressed or implied, or assumes any legal liability or responsibility for any third party's use, or the results of such use, of any information, apparatus, product or proc-ess disclosed in this report, or represents that its use by such third party would not infringe privately owned rights.

r 1

4 l

NUREG/CR-4165 EGG-2379 Distribution Category: R1 I

SEVERE ACCIDENT SEQUENCE ANALYSIS. PROGRAM-ANTICIPATED TRANSIENT WITHOUT SCRAM SIMULATIONS FOR BROWNS FERRY NUCLEAR PLANT UNIT 1 1

R. Jack Dallman Richard C. Gottula Edward E. Holcomb Wayne C. Jouse Steve R. Wagoner  ;

Philip D. Wheatley t Published May 1987 EG&G Idaho, Inc.

if Idaho Falls, Idaho 83415 1 4

Prepared for the Division of Reactor Accident Analysis Office of Nuclear Regulatory Research U.S. Nuclear Regulatory Commission Washington, D.C. ?Z55 Under DOE Contract No. DE AC07-761DO1570 FIN No. A6%4

)

I f-1 ABSTRACT An analysis of five anticipated transients without scram (ATWS) was conducted at the Idaho National Engineering Laboratory (INEL). The five detailed deterministic simula-tions of postulated ATWS sequences were initiated from a main steamline isolation valve (MSIV) closure. The subject of the analysis was the Browns Ferry Nuclear Plant Unit 1, a boiling water reactor (BWR) of the BWR/4 product line with a Mark I containment.

The simulations yielded insights to the possible consequences resulting from a MSIV closure ATWS, An evaluation of the effects of plant safety systems and operator actions on acci-dent progression and mitigation is presented. j i

i l

1 1

1 l

i FIN No. A6354 - Severe Accident Sequence Analysis ii I

EXECUTIVE

SUMMARY

Under the auspices of the U.S. Nuclear Regulatory damage was analyzed using the FRAP-To and SCDAP Commission (NRC), simulations of anticipated tran- computer codes. The fuel damage analysis included zir- l sients without scram (ATWS) were performed at the caloy oxidation and fission product release. However, l Idaho National Engineering Laboratory (INEL). The it did not consider vessel melting or the transport of I Browns Ferry Nuclear Plant Unit 1 (BFNP1) was fission products outside of the RPV, l selected as the specific subject of this work because A probabilistic approach was first undertaken to of the cooperation of the Tennessee Valley Authority isolate significant phenomena, events, and system interactions. A sequence event tree (SET) was devel- 1 (TVA). Also, a probabilistic risk assessment (PRA) was available for BFNPl. BFNP1 is a boiling water oped, which formulated the relative frequency of reactor (BWR) representative of the BWR/4 product occurrence of sequences and the expected conse-line with a Mark I containment. This work is part of quences. Analysis of the SET revealed that even though a cooperative effort coordinated by the NRC's Severe sequence event timing can show variations, the tran-Accident Sequence Analysis (SASA) Program. Work sient signatures are limited and distinct. A plant auto-performed at the Oak Ridge National Laboratory matic simulation (no operator actiorc) was chosen to tORNL) and Brookhaven National Laboratory (BNL) illustrate the accident event progression. This simula-is also part of the BWR SASA effort on ATWS. tion would also serve as a base case from which to Several objectives were identified in the present determine the effects of certain operator actions. It was ATWS study. These include establishing the method- observed from the plant automatic simulation that ology for and perfonning comprehensive deterministic mitigative operator actions are required to prevent con-analyses of the postulated accident, determining acci- tainment failure and severe fuel damage. Primary dent event progression, and evaluating the effects of containment failure was predicted from static over-plant safety systems and operator actions on accident pressurization 45 min after transient initiation.

progression. An additional objective of the study was Four other simulations were performed that modeled j

to define requirements for improved analytical methods various combinations of operar actions and assumed or for experiments to resolve uncertainties. Proposed effectiveness of the standby liquid control system Emergency Procedure Guidelines (EPGs) were used (SLCS). Each simulation modeled an operator action as a basis for modeling operator actions. The ATWS called level control, whereby the operator is instructed simulations performed for this study thus served as a to lower downcomer water level to the top of the ac-partial evaluation of the EPGs. tive fuel (TAF). Several conclusions resulted from the Probabilistic analyses have asserted that the ATWS analysis of those simulations. The steps proposed in at BFNPI (and other BWRs) is included in a group the EPGs, as they apply to the RPV and primary con-of dominant transients relative to core damage fre- tainment, would result in the mitigation of a MSIV quency. Although low in probability, ATWS accidents closure ATWS if carried out successfully. The need ,

are of concern because they could lead to core damage for operator training in the use of the EPGs was ob- l and fission product release to the environment. The served because actions must be taken promptly to J transients analyzed were initiated with a main steamline reduce the risk from the accident.

isolation valve (MSIV) closure followed by a complete The SLCS was modeled using a bounding appmach.

failure to scram. In all cases it was assumed that the Maximum SLCS effectiveness was modeled by assum-recirculation pumps tripped automatically. This ing that the boron solution was transported isotopically scenario results in the most severe duty to the centain- with liquid in the RPV. Minimum SLCS effectiveness ment because the main condenser is unavailable to con- was modeled by assuming that the boron solution was dense vessel steam. Steam produced in the vessel is completely stratified in the lower plenum, until a suf-dhcharged to the pressure suppression pool (PSP), fic ient amount had been injected to effect a hot shut-creathg a threat to the containment unless reactor down. Three simulations were performed assumir.g l power can be reduced to decay heat levels. The simula- that the operator initiated SLCS (50 gpm capacity) and tions are limited to the response of the reactor pressure began level control at 120 s. With maximum SLCS ef-vessel (RPV) and primary containment (PSP and festiveness, the predicted PSP temperature at the time drywell). RELAP5/ MODI.6 was used to model the of reactor shutdown was 140*F. With minimum SLCS RPV and associated recirculation loops and plant effectiveness, the PSP reached 195 F with RPV systems. CONTEMlYr/LT-028 was used to model the depressurization and 218"F without RPV depressuriza-primary containment. An iterative approach was used tion when the reactor was shutdown. A fourth simula-to exchange information between the two codes. Fuel tion was performed with the SLCS capacity increased iii

l 1

to 86 gpm. The 86 gpm corresponds to the published 3. Operator training should be an integral part of NRC final rule on ATWS. Assuming minimum SLCS the implementation of the EPGs. These sim-effectiveness and no RPV depressurization, the in- ulations indicate that the operator must act creased capacity of 86 gpm reduced the PSP temper- properly and promptly to reduce the risk from ature at shutdown from 218 to 173'F. The simulation a MSIV closure ATWS. The process of follow-matrix and predicted PSP temperatures are illustrated ing the EPGs through their various parts could in the following table. be very time consuming unless the operators Conclusions, results, and recommendations resulting had a thorough understanding of them before-from the study are summarized below. hand.

1. Mitigative operator actions are required to pre- 4. More guidance should be given in the EPGs vent containment failure and severe fuel damage relative to the injection oflow pressure systems during a MSIV closure ATWS. During the plant to the RPV. The operator is told to " slowly automatic simulation, primary containment fail- inject" when the RPV is depressurized. More ure was predicted from static overpressurization definitive instruction is required, as are mea-45 min after transient initiation, in less than four sures to prevent automatic injections.

hours, over half of the core was predicted to have liquefied and relocated. Sigmficant hydro-

5. Level control as advocated by Contingency #7 is an effective although limited means of redoc- i gen production and fission product release from the fuel was predicted to occur.

i or pwer. Lowering RPV water level m BF M 1000 pia was calculated to reduce

2. It is predicted from these simulations that if all reactor power fmm 30% to approximately 17%

of the actions proposed by the EPGs are of rated. This rawes the PSP heatup rate from completed successfully, then a MSIV closure 6 to 4*F/ min. Although several analyses have ATWS would be brought under control. Some converged on the predicted power level range of the actions, however, may be very difficult of 17-20%, uncertainties remain and the actual to accomplish. power could be higher.

Summary of MSIV closure ATWS simulations PSP 4 Temperature l Simulation Boron Level RPV at Shutdown j Description Modeling Control Depressurization (* F)

Plant automatic None No Yes (automatic Shutdown not ADS) achieved EPG nominal- 50 ppm Yes No (pressure 140 I maximum SLCS isotropic control, not '

effectiveness depressurization)

EPG nominal- 50 gpm Yes Yes (manual 195 minimum SLCS stratified blowdown) effectiveness Minimum SLCS 50 gpm Yes No 218 effectiveness stratified q without depressurization EPG nominal- 86 gpm Yes No 173 increased SLCS stratified capacity and minimum j effectiveness -

iv ,

6. The action of depressurizing the RPV to avoid 218 to 173*F for minimum effectiveness assumptions and no RPV depressurization. The j violation of the heat capacity temperature limit reduction in temperature is significant in that (HCTL) should be evaluated further. RPV depressurization should result in a power reduc- complete condensation in the PSP should ensure tion during an ATWS, and it may also preclude that no steam breakthrough would occur, thus l the need to blowdown the RPV when the PSP preventing drywell pressurization.

has no condensing capacity. However, the potential for low pretsure power oscillations and 9. The uncertainty of assuming uniform control rod unthrottled low pressure ECC system injections worth is currently too large to allow reliable .

could cause fuel damage. modeling. It is recommended that the effect of individual rod insertion be quantified in terms 7.- Even w.it h minimum SLCS effectiveness, it is nega r ctivitywrsm time, j predicted that PSP temperature would be less than 200'F at the time of shutdown Using 10. Use of the torus cooling mode of the residual-50 gpm of 13% sodium pentaborate solutioit in heat removal (RHR) system is not assured dur-conjunction with level and pressure control ing a level control transient. Once RPV level results in a PSP temperature of 195'F at shut- is raised back up, torus cooling should be readily down. Assuming maximum SLCS effectiveness

".## *E '. The importance of torus cooling results in a PSP temperature of 140'F. hes not only in reducing PSP heatup rate but

8. Increasing SLCS capacity to 86 gpm reduces the also helping ensure that the torus remains well predicted PSP temperature at shutdown from mixed.

4 i

i v

1 l

l ACKNOWLEDGMENTS The support of the U.S. Nuclear Regulatory Commission is gratefully acknowledged, particularly that of Dr. Bharat Agrawal. The cooperation and valuable discussions of several others also contributed significantly to this study: L. Claassen (General Electric Co.).

l B. Chexal (EPRI), S. Hodge and M. Harrington (ORNL), M. Miller and K. Keith (TVA),

and P. Saha and G. Slovik (BNL).

The authors thank Gary E. Wilson for his technical review and management support i of this work. The efforts of Louise Judy and Joan Mosher in text processing are greatly {

appreciated. Special thanks go to Nancy Thornley for generating drawings and plots j throughout this analysis.

i vi

CONTENTS ABSTRACT. ii

SUMMARY

. iii ACKNOWLEDGMENTS . . Vi Xi NOMENCLATURE . .

1

1. INTRODUCTION .

2. EARLY PLANT RESPONSE TO A MSIV CLOSUfiE ATWS. . 2 2.1 Normal Operation . . . 2 2.2 System Alignment During MSIV Closure ATWS . 2 i

2.3 Sequence of Events Dt. ring MSIV Closure ATWS 7

3. EMERGENCY PROCEDURE GUIDELINES. . 9

4. ANALYSIS APPROACH . . . 16 4.1 Operator Actions and System Interactions . 16 4.2 Sequence Event Tree 17 4.3 Application of the SET . . 17

5. RPV AND PRIMARY CONTAINMENT RESULTS . 19 5.1 Plant Automatic Simulation . 19 5.2 Simulation of Operator Actions . . 24 5.2.1 EPG Nominal-Maximum SLCS Effectiveness . 24 5.2.2 EPG Nominal-Minimum SLCS Effectiveness. 25 5.2.3 Minimum SLCS Effectiveness-Without Depressurization . 30 5.2.4 EPG Nominal-86 gpm SLCS . . 31 5.3 Effectiveness of EPGs. .. 31

6. FUEL DAMAGE ANALYSIS . . . . . . 34 6.1 Power Excursions: FRAP-T6 Results . 34 6.2 liigh Pressure Boiloff: SCDAP Results . . 35

7. ANALYSIS UNCERTAINTIES. . . 38 7.1 SLCS Effectiveness. . 38 7.2 Effectiveness of Level Control. .. 38 vii

l l

7.3 Manual Rod insertion . . . . . . ..... .. . .... .......,,,,.. . .. 40 1 7.4 Containment Related Uncertainties . . . 41 I

8. CONCLUSIONS AND RECOMMENDATIONS . .. . . . 42

9. REFERENCES. . . . . . . . . . 43 APPENDIX A-RELAP5/ MOD 1.6 MODEL OF BFNPI REACIOR PRESSURE VESSEL. . A-1 APPENDIX B-DESCRIPTION OF BFNPI PRIMARY CONTAINMENT AND 4 l

CONTEMPT /LT-028 MODEL , , B-1 APPENDIX C-DESCRIPTION OF THE SCDAP CODE AND THE BFNPl SCDAi MODEL , .. . . . . . , C-1 APPENDIX D-FRAP-T6 CODE AND MODEL DESCRIPTION D-1 FIGURES

1. BFNP normal operation schematic . . .. . 3

2. Reactor pressure vessel . . . . . . . 4

3. Flow path schematic during early stages of MSIV closure ATWS . 5

4. Primary containment schematic. . .. . . 6

5. EPG level and power control flowchart. . .. 10 ,

l

6. EPG pressure control extracted flowchart . .. . . . . I1

7. Heat capacity temperature limit . . . 12

8. Contingency #7 extracted flowchart . . . . .. . . 13

9. EPG primary containment flowchart .. . . . . .. . . . 14

10. MIR alignment . . . .. . .. . . . . 15

11. Information exchange between RELAP5 and CONTEMPT. , . ... 19

12. Plant automatic transient-RPV pressure . . . . . , , .. 21

13. Plant automatic transient - normalized core power . . . . .. . 22

14. Plant automatic transient - pressure suppression pool temperature . . . . . . . 23

15. Plant automatic transient - drywell pressure . . . . 23

16. Isotropic boron - downcomer water level .. . . . . 25

17. Isotropic boron - injection rate to the RPV . .. . . . 26 viii

i Isotropic boron - normalized core power. . . . . .. . . . .. 26

18. . ..

Isotropic boron - pressure suppression pool temperature . . .. . .. . . . .. 27 19.

. . . 28

20. Stratified boron - downcomer water level . . . . .. . . . . . . .

Stratified boron - normalized core power . .. . . .. 28

21. . .

Stratified boron - pressure suppression pool temperature. . ........ 29

22. . .

23. Stratified boron '- RPV pressure . . ..... .. . . .. . . . .... . 29

24. Normalized core power without depressurization . . .... .. . .. . . . 30 -

Pressure suppression pool temperature without depressurization . .. .. .. 31

25. .

26. Pressure suppression pool temperature with 86 gpm stratified boron . . .. . . . 32 PSP temperature comparisons . .. . 33

27. . .... . .. .

28. FRAP-T6 input power . . ,. . . .. .. . . . .. . 34'

29. SCDAP fuel rod cladding temperature histories of the top six core axial nodes. . . 36

30. SCDAP mass flow rate of steam in the bundle at nodes four and ten . . . . . 36 '

31. The effect on reactor power of lowering downcomer liquid level in steady-state operations . 39 A- 1. Vessel nodalization diagram . . . . . . .. . . , A-4 A-2 Jet pump performance curve . . .. .. . , , .. A-5 A-3. Core nodalization . . . . . . . . . . .. A-6 A-4. Core axial power profiles .. . ., .. .. . . . . A-7 A-5. Core axial void profiles at 20 F subcooling . . . . . . . . A-9 A -6. Pump trip test data comparisons - pump speed . ...... ... . . . .. . . . A-13 A-7. Pump trip test data comparison - core inlet flow . .. . ...... .... . . . ... . A-14 A-8. Pump trip test data comparison - reactor power . . . . . . ... . . A-14 A-9. Pump trip test data comparison - steamline flow . . . . . . . ... .. A-15 A-10. Pump trip test data comparisor - steam dome pressure . . .. . . .. . A-15 ,

A-ll. Pump trip test data comparison - wide range level. . .. . ... . . .. A-16 A-12. Generator load rejection data comparison - steamline flow . . . . .. .. . .. A-16 A-13. Generator load rejection data comparison - reactor power . .. . . . .. . ., A-17 A-14. Generator load rejection data comparison - steam dome pressure. .. . . A-17 ix

A-15. Generator load rejection data comparison - core inlet flow . . . . . A-18 A-16. Generator load rejection data comparison - wide range level . . A-18 B-1. Mark I drywell/ torus configuration . . . . .. B-3 B-2. Cross section of the torus . . .. . . .. . . .. . . .. . B-4 B-3. CONTEMPT Browns Ferry containment model . .... . B-6 D-1. FRAP-T6 input power . . . .. . . . . . . . .. D-4 TABLES

1. MSIV closure ATWS sequence of events-first 80 s . . ... . . 8

2. MSIV closure ATWS simulations. . . ... . 20

3. Plant automatic transient event timings after 100 s . . . . .. 24

4. Summary of predicted PSP temperatures. . . . . 33 A- 1. Plant modeling parameters . . . A-3 A-2. Core lattice modeling parameters . . . .. A-6 A-3. Reactor kinetics data . . . . . A-7 A-4. Fuel assembly data. . . . . A-8 A-5 Safety / relief valve characteristics . A-9 A-6 RPV geometric data. . . . . . A-10 A-7 RPV vessel elevations . A-12 A-8. Steady-state conditions at 100% power . .. . A-12 B.I. CONTEMI'T/LT-028 Browns Ferry heat structure modeling . . . .. B-7 B-2. ATWS calculation initial conditions compared to plant specifications .. . . B-8 C-1. SCDAP input parameters . C-4 l

D-1. Additional input for the FRAP-T6 analysis . . . . . . . D-4 x

l

NOMENCLATURE l l

ADS Automatic depressurization system NSSS Nuclear steam supply system ATWS Anticipated transient without scram ORNL Oak Ridge National Laboratory {

BFNP1 Browns Ferry Nuclear Plant Unit i PC Primary containment  !

BNL Brookha,en National Laboratory PC/P Primary containment pressure BWR Boiling water reactor PRA Probabilistic risk assessment CRD Control rod drive PSP Pressure suppression pool CST Condensate storage tank RCIC Reactor core isolation cooling DW/T Drywell temperature RC/L RPV control / level ECC Emergency core cooling RC/P RPV control / pressure EPGs Emergency Procedure Guidelines RC/Q RPV control / power EPRI Electric Power Research Institute RHR Residual heat removal FIST Full Integral Simulation Test RPV Reactor pressure vessel HCTL Heat capacity temperature limit SASA Severe Accident Sequence Analysis HDWP High drywell pressure SCDAP Severe Core Damage Analysis Package HPCI High pressure coolant injection SET Sequence event tree j i

INEL Idaho National Engineering Laboratory SLCS Standby liquid control system j LLLWL Triple lo level water level SP/L Suppression pool water level LOCA Loss-of-coolant accident SP/T Suppression pool temperature LPCI Low pressure coolant injection SRV Safety relief valve LPCS Low pressure core spray TAF Top of active fuel MSIV Main steamline isolation valve TVA Tennessee Valley Authority NRC Nuclear Regulatory Commission xi

SEVERE ACCIDENT SEQUENCE ANALYSIS PROGRAM-ANTICIPATED TRANSIENT WITHOUT SCRAM SIMULATIONS FOR BROWNS FERRY NUCLEAR PLANT UNIT 1 i i

1. INTRODUCTION The subject of this report is the analysis of an an- ysis methodology. Using probabilistic methods, a l

I ticipated transient without scram ( ATWS) in a boiling sequence event tree (SET)4 was developed which water reactor (BWR). With the cooperation of the Ten- described the various paths (or sequences) that an nessee Valley Authority (TVA), the specific subject ATWS could follow in terms of plant systems and of this effort is the Browns Ferry Nuclear Plant Unit 1 operator actions. Four dominant (higher probability of (BFNPI). The work is sponsored by the U.S. Nuclear core or containment damage) sequences were iden-Regulatory Commission (NRC) through the Severe tified. Analysis of the dominant sequences yielded in-Accident Sequence Analysis (SASA) Program. SASA sight to the possible consequences from a MSIV closure objectives ir.clude establishing the methodology for and ATWS. Even though timings of events can show con-performing comprehensive deterministic analyses of siderable variation, the transient signatures are limited severe accidents, determining accident event progres- and distinct.

sion, evaluating the effects of plant safety systems and Section 2 of this report describes the BFNPI with operator actions on accident progression, and defin- regard to systems that are important during an ing requirements for improved analytical methods or ATWS. Operator actions are also discussed as they for experiments to resolve uncertainties. relate to Emergency Procedure Guidelines (EPGs)in An ATWS occurs when an expected operational oc- Section 3. Plant systems and operator actions are tied currence is followed by a failure to scram the reactor. together by describing their interaction during the Previous probabilistic analyses' have asserted that the tranient.

ATWS at BFNPI is included in a group of dominant The analysis methodology is outlined in Section 4.

risk transients. Although low in probability, ATWS Section 5 presents key calculational results as they  !

accidents are of concern because they could lead to core relate to the reactor pressure vessel (RPV) and primary  !

damage and fission product release to the environment. containment. Fuel damage analysis of the plant auto- ,

The ATWS studied here is initiated by a main steamline matic transient (no operator actions) is given in Sec-isolation valve (MSIV) closure. This event leads to the tion 6. Uncertainties with the deterministic analyses most severe ATWS sequence 23 because the power are outlined in Section 7, and conclusions and conversion system is lost from the start of the tran- recommendations resulting from the analysis are given sient. Steam produced in the vessel is discharged to in Section 8. A list of references is provided in the pressure suppression pool (PSP), creating a threat Section 9. l to containment unless power can be reduced to decay Appendix A describes the RELAP5/ MODI.6 levels. The outcome of the transient is governed by computer model of the reactor pressure vessel of the ability of plant systems and operator actions to BFNPl. Appendix B describes the primary contain-maintain core cooling and containment integrity until ment and CONTEMPT /LT-028 model. Appendix C the accident can be controlled. describes the SCDAP code and the BFNPI SCDAP In line with SASA objectives, the MSIV closure model, and Appendix D describes the FRAP-T6 code ATWS at BFNPI was studied by developing an anal- and model.  ;

J h

1

i l

2. EARLY PLANT RESPONSE TO A MSIV CLOSURE ATWS i

In this section, the plant response to a MSIV closure 2.2 System Alignment During ATWS is described. First, the system alignment dur-ing normal operation is discussed. Next, the automatic MSlV Closure ATWS plant response for the first 100 s of the ATWS event Closure of the MSIVs isolates the turbines and con-is presented. Because operator actions are assumed to denser from the pressure vessel. The main condenser l begin at 120 s, presentation of the early event timings is no longer available to condense vessel steam, and j applies to simulations with and without operator ac- the feedwater pumps will not function because steam tions. Results from the complete simulations are for feedwater turbines is not available. Figure 3 il-presented in Sections 5 and 6. lustrates pertinent flow path during the early portion of a MSIV closure ATS S.

Vessel steam is directed through the safety relief 2.1 Normal Operation valves (SRVs) to the PSP, where it is condensed. Part A schematic of the pressure vessel, main steam of the steam is diverted from the main steamline to system, and condensate and feedwater system is shown drive the high pressure coolant injection (IIPCI) and in Figure 1. Flow paths indicate the alignment during reactor core isolation cooling (RCIC) turbines, which normal operation of the BWR direct steam cycle. Steam provide high pressure makeup water to the vessel. The generated in the core exits the pressure vessel through HPCI and RCIC pumps, as well as the CRD pump, four main steanlines. Most of the steam is used to drive take suction from the condensate storage tank (CST).

the main turbines for electricity generation. After leav- PSP cooling (N2-1/2% of rated power) can be pro-ing the turbines, the steam is condensed and becomes vided with the residual heat removal (RHR) system makeup for the feedwater system. Steam that is not following manual alignment.a Low pressure coolant used by the main turbines is diverted from the main injection (LPCI), core spray, and condensate booster steamlines and used to drive the feedwater turbines, pumps would also take suction from the Lar.

A drawing of the RPV is shown in Figure 2. Dur- During a transient involving MSIV closure, contain-ing normal operation, two recirculation pumps provide ment interactions are important. Figure 4 illustrates the the driving potential for the core inlet flow. Taking primary containment at BFNPl. The primary contain-suction from the lower downcomer, ~1/3 of the total ment encloses the reactor pressure vessel, recircula-core flow passes through the recirculation loops, tion loops, and other components including the SRVs. 4 discharging through the jet pump drive nozzles. The The primary containment consists primarily of a dry- f remaining 2/3 enters the jet pump suction from the well and pressure suppression pool. Connecting vent downcomer. This combined flow is discharged through systems, isolation valves, cooling systems, and other the jet pumps (10 per loop) into the lower plenum service equipment are also included. Appendix B in-region. Entering the core inlet, ~10% of the flow is cludes a more complete description of the BFNPI  ;

diverted through the interstitial bypass region (the Mark I containment. '

region inside the core shroud, but outside the fuel chan- The main function of the primary containment j nels). Recombining in the upper plenum region with syt; tem is to provide radiological shielding, first be- {

flow through the fuel channels, the saturated two-phase tween the nuclear boiler and the reactor building, and l mixture passes through the steam separators. Water is ultimately the environment. The drywell is a steel separated and returned to the downcomer. Steam con- pressure vessel encased in concrete. Its internal design tinues upward through the dryers and leaves the pressure is 56 psig at 281*F. Cooled by 10 large fan pressure vessel via the main steamlines. The mass of cooling units, it normally operates at an internal steam leaving through the steamlines is replenished by temperature ofless than 135*F. The drywell is nitrogen the main feedwater system. filled, the containment being inerted to prevent burn-Also shown in Figure 2 are elevations corresponding ing of hydrogen in the event of its release from the to downcomer level trips. These will be referred to in vessel during an accident. Via eight large vent pipes, subsequent discussions. Also shown are injection points the drywell is connected to the torus, which includes for the control rod drive (CRD) and standby liquid con-trol system (SLCS). When activated, the SLCS injects

a. As pomted out by S. A. Hodge (ORNL), the PSP cooling capa-sodium pentaborate thmugh a single sparger below the city is a function of pool temperature and can reach 4-l/2% of rated core support plate. power at very high pool temperatures.

2

1 i) 5 0

4 0

R 0 T

O S T N A

R E

N E

G E T

A S

N E

D N ,

O yay #E t R

E S

N C

D E

2 B

8 R

L E

O E%

E F mj' R U

T t T O .

d C M

E T

S Y

S R .

E ic M

A m wg ET t a

E g T A AM m T E t SWE e S

N T

R tg NDT EES h

c s

I A E oFEY r S n o

M T _-

o i S

O m- CCf t

a r

A e OW B _

p o

l l

a m _

,' r o

P n -

2 N _

v 6 F m 'S i D

E t

t E Al B _

w7u F E

1 F e r

u ig F

l

~

\

1 9

I L

J

)/ . .

S E

V I

R O

P h _..

O C

R

( g

\

I l

N O

A TI A

m V

8 H

C T N 9 p_

C O A I T

E AM R iL d

Y PS C

E R

w

ll 1

A 6

0 4

R O O

T O

T C S _

A N E _

) R _

N F _

fj

)

O _

r I 5 N 4 E 0

I 8 DD I A 1 0 3

( SE 6 6 NF I

5 7 L

(

4 W E OM L

(

O V )N TO R E L L E[ T E V E F L 1 DT EO Y E V R R L E U L

O L 6 3 CB N

E O L L G L - 6 tL R A

M A O F O

L (3 h. E tS A M L- N S P E R F S T O O MA I

OA uE eV S N L LT . S C

I L

S I

- 4< y.' l.

e s

s g

\ -

7 T

e v

e r -

j u s

- _ s P

J

_ OS e

r p

r MSA t o

c

- a

> _ EP e

- _ RY R

_ OB -

C ~ .

_ 2

- - e r

g u ig

/s- _.-

~ .

7 T

F I

(

i N '

^' '

S E

< S R

3 4

A.

o u

D O T

A > R '

T D

R T E

N )2 C R A J c TF I

W i P E ,! AO O

L F < S W , L U1 D M O C(

ME A L R IP E F AN T CO EI B S E EO RL T V SWT I R

D N

AO MT r

u

4 0

4 0

R 0 T

H S R N

-,'l$+i

- ~

N

,s N

s m p T g S O

p C S

, W T

A e

' a, r u

- s

' o

- ' lc

'( : 4 i ll!

V I

- _ S

_ M f

o

- s

_ e V

I g E, V

,l sil t

g a

s l

R So M m E I S ly _

r _

a e

g

, w T n y ,

i r

r / d u

S ic _

/ M t a

B m L @ h e

c 0 2 e P s 0 b S ht l 8 _

1 1

v P a p

O O w T T lo l\ F W d - _

T e " 3 M

i' d

e r .

?

5 2

9 A . -

- T m m F u

ig lf' -

l w

M P M

,iIli C P G

I 0 C 0 IC 0 S P 0 C 0 H 5 R 6 C

L S - e -

1l!ll

I C M P

H A E

T S

LY EL SP SP EU n  !!

E._ VS p E N

V IC B I

P P R

_ R H i U

_ , T I e a !l3, 8 -

t

\

n p

C

_ ,\

_ , l P U

_ r, " P m/

_ /

L L

E W

% I C

VE RN SL I

R E

H h t

e a

m Y

R D

P L S T

N \

E

\ ]i C

N E

yM c s

t n

e m

_ V U in O T i

a

- S t n

C o

_ F T

M c

y r

a E

l O m l R i

_ ES F P RE ,

V 4

_ ML y e r

_ UA ,

i u

_ UV F C

A ,

V N EO

_ R IS L

_ US 1\s G

_ SE O S N P P SRO H E PP RP + gR UI L H M RP U OO PU TO '

. r S C y

M TR

_ AE

_ EG HN R A

_ HH RCX E

the PSP holding nearly 1 million gallons of water. The pumps rapidly coast down, turbine-driven feedwater function of the PSP is to provide a heat sink for energy is lost. Continued core steaming and loss of feedwater released from the vessel during a design basis loss-of- causes the vessel downcomer level to decrease. When coolant accident (LOCA). The PSP also provides a 10-10 level (see Figure 2) is reached, HPCI and RCIC source of makeup water for the low pressure emer- pumps start automatically. The HPCI and RCIC take gency core cooling systems (ECCS). suction from the CST, and pump water into the RPV Reactor vessel SRVs discharge to the PSP when re- via the feedwater line. HPCI flow accounts for quired. Steam generated in the vessel during a MSIV 5000 gpm and RCIC for 600 gpm of high pressure closure ATWS is relieved entirely through the SRVs makeup.

to the PSP. Steam is discharged to the suppression pool A scram would be signaled on four situations: MSIV via spargers called T-quenchers. The T-quenchers are closure, high RPV pressure, high neutron flux, and designed to promote complete condensation in the pool, lo downcomer level (539.0 in. above the inside of the l RPV bottom head). It is assumed for this ATWS study l 2.3 Sequence of Events During that all of the control rods fail to insert. Following the scram signal H2 gpm of wata is pumped into the MSIV Closure ATWS RPV by the CRD system.,

Vendor analysis (see Reference 2) of ATWS occur- Table I lists timing of key events during the early rence concluded that the MSIV closure without scram portion of a MSIV closure ATWS. Note that all events j l

l " transient results in higher vessel pressure, more have occurred without operator action. )

severe fuel duty, and a larger amount of steam being After recirculation pump coastdown, the RPV is m a natural circulation mode of operation. With makeup dumped into the (pressure) suppressimi rool than any other moderately frequent transient " Thus, analysis provided by the HPCI, RCIC, and CRD systems, a of MSIV closure without scram provides a bounding pseudo steady state condition exists in the RPV. Reac- l case for ATWS events in BWRs, and the scope of the tor power stabilizes at ~30%, which results in a steam- I present analysis was restricted accordingly. A descrip- ing rate to the PSP of ~21 % of normal steamline flow. l tion of te subject transient follows. The difference in power and steaming rate is accounted  !

MStV c!mure causes a rapid pressurization of the for by heating of the subcooled ECC fluid to satura- I RPV because the core continues to produce steam at tion temperature in the core.

near rated conditions. Increased pressure collapses The subsequent transient can progress through many voids in the core, which introduces positive reactivity different sequences, depending on plant system and causing increased power and steam generation. The in- operator action interactions. Operator actions as they creasing pressure causes the recirculation pumps to trip relate to an ATWS are discussed in the following automatically (at 1135 psia), and the SRVs to lift. section.

These two actions relieve overpressurization by vent-ing steam and reducing core inlet flow, which reduces power by decreasing core void content and introducing negative reactivity into the core. a. The CRD hydraulic system pumps water continuously into the R m a scram signal, Wat water Gow is M gpm. FoHowing The MSIV closure isolates the steam suPP lY to the a scram signal, the flow is increased and taken to be a constant value three feedwater turbines. As the turbines and feedwater of 112 ppm in this study.

}

7

f-Table 1. MSIV closure ATWS sequence of events-first 80 s' Time (s) Value Event 0.0 -

Transient initiation, MSIVs begin to close.

0.4 10 % MSIV fractional area has decreased 10%

causing scram signal. I 2.67 104.7% Jet pump discharge mass flow peaks.

2.71 253.0 % Peak reactor power.

3.03 110.0 psi /s Maximum rate of change of steam dome pressure, SRVs begin to open.

3.26 1135.0 psia Recirculation pump trip signal on high steam dome pressure.

3.79 -

Recirculation pumps trip, begin flow coastdown.

5.0 -

MSIV completely closed.

6.7 2582*F Peak fuel temperature.

6.7 82 % SRV flow peaks at 82% of normal steaming.

7.3 608.9'F Peak cladding temperature.

10.5 1272 psia Peak steam dome pressure.

34 476.0 in. Downcomer level reaches 10-10 level, HPCI i and RCIC actuate.

78 110 F PSP temperature reaches 110*F, SLCS in-jection called for by EPGs.

79 5600 gpm HPCI and RCIC at full flow.

a. Event timings as predicted by RELAPS/ MODI.6.

l 1

8 l

3. EMERGENCY PROCEDURE GUIDELINES As described in the previous section, a quasi steady Contingency #7 is discussed in more detail at the end state condition would be automatically reached in less of this section.

than two minutes following MSIV closure and control While executing RPV water level control, the rod scram failure. With high pressure makeup provided operator is also told to monitor and control reactor by the HPCI, RCIC, and CRD systems, the RPV power [RPV control / power (RC/Q)]. Two decision would be steaming at approximately 21% of its full points are passed immediately (see Figure 5) because power rate. The steam is condensed in the PSP, which reactor power is not less than 3% and the MSIVs are l

I contains highly subcooled water. Without operator ac- closed. The operator is told to trip the recirculation tion, this scenario would continue until either high pumps to reduce power," and then to proceed down pressure makeup or PSP condensing capacity was lost. two parallel branches. One branch consists of pro-A detailed analysis of the plant automatic transient is cedures intended for the insertion of control rods.

contained in Section 4. The purpose of this section is These procedures manifest themsel o as attempts to to present probable operator actions within the con- manually scrum the reactor. If the manual scram at-text of proposed EPGs.5 tempts are unsuccessful (which is assumed for these The proposed EPGs were developed by the Emer- simulations), the procedures call for manually insert-gency Procedures Committee of the BWR Owners ing individual control rods. The second branch, which Group to provide detailed guidance for plant operators is executed concurrently, tells the operator to monitor during abnormal or accident conditions. They are the PSP temperature. If the PSP temperature exceeds designed to be symptom oriented, and as such are quite 110"F, he is told to initiate SLCS and prevent elaborate in order to address all postulated plant con- automatic actuation of the automatic depressurization ditions. Those parts of the EPGs that relate to the system (ADS). When the PSP temperature exceeds predicted plant status during a MSIV closure ATWS 110*F (approximately 80 s after transient initiation),

were extracted accordingly ror this analysis. Specific- boron injection is required and the operator is directed ally the EPGs were used as a basis for modeling to enter Contingenew #7.

operator actions during the postulated ATWS scenario. The third path (see Figure 6) which is executed con-A flowchart of RPV control guidelines is shown in currently with RC/L and RC/Q, is entitled RPV Figure 5. It is emphasized that the actual EPGs are pressure [RPV control / pressure (RC/P)]. If any SRV much more detailed. Onh those parts that are perti- is cycling, the operator is told to minimize cycling by nent to the present analysis are discussed in this sec- reducing RPV pressure below the minimum SRV open-tion. Referring Figuie 5, it is seen that entry to the ing setpoint. It is preferable to relieve pressure through EPGs is gained following MSIV isolation. The operator the turbine bypass valves (blowdown to main con-is first told to ensure reactor scram, which would be denser) rather than the SRVs, However, it is assumed automatically initiated. For these simulations, however, here that the main condenser is not available because it is assumed that reactor scram is unsuccessful. Pro- of MSIV isolation. Although not accounted for in this cceding down the figure, the operatm is told to monitor analysis, the EPGs imply that the operator should at- .

I and control (concurrently) RPV water level, RPV tempt to reestablish the main condenser as a heat sink.

pressure, and reactor power. These three paths will This would be accomplished by opening the MSIVs, be discussed individually. which requires that there is no indication of gross fuel While following the path entitled RPV water level failure and no indication of a steamline break. In addi-  !

[RpV control / level (RC/L)], the operator is told to en- tion, boron injection must be required, and the main sure automatic actions. Specifically, this creails con- condenser available, before isolation interkicks should firming MSIV isolation and high pressure system be overridden to open MSIVs. ,

(HPCI and RCIC) initiations. To ensure RPV level DurinF pressure control, the operator may find control, the water level must be restored and main- that either the heat capacity temperature limit (HCTL) tained above the top of the active fuel (TAF). This is or the suppression pool load limit is reached. These accomplished automatically with the recirculation limits are plots of PSP temperature and level respec-pump trip and initiation of the HPCI and RCIC tively, as functions of RPV pressure. Figure 7 illus-systems. Combined with CRD flow HPCI and RCIC trates a plot of the HCTL as applied to the Browns flows result in a reactor power of approximately 30%

and a vessel water level above TAF, An alternate deci-sion branch is provided which tells the operator to pro- a. The recirculatum pumps would have tripped autonmtically on ceed to Contingency #7 if boron injection is required. high RPV presure.

9

m o _

liI T

o _

T S '4 s D g C a DOL O n

R y TnRN T TO NI f

a f AmOT ORC T C R

E l

Nm T

sLAN lU aOD S

l R cVC IR l O E

/ W C O R P R

R3 O

O N

?

D E

S E

N R

O D

l!

E S

N m NuPnO T< O. T h R CR? Q C Y

L O

T AE EW S A E

T C RO V N A P I

S E E

R R

J 1

l E M C U

D

• y N

R O

T C E O N R I T C C CD A C

l PE EE nE EST R Jn E BPA O Nu N T

U j TETI f

T C

I O W C

E fRN AO A E

NE O OR O X p CF R R T E f E T, C u RB

/ N B N t

r c O 0 D

N L

A h

a N TN1 CW1 AO>

ED CL t W

O D

T J

s h l

a c

w o

O R m R T 3W.3 FHE J F S

f l

o T

N O

ST / r t

n C 7

( o N c O O r Y

l I e M T CD v C N

w o

E p R

O ER E G _

T FAI RU !l, lllI:l9 N d n

I N

Q I T a E ._

O R N l e

M O v l

C le

> O G T P

\ E 5

e r

N u D G O ig E P N I

T MT E O C F AI A LF D I U

RTm C T EA VT E T A D E N SNCTA E.

t E

D A D. ZIR R W O

RNE R L S VV L T

N E T RE D OER T N &l L O PO O I A wi n n

T T A

TE CDM E l

/

E V A I

I T RB A RT MOI T _

J H

C E T C s e

_ A TA T I A NR N E MN S DF m_D R L AO O E R S RO C I C aA V D

_ NS R I

F Q R E $ N" O T !

L L 1y ASt NTA G P f L O

FS R O A CEHVN T

A H O A

W R[

M R

U A

E Ot E E V V E. E L

VRR LOE P

n r

C O

T V

I F

4 J FL EDMO V IT tx. r P 7 e AA V EAT Um i D R X[ M P L 9A w E

/ / C D CR R V T E C

P DATB Y T

RO N O

)S ON T R N O R C N S T E C P R O E V (I

N R A' F E R W L!T O F O OO R IVIT L

L TO SA.** A NC ML O O CY S R I VT PN RE

\

5

R O E

/C WO R P 0 R O

Y T N L C W T A O N E R D E R O L V

_ R O P l

R l

U ) D. O OR C S ETR OD C N G

N N N H M T RV .N N EPN A &W O I

E T W S E O L P

O O

_ C l O G IWETS RE SRWN E D O D G

_ E .N N GWI VY U ELS S NNOW DO US O RL C N T U

_ T E ]NO L

C RA S SITD O N IL F U P R LVRE ER L D Y S DNJTVS C W O S

_ C E / U CWC S OOFUPE t O O

_ X C S R S YO ESH P R ST RRSHRR P o D

TE C D L

CL V PAO O S P( P V t

E E A l( U C V hO OR LBD LS NV V V D h RP C

_ S ID S Y R O Hlv P

_ N NE VBY P RRTOI TNC NR P EZ SR R O A V R EB LO E M P VAZ R ENE RO U R H S R T

T F

CITORE d fu

- ND T R I

L R SV M LIE ICICTOSC D O RNI E F ECBA R N TE E R YSI CBT N d.

I E

T l N M ~ RRN TU O NLLFFA FEOR E e GE G C N A NT E C EASERNC t

AD N O O O A H E M. LO O R C F C U C

A W** = D u t PM L P

- ot l

l A G R L V O

T C/ E L I

N hE R R O T i ! ! i*j j I9 i.I j

l j lj A A W

_ > V

_ P R

T W T O l

_ Yu! f L E

TU I

f AB C

AE D C 1 D MG PRE D E AP AUF T C LE 7EF MTA AIRC RE rC AAE V TA D E A RTOSO CIT T TRR AE L

NO N P PA SICR EM AOY RNE T R F

E SM P

T OERCT TE AN AT E B

E CBME A AR T ET R OR ES FS NCSRI LF R

1!  !

• I il S

A NN H EO <

T NS ED W NA P

O AR CN VE L TO R

(O P B NF R O N CE LITI NE L

OD O E B RN !VI T AL TSST EA FL T O 1V SA .

• LI NC ML N EI SA C O AR M SV I

CY SI MPL AA R

VT PN RE

\

-~

210 4 i i 8 V-v e

u 200 -

)

i D j l

O i b

a.

190 - -

E OPERATION ABOVE THIS o / CURVE SHOULD BE AVOIDED 180 - -

]

1 O

O n.

c 170 - -

.9 E

e k 160 - -

e in 150 ' ' ' '

O 250 500 750 1000 1250 RPV Pressure (psio)

Figure 7. Heat capacity temperature limit.

Ferry containment. If either limit is reached, the reaches TAF, it is to be maintained there with available operator is told to reduce RPV pressure to stay below high pressure systems. This action is intended to reduce l the limit. reactor power by reducing core inlet flow. If the water i For the simulations presented he7, the operator level cannot be maintained at TAF, an alternate branch would be monitoring and controlling RPV water level, indicates that depressurization is required. This would '

RPV pressure, and reactor power. When the PSP allow the use oflow pressure systems to maintain level temperature reached 110*F,8 the operator would be at TAF.

instructed to initiate SLCS and proceed to Contin ~ So far, the actions discussed have been concerned pency #7 of the EPGs. From the previous discussion, with RPV contrel. In addition to RPV control, the Contingency #7 is entered from paths RC/L and RC/Q. operator would also be concerned with primary con- i RPV pressure would continue to be monitored and con- tainment (PC) control. Figure 9 illustrates actions from l

trolled according to RC/P. the primary containment guidelines, which are perti- l Tne expressed objective of Contingency #7 is to nent to the transients documented in this report. As with

" minimize heatup of the suppression pool during the the RPV guidelines, the PC guidelines presented here time of boron injection, thus avoiding the need for are considerably reduced from the actual EPGs, emergency RPV depressurization." Figure 8 illustrates For the ATWS, entry to the PC control would oc-pertinent actions and decision steps extracted from cur when the PSP temperature exceeds 95*F. At that Contingency #7. Following the entry conditions, with time, the operator is told to monitor and control con-power >3 %, the operator is told to terminate and pre- currently drywell temperature (DW/T), suppression vent all injection into the RPV except CRD and SLCS. pool temperature (SP/T), primary containment pres.

This action will cause the RPV water level to drop sure (PC/P), and suppression pool water level (SP/L).

I because steam lost through the SRVs will exceed Only the SP/T path requires attention past the first ac-

! makeup from CRD and SLCS. When the water level tion step, and as such is the only path detailed in Figure 9.

For the DW/T path, the drywell coolers are assumed l

a. This temperature is predicted to be reached approximately 78 h operable. Because there is no containment pressuriza-af ter transient initianon. tion calculated durieg the operator action transients.

12

\fROM RC/O[

I: \FROM RC/L /

F OORON INICTION REOUIRED .

I l

I (CONTIMENCY 1

1 IF REACTOR POWER > 3%

OR , . .

l TERMNATE NO PREVENT ALL IKECTION NTO RPV EXCEPT CRD NO SLCS REDUCE REACTOR POWER 4TIL RPV WATER LEVEL REACFES TAF MANTAN RPV WATER - ~ ~ ~ - - - - - -- ---- - - - - - - - - - - - -

LEVEL AT TAT RPV LEVEL CONTROL

........ .........q

- i WHEN StJFICIENT BORCid HAS BEEN IMICTED TO PLACE REACTOR IN HOT SHJTDOWN {

OR ALL RODS ARE IN F RPV i a RESTORE NO MANTAN LEVEL CANT BE i NCHMAL RPV WATER LEVEL MANTAPED AT TAF 4

ENERGENCY RPV DEPRESSURIZATU4 REQUIRED MINIMlZE CORE DAMAGE j TERMNATE NO PREVD4T ALL NECTION j NTO RPV EXCEPT CRD N O SLCS i U4Tk. RPV PRESSLEE < NNIMN i ALT RPV FLOODING PRESSLEE i

5 SLOWLY INICT INTO RPV g

o o RPV LEVEL CONTROL NO CORE COOL ,

-T COto SUTOOWN

.j I

Figure 8. Contingency #7 extracted flowchart, i

13 l i

8  !

E f 8 f! ll  !

g - ...

j ll l!

s

- b i 3

} s I r 1 by E l?, s

~

n

=

f

~ **

p ]

s e 3

l.u t i

!g8 4 l jii e si i !

g9. . . . l, gg! c t

-lEl ! g Il ! e}g 8 _

~

E $ l hI ~

l

~5!! j,! lgll 'Ns)l' l di!l~ l lEl

$ E $

llh

_ l l l lh l _ _..

I 1118 14

_____-__-_a

path PC/P requires only monitoring. Similarly, path The timer is a LOCA consideration to ensure that LPCI SP/L is only monitored because the PSP water level is available to provide makeup to the RPV if necessary, is calculated to remain within normal operating limits. After five minutes the operator may realign for pool Following the SP/T path, the operator is told to cooling. However, if the RPV water level rises above operate suppression pool cooling when the PSP LLLWL, the timer will reactivate each time it passes temperature exceeds 95'F. With the PSP initial below the setpoint and prohibit pool cooling. Because temperature assumed to be 90*F, this occurs early this level movement is predicted to occur, no credit

(<25 s) in the ATWS. It is unlikely that the operator for pool cooling was taken in these simulations. The would have time to establish suppression pool cooling possibility for RHR alignment to pool cooling is 11-(torus cooling mode of the RHR system) during the lustrated in Figure 10.

first few minutes of an ATWS as explained below. At Continuing down path SP/T, the operator must approximately 220 s, the RPV level would be lowered monitor PSP temperature and keep it below the HCTL.

to TAF in accordance with lesel control procedures This is the same action required as in the RC/P path in Contingency #7. When the RPV level passes discussed earlier (see Figure 6). Guidelines for second-through the triple lo level water level (LLLWL) set- ary containment and radioactivity release control are point (Figure 2), the RHR pumps (Figure 4) are also provided bv the EPGs. However, their evaluation automatically aligned for LPCI injection and inter- is beyond the scope of this analysis, kicked on a five minute timer. This prohibits realign- The effectiveness of these guidelines in mitigating ment to pool cooling even though LPCI is not needed. an ATWS is discussed in Section 5 of this report.

1.loE LP FOR RFft AUTOMATIC SYSTEM SYSTEM MAY EE RE-ALIGNED POOL COOLING POSSIBLE ALIGNMENT FOR LPCI FCR POOL COOLING IF WITH OPERATOR ACTION INJECTION ONLY DOWNCONER LEVEL DOES t4T PASS THROUCH LLLWL OPERATOR DOWNCOMER LPCI TINER TRANS!ENT ASSOES LEVEL AT TIMED OUT INITIATION LEVEL CONTROL TAF (RH1 ALIGNENT POSSIBLE)

I f f l 0.0 120.0 220.0 520.0 (Trne not to scale)

TIME (1) j NSTOO3M Figure 10. RHR alignment.

l l

15

4. ANALYSIS APPROACH This section describes how significant parameters, 50 to 56 gpm. Assuming that 265 pounds of boron operator actions, and accident sequences wete isolated solution is required for hot shutdown (see Refer-for analysis. This isolation procedure was necessary ence 3), the SLCS would have to inject for ~24 min-because of the many and varied combinations of utes. While injecting boron, the operator would be operator actions and accident sequences that could following procedures outlined in Contingency #7 of the possibly result during a MSIV closure ATWS. EPGs.

The objective of Contingency #7 is "to minimize 4.1 Operator Actions and heatup of the suppression pml during the time of bomn System Interactions in.iection, thus avoiding ge need for emergency RPV depressuriz.ation. This is because ' reactor instabil-During a MSIV closure ATWS, the primary respon- ities are likely to be associated with a blowdown at sibility of the operator is to effect a safe shutdown. To power, (and) it is desirable to have the reactor shut-be successful, the operator must either insert control down prior to depressurization."5 The reactor system rods into the core or poison the core with the SLCS. is not necessarily stable because low pressure vessel The ability of the operator to fulfill this responsibility water inventory systems can fbod the RPV if depres-is dependent on the availability and reliability of the surized, thus inducing reactor power excursions and sarious plant sy stems, and on the state of the reactor. increased steaming rates to the PSP. According to the The effectiveness of the sy stems depend on the ability plant automatic chronology presented in Section 5.1, of the operator to recognize the signature of the tran- the vessel could be automatically depressurized as early sient and take appropriate mitigative actions in a timely as 18 minutes alter transient initiation, manner. Contingency #7 proposes to avoid RPV depressur-Imtially, the operator would attempt to scram the ization while borating by reducing the lesel in the reactor. The EPGs proude guidance for alternate downcomer to the lowest practical level. TAF. Accom-scram procedures. Depending on the cause of the panied with a corresponding HPCi throttling technique, original scram failure, the operator ma) very well be this action should reduce the reactor power. This ae-sueecssful and shutdown the reactor. Ilowever, for the tion of level control is sery important. Normal HPCI purposes of this analysis. it is assumed that reactor system flow alone induces a sessel steaming rate of scram cannot be whievcd. The operator woulti then approximately 209f of normal steandine now to the be expected to anempt other means of inserting con- PSP. Level control should reduce this rate, and delay I trol rods. One such procedure is termed manual rod the PSP heatup rate. If the PSP gets hot enough even- l insertion, and insolves inserting control rods in- tually the llPCI will fail on high turbine lube oil dividually from the control room. This procedure was temperature." Thus, by reducing the PSP heatup rate not accounted for because of the uncertainty in model- with level control, HPCI failure due to high PSP ing individual rod reactisity. A discussion ofits poten- temperature would be delayed or avoided. The impor-tial effectiveness is included in Section 7.3. tance of preserving 11PCIintegrity is related to boron Scramming the reactor is the simplest and most ef fectiseness. If HPCI should fail before shutdown reliable means of transient mitigation. In the absence from boron occurs, it is unlikely that the remaining ,

of control rod insertion, the operator must initiate the high pressure systems (RCIC and CRD) would be suf- l SI CS to effect a shutdow n. The success of this system ficient to efHeiently transport baron into the core. The is howes er, contingent upon system effectiveness and combined Dows of HPCI, RCIC, and CRD result in the presence of other, complementary operator actions. a core inlet now that is approximately 109i of normal These actions are directed to maintaining core cool- How. RCIC and CRD result in a core inlet Dow that ing and containment integrity while the operator is approximately l q of normal flow. Following HPCI borates the core. For example, enough core How to failure, RPV water level would drop below TAF, transport the boron into the core must be provided. If necessitating emergency RPV depressurization per that is not accomplished, the boron could stagnate in EPG direction.

the lower plenum and have little or no af fect on re-ducing power.

At Browns Ferry, the SLCS must be manually ac- a. Once IIPcl shihs suction m the PSP on high PSP water level, tivated. It is designed to inject ~139; by weight of "urNne lobe ou n, uioled w uh PSP w ater An estimate of 19n"I-in the PSP was used to predict iiPCt failure While this value is sodium pentaborate solution into the lower plenum of beliend m be conservatne, ilPCI failme wwlJ occur at some the RPV through a single sparger at the rate of clesated psp temperature.

16

While controlling level at TAF, the operator would 1. Sequence 4838 is essentially a plant automatic also be monitoring PSP temperature. If the HCTL transient. The high pressure ECC systems main-(Figure 7)is reached, the operator would be directed tain core cooling; however, continual large by the EPUs to depressurize to stay below the limit. steaming rates eventually threaten containment If depressurization is required before the reactor is integrity.

shutdown, reactor instabilities are possible. In addi-

2. Sequence 551 describes a transient with early tion, unless the low pressure injection systems are failure of HPCI and no RPV depressurization.

throttled or their mjection prevented, large power ex- A high pressure boiloff ensues, leading to core cursions could occur when they flood the RPV. damage. (Subsequent analyses have indicated The objective of the simulations presented in Sec- that RCIC and CRD flow together may be suf-tion 5 is to answer questions about the plant respons f cient to limit fuel rod heatup.)

to a MSIV closure ATWS. In particular, how effec-tive is the SLCS in accomplishing shutdown? How ef- 3. Sequence 465 is similar to Sequence 483 except fective is level control in reducing reactor power? Can that the operator takes level control by throttling an emergency depressurization be avoided? How much the HPCI. Core steaming is reduced; however, time does the operator have before taking action? What containment integrity is threatened.

are the consequences of emergency depressurization

4. Sequence 482 is similar to Sequence 483 except The progression of an ATWS sequence is com the operator takes pressure control. Again, core plicated by the interdependence of operator actions, steaming is sufficient to pose a threat to contain-plant systems, and phenomena. Operator actions are tuent integrity contmgent on the operator's perception of the state of the reactor system which, conversely, depends on operator actions. Because of this complexity, a prob. 4.3 Application of the SET abilistic rather than a deterministic analysis approach Evaluation ot the four most likely sequences from was adopted to isolate sigmficant phenomena, events, the SET revealed a striking similarity between three

"" "W#** of them. The similarity is best described in terms of consequence level. A qualitative discussion of the plant 4.2 Sequence Event Tree automatic transient fonows. which delineates this point.

e p ant automat c transknt @ qu q 483, From the preceding subsection. it is apparent that described quantitatively (m Section 5.1) begins a MSIV closure ATWS could proceed through many mnim n tr ns nt imtiamr,ix. M cbsum, suam different paths. For that reason, a structured approach fanum, p u ate as gne , and m&ula&m to the analysis of ATWS sequence progression was re-pumps trip. In the absence of boration, the RPV steams quired. Specifically, it was desired to formulate prob-through the SRVs to the PSP at ~21 % of rated flow.

abilistically the frequency of occurrence of sequences a eup is pmvided by the HPCI, RCIC, and CRD and the expected consequences. To that end, a SET systems. HPCI will shift suction from the CST to the was developed PSP on an indicated PSP high water level. HPCI even-The SET used traditional event tree analysis to depict logically the sequence of events that must occur to jually fails due to high PSP water temperature, which is used to cool lube od for the HPCI turbine bearings.

recover the plant or to induce a severe accident. Ir all Following HPCI failure, reactor power is reduced to cases, entry to the SET involved four assumptions: the less than 5% as RCIC and CRD provide the only MSIVs close and are not reopened, the control rods

"" "E' fail to insert, the SRVs function as designed, and the PSP heatup continues, which pressurizes the PSP, recirculation pumps trip automatically. Subsequently, and pressure is reheved to the drywell. The high dry-the SET encompassed plant phenomena, plant systems, wcH pmssum wmoined with the already low down-operator actions, and EPGs. Reference 4 describes comer level trips the ADS. Following a 120 s delay, each event tree heading and corresponding success the RPV depressurizes through the ADS valves to the criteria FSP. Low pressure ECC systems flood the RPV Four sequences were identified in the SET as being resulting in p wer excursions, which increase the most likely to result in core and/or containment dam-age. All four satisfied the entry conditions described in the previous paragraph, and all four included failure of boration and failure of manual rod insertion. Unique a. Sequence nunbers are pndd for constency with Reference 4.

'pressure

" '" **relief ""'""*"'

features of the four most likely sequences are outlined valvesd#P'*"" *"* ","Y"'"'

open autonmucally relieving RPV prenure below. to the Pse.

17

I steaming rate to the PSP. The ADS valves will close situations, the downcomer level may recover on high containment pressure (115 psia). The RPV such that HPCI and RCIC are tripped off. A then pressurizes above the low pressure ECCS shutoff manual reset is required to restart RCIC injec-heads. HPCI has failed, RCIC has isolated (at 40 psia tion, drywell pressure), and containment failure is imminent'

3. During simulator sessions conducted by the Oak Core damage is expected from the assumed loss of all Ridge National Laboratory (ORNL) at the Ten- l ECC5 mjections following containment failure.

V Mo VR h was observed j The level control transient (Sequence 465) differs i in s me situations that the operator would shif.t from the plant automatic transient only in the timing l RCIC suction to the PSP following the automatic of events. It is assumed that at some time the operator HPCI shift. This would lead to RCIC failure on lowers downcomer water level by throttling HPCI. .

high lube oil temperature.

This action reduces the RPV steaming rate, and delays the time of HPCI failure. Following HPCI failure, the 4. If the primary containment pressurizes above events are identical for the two transients albeit shifted 40 psia, high turbine exhaust pressure will in time. isolate the RCIC turbine.

During the pressure control transient (Sequence 482)'

The plant automatic transient provides a description the operator depressurizes the RPV m accordance with MSIV closure ATWS without operator action, it the heat capacity temperature limit. The depressuriza-exhibits the same consequence level as transients with tion results in a power reduction during HPCI opera-level or pressure control only. In addition, it was found tion. However, with pressure control, the low pressure that the transient degraded into a high pressure boiloff ECC systems vill start mjection when RPV pressure when the ADS valves close because of high contain-drops below the shutoff head of the pumps.

ment pressure.. The RPV then pressurized above the Although shifted in time, the three transients exhibit

"* E""" "1"'#* "I"" #" "'# ""

the same consequence level. This consequence level

. RCIC were no longer available, a high pressure boiloff is characterized by containment failure, and subsequent ensued. For that reason, the plant automatic transient expected core damage (see Reference 4).

was chosen as the vehicle for describing the dominant The SET formulation of the high pressure boilof.t.

sequences identified by the SET. Results from the plant transient (Sequence 551) assumed that HPCI failure was sufficient for entry into a potential boiloff situa- ".utomatic transient simulation are presented in Sec-tmn 5.1 (RPV and primary containment) and Section 6 tion. Subsequent analyses showed that RCIC and CRD Uuc damage).

flows were sufficient to delay and limit core heatups.

Four situations were postulated, which would result in RCIC unavailability and a resulting core boiloff.

a. In the rehef mode of SRV operatum Oncludmg ADS), air pressure

1. Because the RCIC system is nonsafety Frade is apphed to the air actuator by energiring a solenoid operated con-there is a remote possibility that it would be out noi air salve. The resultant pressure differential across the bellows of the remote air actuator causes the SRV to open. Because the com-of service at the time that the MSIV closure prehsor supplying control air hes outside the primary containment.

ATWS occurred. when the pressure inside the containment rises to the pressure of the cantrol air, the remote air actuator will no longer be capable

2. During combined HPCI, RCIC and CRD flow of maintaining the SRV in an open position.

l 1

l l

i 1

i I

1 l

J 18

5. RPV AND PRIMARY CONTAINMENT RESULTS Results from several MSIV closure ATWS simula- NRC final rule on ATWS)" is discussed in Sec-tions are presented in this section. The simulations were tion 5.2.4. Table 2 summarizes the simulations 6

performed witt RELAP5/ MODI.6 and CONTEMPT / presented.

LT-028.7 RELAP5 was used to model the RPV and associated internals, the feedwater line, and the main steamlines. Also modeled were the SLCS, CRD 5.1 Plant Automatic Simulation system, and ECC systems. Appendix A contains details The plant automatic simulation illustrates the auto-of the RELAPS modeling. CONTEMPT modeled the matic progression of a postulated ATWS in the absence primary containment, specifically the drywell and PSP. of operator actions. Without mitigating actions, the Also modeled were associated heat structures and flow containment could fail by overpressurization and it has paths. The CONTEMPT model is discussed in more been common practice in PRAs to assume that this detail in Appendix B. would lead to the loss of the ECC systems. Loss of The simulations were performed in an iterative injection to the RPV leads to core uncovery and subse-manner, RELAP5 provided boundary conditions to quent fuel damage. This simulation is bounding in that CONTEMPT in the form of mass and energy rates it provides a realistic estimate of the minimum time through the SRVs. CONTEMI'T in turn provided event to containment and fuel damage.

timings back to RELAP5. Figure 11 illustrates the in- Events occurring during the early portion of a MSIV -

formation exchanged between the two codes. closure ATWS (Table 1) are well defined. MSIV Five MSIV closure ATWS simulations are pre- closure causes a rapid pressurization of the RPV sented. The plant automatic transient is discussed in because the core continues to produce steam at near Section 5.1. Section 5.2.1 discusses an EPG nominal rated conditions. Increased pressure collapses voids in j

simulation with maximum SLCS effectiveness, and the core which provides positive reactivity feedback.

Section 5.2.2 considers the case with minimum SLCS Increased power results, which increases the core effectiveness. Section 5.2.3 kioks at the effect of not steaming rate. The increasing pressure causes the recir-depressurizing the RPV when the HCTL is reached. culation pumps to trip (at 1135.0 psia), and the SRVs The effect of increasing SLCS capacity to 86 gpm (per to open.

RELAP5  :

NSS PARAMETERS: CONTAINMENT PARAMETERS:

HPCI, RCIC TURBINE EXHAUST EVENT TIMINGS SRV, ADS FLOWS

• HPCI SUCTION SHIFT

• MASS FLOW RATES

• HPCI FAILURE

• FLUID ENTHALPIES

• HCTL MONITORING

= LIOUID LEVEL MONITORING ECC EXTRACTION MASS FLOW RATES

• HDWP

• ISOLATION PRESSURES (RCIC, ADS)

CONTEMPT l L226-KW263-05 Figure 11. Information exchange between RELAP5 and CONTEMPT.

19

l l

Table 2. MSIV closure ATWS simulations I

Boron Level Simulation Description Modeling Control Depressurization l

l 1 Plant automatic None No Yes (automatic ADS) .,

2 EPG nominal-maximum 50 ppm Yes No (pressure control, )

isotropic not depressurization)

SLCS effectiveness l

3 EPG nominal-minimum 50 Fpm Yes Yes (manual blowdown) l SLCS effectiveness stratified 4 Minimum SLCS 50 gpm Yes No effectiveness without strati 0ed i depressurization 5 EPG nominal- 86 gpm Yes No increased SLCS stratified capacity and minimum effectiveness MSIV closure isolates the steam supply of the feed- is used to cool the lube oil for the HPCI turbine bear-water turbines from the RPV. As the feedwater tur- ings. These bearings will eventually fail as the PSP bines and pumps rapidly coast down, turbine-driven water temperature continues to increase, resulting in feedwater is lost. Continued core steaming causes the the loss of HPCI How. For the purposes of this simula-vessel downcomer level to decrease. When lo-lo level tion, it was assumed that HPCI failed when the PSP (476.0 in.) is reached, HPCI and RCIC systems are temperature reached 190*F (at 830 s),

activated. Taking suction from the CST, HPCI pumps Following HPCI failure, the core power decreases 5000 gpm and RCIC pumps 600 ppm of high pressure as only RCIC and CRD flow enter the RPV and down-makeup into the RPV via the feedwater line. For the comer level decreases. Together, they provide suffi-purposes of these analyses, it was assumed that a scram cient core flow to maintain power at 3-4% of rated, signal was penerated but no control rods were inserted. Even though the downcomer water level drops below Following the scram signal,112 ppm of water was TAF, sufficient core cooling is provided to delay rod pumped into the RPV by the CRD system. Figure 3 cladding heatup from occurring.

illustrates the flow alignment described above. The PSP pool temperature continues to increase, but Following recirculation pump trip, the reactor at a slower rate following HPCI failure. The bulk PSP stabilizes in a natural circulation mode. Core power temperature is calculated to reach 200'F at 924 s, and is determined from the mass flow rates and enthalpies at that time complete condensation is no longer en-of the injected water. The resulting power with HPCI, sured. When the RPV steam is no longer completely RCIC, and CRD injection is ~30% of rated. This condensed in the pool, some of it will flow through causes a steaming rate of ~21% of normal steamline the pool directly to the PSP atmosphere. The pressure flow to be relieved to the PSP. The difference in core in the PSP increases and is relieved to the drywell. The power and RPV steaming rate is accounted for by the drywell in turn is pressurized, with a high drywell heating of the highly subcooled ECC fluid. pressure (HDWP) signal occurring when the drywell j The RPV stabilizes to a quasi steady state condition reaches 2.45 psig. This is predicted to occur at 985 s in less than 100 s after transient initiation. Steam pro- based on linear departure from complete condensation duced in the core is relieved to the PSP through the at 200'F "

cycling of two to four SRVs (SRV setpoints are listed Two signals are required for initiation of the ADS:  ;

l in Table A-5). As the steam is condensed in the PSP, hwkulo water level (Figure 2) in the RPV and HDWP.

the PSP heats up at the rate of ~6F'/ min. At 255 s, l the PSP water level has risen to the upper limit of its i

a. For this matel, complete condensation is calculated in the pool normal operating range (181.25 in.), This causes the

. . until a reaches 200 F. Condensation efficiency is then calculated HPCI suction to automatically and irreversibly shift to decrease linearly to rero as the pool approaches saturation l from the CST to the PSP. After the shift, PSP water temperature (N218"F). l 20 1 1

I l

______o

Low water level is predicted to occur early in the tran- cycle of core flooding, power excursion, and pressure sient (<100 s). Later, the HDWP signal initiates the relief thus repeats itself.

ADS timer, and the ADS valves open following a time Primary containment pressure increases rapidly as delay of 120 s. At i10.5 s. the six relief valves a result of the power excursions. At 1800 s, contain-dedicated to the ADS opea and rapidly depressurize ment pressure reaches 40 psia, which isolates RCIC the RPV. When RPV pressere reaches the shutoff on high turbine exhaust pressure. At 2100 s, contain-heads of the low pressure ECC systems and conden- ment design pressure (71 psia)is reached. After seven sate booster pumps, they flood the RPV with cold water power cycles, the ADS valves close when containment (210*F). Shutoff heads for LPCI, core spray, and con- pressure reaches i15 psia at 2470 s.

densate systems are 346, 355, and 418 psia, respec- With the ADS valves closed, the RPV remains at tively. The design capacities of LPCI, core spray, and high pressure (~1100 psia) with SRV cycling. This the condensate booster pumps are 40,000,12,500, and prevents the injection of water by low pressure makeup 30,000 gpm, respectively. RPV flooding results in a systems. Because HPCI has failed and RCIC has large positive reactivity insertion. Core power in- isolated, the CRD system provides the only injection creases rapidly, producing steam faster than it can be to the RPV. The downcomer water level drops to the relieved through the ADS relief valves. The increased jet pump suction elevation as water is boiled off in the steam production raises RPV pressure above the shut- core. Pressure is maintained around i100 psia by the off heads of the low pressure injection systems, cycling of one SRV. The continued core steaming is terminating their injection. The pressure increases until relieved to the primary containment, which is predicted the remaining SRVs !!ft. With low pressure injection to reach failure pressure (132 psia) at 2700 s. CRD to the RPV stopped, moderator voiding in the core flow alone is not sufficient to maintain core cooling terminates the power excursion. RPV water level even at very low decay power levels. As a result, rod drops, and pressure is relieved through the ADS valves heatups are predicted to begin at approximately 4000 s.

which are automatically held open by control air. When RPV pressure during the plant automatic simulation RPV pressure drops below the low pressure injection is shown in Figure 12. The predicted peak pressure system shutoff heads, they again begin injection. The of 1272 psia occurs at i1 s. Pressure is maintained near 1500 i i PEAK RPV PRESSURE 1272 psio o ADS s .

SRV CYCLING

~

n a M} t i

I ,

M%WM%MWMMWMWW%%

1000 -

• ADS VALVES m

s CLOSE E

o L

CL L

.3 500 - -

8 o

O' 0

O 2000 4000 6000 Time (s) l l

FiFure 12. Plant automatic transient-RPV pressure, j l

21

1100 psia by SRV cycling until ADS blowdown begins the power excursions, with failure pressure (132 psia) at 1105 s. The pressure drops to almost 200 psia before predicted to occur at 2700 s.

it turns mund from increased core steaming caused With the fuel rod cladding heating up due to lack by low pressure injection to the RPV. Following of coolant, the simulation was terminated at 5600 s.

repressurization, the cycle of blowdown, RPV Dood- At that time peak cladding temperature was ~1600 F, ing, and repressurization repeats. In a!!, seven cycles which is above the threshold temperature for cladding were predicted until 2470 s when the ADS valves oxidation. At that temperature, zircaloy fuel rod clad-closed. Following ADS valve closure, pressure was ding will react with steam to oxidize the cladding and again maintained near 1100 psia by SRV cycling. produce hydrogen. Core inlet conditions were taken The predicted core power is shown in Figure 13. from the RELAPS calculation and input to the Severe MSIV closure causes a peak power of 253% at 3 s Core Damage Analysis Package (SCDAP). SCDAP (subsequent pressure spikes ar caused by low pressure was then used to predict fuel damage during the high injections). The power levels off following recircula- pressure boiloff. Results from the fuel damage analysis tion pump trip to 30% as the high pressure systems are presented in Section 6.2.

maintain injection. HPCI failure at 830 s causes a in addition, the possibility of fuel damage during the power reduction until low pressure injections begin at power excursions caused by low pressure injections 1240 s. The seven predicted injection cycles are clearly was examined. A best estimate fuel rod analysis pro-illustrated with the corresponding power excursions. gram (FRAP-T6) was used for that analysis, and results Following ADS valve closure at 2470 s, the power are included in Section 6.1.

decreases to decay levels with only CRD flow injec- Table 3 lists predicted event timings after 100 s for ting to the RPV. the plant automatic transient. It is realized that the latter PSP pool temperature steadily increases as shown stages of the plant automatic transient could very well in Figure 14. Th: temperature reaches 200*F at 924 s, proceed differently than in this simulation. For in-and saturation temperature (218'F) at 1220 s. Fig- stance, containment failure could possibly affect RPV ure 15 illustrates the predicted drywell pressure pene:r mns, resulting in RPV depressurization. These response. The pressure increases rapidly as a result of possiinties, however, are beyond the scope of this 400 , i O

v u 300 - -

o Q-

[ PEAK POWER 2537.

o L

o 200 - -

l o

R

=

0

[

O 10 0 - -

Z t . bl 0

0 2000 4000 6000 Time (s)

Figure 13. Plant automatic transient - normalized core power.

22

400 i ,

p v

300 -

o w

3 SATUR ATION o -

200 -

a E

e F- -

O.

U) -

a. 10 0 0

0 '.000 2000 3000 i Time (s) j Figure 14. Plant mitomatic transient - pressure suppression pool temperature.

1 I

I 150 I i

.o-m O 10 0 - _

o u

a e .

M j o l Q.

=

o 50 - -

m A

L o

HDWP

' f 0

0 1000 2000 3000 Figure 15. Plant automatic transient - drywell pressure.

23

Table 3. Plant automatic transient event timings after 100 s a (s) Event 0 Transient initiation 255 HPCI suction shift to PSP 830 Assumed HPCI failure (PSP pool at 190*F) 985 High drywell pressure (2.45 psig) 1105 Automatic depressurization (ADS) 1220 PSP at saturation temperature 1240 First low pressure injection 1270 First power excursion 1800 Drywell pressure at 40 psia (RCIC isolation) 2l00 Primary containment at 71 psia (design pressure) 2470 ADS valves close 2700 Primary containment at 132 psia (predicted failure pressure) 4000 Rod heatup from boiloff begins

a. Event timings as predicted by RELAP5/ MODI.6 and CONTEMPT /LT-028.

analysis because of uncertainties inherent in the se- dition would be reached autonutically in less than 100 s quence progression. For the purposes of this study it following transient initiation. During this time, the was assumed that the CRD hydraulic system remained operator has entered the EPGs (upon MSIV isolation) operational throughout. but all other injection systems and is monitoring the status of the plant. While fail when the containment fails, monitoring reactor power (RC/Q path on Figure 5),

the operator would be directed to initiate SLCS when 5.2 Simulation of Operator the PSP temperature reached 110*F, which occurs 78 s Actions after transi nt initiati n. It is assumed, h wever, that operator actions will not begin before 120 s. This is The plant automatic transient described in Sec- believed to be a reasonable estimate of the minimum tion 5.1 illustrated that containment and fuel damage time required to assess the situation and act appro-are expected to occur during a MSIV closure ATWS priately.

without operator actions. Describing the automatic This simulation assumes maximum SLCS cffec-response to an ATWS is valuable in understanding tiveness. When the boron soluti.m begins injection, the possible event progressions. It also provides an esti- 50 ppm How is modeled to be distributed isotropical-mate of the minimum time to damage. It is, however, ly in the RPV. This is accomplished by assuming that more realistic to assume that the operator would take the boron solution is transported with the liquid in the actions to mitigate the accident. The effectiveness of RPV. It is acknowledged that this modeling is idealistic the operator to shutdown the reactor in a timely manner and will result in the maximum possible effect of the is the subject of this section. It is recognized that the boron injection. Section 5.2.2 presents a simulation of timings and extent of operator actions are infinite. For the same transient, but the boron injection is modeled the purpose of these simulations, a series of actions in such a way as to minimize its effectiveness.

were modeled. That series was termed "EPG nominal,' At 120 s, when the operator first takes action, the and is described in the folhiwing paragraphs. core power is ~300 HPCI, RCIC, and CRD dows are providing 5712 gpm of high pressure makeup from 5.2.1 EPG Nominal-Maximum SLCS Effective- the CST to the RPV. Steam is being relieved through ness. As described earlier, a quasi steady state con- the SRVs to the PSP, which has a pool temperature 24

of 117'F and is heating up at ~6F'/ min. As directed tingency #7) to restore and maintain normal water by the EPGs, the operator initiates SLCS and enters level. This is simulated by injecting HPCI and RCIC ]

Contingency #7. It is also assumed that at 120 s the at full now. The level recovers to normal level at operator terminates HPCI and RCIC Dow as directed ~2000 s. Figure 17 shows the combined injection rates by Contingency #7. In line with the. RC/P control path of the high pressure systems.

(Figure 6), pressure control by the operator is also The effect of boron injection combined with level assumed. Pressure control means to reduce pressure control is dramatic. Core power (Figure 18) decreases to eliminate SRV cycling; it does not mean to less tnan 10% by 200 s. and approaches decay levels depressurization. by 600 s. The PSP heatup rate is reduced considerably Figure 16 illustrates the effects of these actions on as a result of the low power levels. As shown in the downcomer water level. When HPCI and RCIC Figure 19, the PSP pool reaches a maximum temper-are shut off, the level rapidly falls, reaching TAF ature of 144 F when the simulation was terminated at (366 in.) at ~220 s. The operator, following Contin- 2800 s. After RPV level recovery at 2000 s, it is gency #7, would then start HPCI and RCIC injection anticipated that the operator would be able to initiate (at reduced rates) to maintain level at TAF. The over- the torus cooling mode of the RHR system. This would shoot at 450 s and subsequent decrease to the TAF terminate the PSP heatup and begin cooling down the level are indicative of the expected time response of pool, the level to changes in injection rates. At 1585 s,265 lb of sodium pentaborate has been injected in the RPV, an amount sufficient to place the reactor in hot shut- 5.2.2 EPG Nominal-Minimum SLCS Effective-down.a At that time, the operator is directed (by Con- ness. When SLCS is initiated at 120 s and the boron solution moves isotopically in the RPV, the reactor 8- ^ C "S'"*'i9 estimate f the toron mass in the RPV required is calculated to be shutdown with a PSP temperature for hot shutdown is 265 lb. At 50 gpm. this would require epprox-imately 1465 s of injecuon. Thus with SLCS initiation at 120 s.

of 140*F. In reality, the boron solution may not be the required amount would be injected 1585 s after transient initiation. well mixed and could stratify in the lower plenum 600 i i i

O

[500 0

HPCI, RCIC OFF o

_J s

v E

o U

c400 m

o O ~ _ . _______________ -.________T_A_F_____________

HPCI TH90TTLED ON HPCI AT FULL FLOW 300 O 1000 2000 3000 Tirne (s)

Figure 16. Isotropic boron - downcomer water level.

25

6000 i i r HPCI + RCIC + CRD m

E o.

v"

• 4000 -

O CC C

,0 _ HPCI THROTTLED I

~ ,

o '

0 i m i C l

- 2000 -

e u

o CRD + SLCS 0~

l 0 1000 2000 3000 Time (s)

Figure 17. Isotropic boron - injection rate to the RPV.

10 0 , ,

2 PEAK POWER:

2537. AT 2.7 s u 75 -

o R

O CL o

L.

o 50 -- -

O y SLCS INITI ATION N

=

0 h

[

O 25 di' -

Z l

1 0 i 0 1000 2000 3000 Time (s)

Figure 18. Isotropic turon - normalized core power.

26

16 0 i i

~

p 140 -

v o REACTOR SHUT 00WN o

a O

tc.120 E

o

>~

Q.

$ 100 l i 80 O 1000 2000 3000 1 Time (s)

Figure 19. Isotropic boron - pressure suppression pool temperature.

during level control. When the downcomer water level throttled), the 17 % core power resulted in a 4F"/ min.

is raised back to normal level, the stratified boron heatup rate.

would be transported into the core. To bound the level control is effective in reducing core power and .

system response, a simulation was performed with hence PSP heatup. However, without negative reac- I minimum SLCS effectiveness. This was modeled by tivity insertion from control rods or boron, the PSP I assuming that the boron would be entirely stratified continues to heatup. At 875 s, the HCTL (Figure 7) in the lower plenum until 1585 s. At that time the is reached with the PSP at 165*F. The operator, 265 lb of boron solution in the lower plenum would monitoring this temperature, is instructed by the EPGs be tr:nsported into the core when the water level was to maintain RPV pressure below the limit. It is assumed raised in the downcomer. Thus, no effect of boron on that the operator would open the six relief valves core power was seen before 1585 s. The remainder of dedicated to the ADS and blowdown the RPV. The the simulation was modeled identical to the previous resulting RPV pressure response is shown in Fig-ase (Section 5.2.1). ure 23. Opening the ADS valves rapidly depressurized The simulation proceeds automatically until 120 s, the RPV to below 400 psia. While depressurizing the when HPCI and RCIC injections are terminated. RPV, Caution #14 of the EPGs " warns the operator At the same time the SLCS is initiated and not to reduce RPV pressure below the isolation set-begins injecting 50 gpm into the lower plenum. The points of steam-driven makeup systems unless motor-boron worth, however, is assumed to be zero until driven pumps are available." Since HPCI would isolate 1585 s. The RpV water level (Figure 20) drops rapidly when the RPV reached 100 psig, the operator could until throttled HPCI flow recovers and maintains it well be expected to terminate the depressurization slightly above TAF. An injection flow rate of before then. For this simulation, the operator was 3800 gpm is required to maintain level at TAF. This assumed to hold RPV pressure between 300 and a results in a core power of ~17%, as shown in 400 psia. That range keeps the pressure below the l Figure 21. The effect of reduced core power on PSP HCTL and also ensures HPCI operation. It should be heatup is shown in Figure 22. With HPCI and RCIC noted that there is some uncertainty in modeling the at full flow, the core power of ~30% resulted in a manual depressurization. ORNL has discussed in detail heatup rate of 6F"/ min. With level control (HPCI the problems involved with staying below the HCTL 27

l 600 , ,

f k

Q I500 o

$ HPCI. RCIC OFT J /

u o

E  :

o E 400 - HPCI FULL FLOW .-

O

! [!

. TAF

\HPCITHROTTLED 300 O 1000 2000 3000 Time (s) l Figure 20. Stratified tmron - downcomer water level.

100 , , I l

i 2

PEAK POWER:

2m AT 27 s 75 - _ j

'e i 3

o O.

o L I o 50 - - i O

o N

4 o

E 25 LEVEL RAISED _ j Z

g LEVEL CONTROL

/

Dh 0 1000 2000 3000 Time (s)  ;

Figure 21. Stratified boron - normalized core power.

28 i l

200 i  ; --- -

REACTOR SHUTDOWN

^ i b i o 150 u 4'FImin. i 3

0 L

E E

v H -

10 0 6* F/ min.

Q.

50 O 1000 2000 3000 Time (s)

Figure 22. Stratified boron - pressure suppression pool temperature.

1500 i i l

4 l

PRESSURE CONTROL o

"E Q. MANUAL DEPRESSURIZATION 1000 -

/

U 5

8 o

6 l

0 l u

3 500 -

I 8

v CE 0

O 1000 2000 3000 Time (s)

Figure 23. Stratified toron - RPV pressure.

29

(see Reference 3). The approach taken here is This simulation was modeled identically to that in reasonable, but certainly not the only way to model Section 5.2.2, except no depressurization was a manual depressurization. modeled. At 120 s, the operator initiated SLCS and During the depressurization, the PSP experiences in- began the level control procedure. The boron was creased heatup (Figure 22). Iloweser, at 1050 s the assumed to be totally stratified in the lower plenum heatup rate is lower than before the depressurization. until 1585 s. The llCTL was reached at 875 s, how-This is because significant void is intnxluced in the core ever, it was assumed that the operator did not during the depressurization, reducing core power. depressurize the RPV.

The downcomer water level is raised at 1585 s when Core power is shown in Figure 24. Following level 265 lb of boron have been injected. This rapidly shuts control, the power steadies out at around 17%. At down the reactor as the boron enters the core. The PSP 1585 s,265 lb of boron is placed in the lower plenum, pool temperature has reached 195 F at the time of and the downcomer water level is raised. Baron is shutdown. rapidly transported into the core, effecting a hot l shutdown. ,

5.2.3 Minimum SLCS Effectiveness-Without During level control, the PSP pool heats up (Fig-Depressurization. As stated previously, the objec- ure 25)at ~4F'/ min. Again the IICTL(165 F in pool) tive of Contingency #7 is to minimize PSP heatup dur- is reached at 875 s, however, no depressurization was ing the time of boron injection, thus avoiding the need modeled. The pool reaches 200 F at 1380 s and satura-  !

for emergency RPV depressurization. Reactor instabil- tion temperature (218 F) at 1585 s. At 1600 s, the ities are possible during a blowdown if the reactor has reactor is shutdown and pnxlucing steam at decay not been shutdawn. The EPGs offer some flexibility, power levels. At this time, torus cooling could halt the and state that "the operator need not resort to emer- PSP heatup and gradually cool it down. Assuming gency depressurization immediately upon reaching the complete condensation in the pool, the drywell llCTL." A simulation was performed to assess the ef- pressurized slightly but ilDWP was not reached.

feet of not depressurizing the RPV even though the For this simulation, it was assumed that ilPCI was llCTL was reached. available throughout the transient. At 1220 s, the PSP 100 i i R

v PEAK POWER:

2m AT 23 s 5 75 -

t 1 o '

O.

L O

8 50 -

o CE

'o o

.d tp LEVEL RAISED o 25 I. LEVEL CONTROL E

o Z

b 0

O 1000 2000 3000 Time (s)

Figure 24. Normalized core power without depressurization.

I 30 L______________

250 i i

~ ~ ~ ~ " - - - "

p200 - , .. - -- g---

v ~ ,,,,, " LWI TH DEPRESSURIZATION O ,- REACTOR SHUTDOWN y .-

a O -

g 150 -

a E

o F-Q -

V)

c. 100 l

50 O 1000 2000 3000 Time (s)

Figure 25. Pressure suppression pool temperature without depressurization.

pool temperature reached 190'F. Because water from to be totally stratified in the lower plenum until 265 lb the pool is used to cool lube oil for the HPCI turbine had been injected. Level control was modeled to begin at 120 s. As before, a core power of N17% was bearings, extended operation above that temperature cannot be ensured. If HPCI should fail before the reac- predicted as a result of level control. At 980 s, RPV tor is shutdown, it is unlikely that the remaining high water level was raised when 265 lb of boron solution pressure systems would have sufficient capacity to raise had been injected into the RPV Figure 26 shows the downcomer water level to transport boron into the effect on PSP pool temperature of increased SLCS core. Should this be the case, the operator would be capacity. With 86 gpm and minimum effectiveness, it required to blowdown the RPV below the low pressure is predicted that the reactor would be shutdown in ECC systems injection pressures. This blowdown 1000 s. At that time the pool would have reached would occur at a time when there was no condensing 173*F, a slight breach of the HCTL. With 50 ppm, shutdown time would be 1600 s with the pool at 218*F.

capacity left in the PSP.

No blowdown was modeled for these two cases. With 5.2.4 EPG Nominal-86 gpm SLCS. By bound- 50 gpm and a manual depressurization, shutdown ing the effectiveness of the SLCS, it has been found occurred at 1600 s, with the PSP at 195'F. The in-that PSP pool temperature would range from 140*F creased SLCS capacity thus resulted in a quicker shut-to 218 F at hot shutdown. This range is determined down and a lower PSP temperature.

by analyzing first, maximum boron effectiveness, and then minimum boron effectiveness. The SLCS capacity 5.3 Effectiveness of EPGs used in these analyses was 50 ppm. The NRC's Final Rule8 for reducing risk from ATWS included the re- The purpose of this section is to consolidate results quirement that BWRs have a SLCS with a minimum from the simulations presented in this report, and capacity of 86 gpm of 13% by weight of sodium pen- illustrate the effectiveness of operator actions in taborate solution. A simulation was performed to assess mitigating a MSIV closure ATWS. Because operator the effect of this increased flow rate. actions were modeled from directions given in pro-For this simulation, it was assumed that the SLCS posed EPGs, this becomes a partial evaluation of the was initiated at 120 s. The boron solution was modeled EPGs.

31

3 18 0 , ,

16 0 - REACTOR SHUTOOWN _

m k'~.-

0 Ba 140 - -

O t

o Q.

! 12 0 - -

Q.

(A Q.

100 -

80 O 1000 2000 3000 Time (s)

Figure 26. Pressure suppression pool temperature with 86 gpm stratified boron.

Because of the uncertainty involved in modeling the hibited. The PSP pool temperature reached 195'F at individual rod reactivity, no credit was given for con- the time of shutdown. If the RPV was not depres-trol rod insertion in these analyses. Section 7.3 dis- surized, the pool temperature reached 218*F at the time cusses an approach to allow the modeling of control of shutdown. Increasing SLCS flow from 50 to 86 gpm rod insertion. Because of the potential to reduce power decreased the pool temperature at shutdown from 195 and shutdown the reactor, the EPGs are entirely correct to 173*F, and no depressurization was required.

in directing the operator to attempt manual control rod level control, as advocated by the EPGs, is predicted insertion. The effectiveness of that action, however, to be effective in reducing reactor power from 30 to is not currently quantified. 17%. This in turn reduces the PSP heatup rate when The EPGs instruct the operator to initiate the SLCS the MSIV is closed. Reactor power is calculated to be during an ATWS when the PSP temperature reaches ~17% with level control and the RPV at 1000 psia.

110 F. This is indeed a prudent action, considering the At lower pressures, level control is more effective.

short time frame required for the ATWS progression. Emergency depressurization, as required by the The present analysis provides data with which to EPGs, may be necessary if the HCTL is violated. This l evaluate boron effectiveness by considering bounding action has been found to be effective in reducing RPV cases. If the operator begins level control procedures pressure. In addition, depressurization aids in reducing at 120 s, and maintains RPV pressure bekw the HCTL, reactor power and PSP heatup rate. It should be noted the ibliowing results are predicted. With boron injec- that operator training may be required to prevent tion at 120 s, assuming isotropic transport (maximum flooding of the RPV by low pressure injection systems effectiveness), the PSP temperature reached 140*F following a depressurization. The adverse effects of when the reactor was shutdown. For that case, no RPV low pressure power oscillations or unthrottled low depressurization was required because the HCTL was pressure ECCS injections could be significant con-not reached. The other bounding case initiated boron tributors to fuel damage.

injection at 120 s, but assumed that it totally stratified Comparisons of predicted PSP temperatures for the in the kwer plenum until 265 lb was injected (minimum five simulations presented in this report are summar-effectiwness). For that case, RPV depressurization was ized in Table 4. Figure 27 graphically presents the PSP required and kiw pressure system injections were in- temperature histories for each simulation.

32

r Table 4. Summary of predicted PSP temperatures Time to Time to Saturation PSP 200'F Temperature Temperature in PSP in PSP at Shutdown Simulation Description (s) (s) ( F) 1 Plant automatic 924 1220 NA 1 NR NR 140 l 2 EPG nominal-maximum SLCS effectiveness 3

3 EPG nominal-minimum NR NR 195 SLCS effectiveness 4 Minimum SLCS 1380 1585 218 effectiveness without depressurization NR NR 173 5 EPG nominal-increased SLCS capacity and minimum effectiveness NA: Not applicable.

NR: Not reached.

300 i i i x CASE 1

+ CASE 2 o CASE 3 l o CASE 4 E

v a CASE 5 ym 200 -

- i e '

u

] ,6' a a a a a a a a a 0 I L

e -

Q.

E e

100 Q. I w

Q.

0 O 500 1000 1500 2000 Time (s)

Figure 27. PSP temperature comparisons, 33

6. FUEL DAMAGE ANALYSIS During the plant automatic simulation described in dix D contains a description of FRAP-T6 and details Section 5.1, the potential for fuel damage existed in of its application to this transient.

two different portions of the transient. Section 6.1 The modeling approach taken here was to model a describes the FRAP-T6' analysis of the power excur- single fuel rod and calculate its failure probability. Data sions caused by low pressure ikxxiing of the RPV. Sec- from the first power spike (1275-1400 s) were used l tion 6.2 describes the SCDAPW analysis of the high as a basis for the analysis. The first power spike was pressure boiloff occurring after ADS isolation. Appen- calculated by RELAP5, and the actual input used in dix C contains details of the SCDAP code, and FRAP-T6 is shown in Figure 28.

Appendix D presents details of the FRAP-T6 code. FRAP-T6 predicted that 10% of the fuel rods would fait during the first power excursioni Failure is defined 6.1 Power Excursions: FRAP-T6 here to be a breach of the fuel rod cladding that allows fissi n pr ducts t migrate int the coolant. The failure Res ds mechanism was cladding overstress, which is predicted Between 1275 and 2600 s in the plant automatic on a comparison of clad hoop stress to the standard simulation (Section 5.1), seven power excursions were deviation of the ultimate hoop stress." Failure was predicted. These power spikes (Figure 13) were caused calculated to occur at 1520 s. A small amount of when low pressure injection systems flooded the RPV hydrogen generation was predicted during the time that with highly subcooled water. To examine the possibil- cladding oxidation temperatures were reached. Based ity of fuel damage during these power spikes, the on the single fuel rod prediction, approximately FRAP-T6 computer code was used. FRAP-T6 is 0.003 lb of hydrogen would have been produced by designed to calculate the thermal and mechanical one 8 x 8 bundle, behavior of light water reactor fuel rods during opera- It should be noted that the prediction of fuel rod tional transients and hypothetical accidents. Appen- failure probability is dependent on the input conditions 15 0 , i ,

)

FRAP-T6 INPUT v

I

$ 10 0 -

i u

O 1 s

o 1 0

x m

E 50 - -

.=

0 E

c O

Z 0

1200 1250 1300 1250 1400 Time (s)

Figure 28. FRAP-T6 input power.

34

supplied by RELAP5. Since the predicted power oscil- A single, high-power fuel bundle was analyzed. The lated substantially during the excursions, it was aver- fuel rods were divided into ten axial nodes, each 1.25 ft aged every five time steps. The resulting FRAP-T6 in- in length. The temperature histories of the top six axial put powr (Figure 28) had a peak power of 114%. A nodes are shown in Figure 29 (node 10 is the upper-sensitivity was performed in which a 200% peak power most node). Each of the nodes exhibits temperature ex-spike was input. That resulted in a prediction that 70% cursions due to runaway oxidation. A peak clad of the fuel rods would fail. Although the integral of temperature of 4300*F was predicted. 'Ihe temperature power was not held crnstant, the 200% spike occurred excursions in nodes five through ten were terminated over a very short time frame and had little effect on by the liquefaction and flow of cladding. The sequen-the integral. Since the rods are near failure, any change tial temperature excursions starting at the top of the in power will cause a much higher failure probability, core are a result of the dropping level in the core dur-ing the boiloff. Only minor differences were seen in 6,2 High Pressure Boiloff: the radial temperature response across the bundle.

CRD now maimained emhng in the lowenhme nodes, SCDAP Results which prevented temperature excursions there, in the plant automatic transient, the ADS valves close Steam mass flow rates in the bundle are shown in on high containment pressure at 2470 s. This closure Figure 30 at nodes four and ten. The mass flow rate keeps the RPV pressure above the low pressure injec- varied between 0.007 and 0.020 lb/s in response to the tion systems shutoff heads. The only high pressure in- core inlet flow. The sharp downward spikes in steam jection system available at that time is the CRD system, flow correspond to the temperature excursions in which injects 112 ppm into the RPV. Even at very low Figure 29. as steam was consumed by runaway oxida-power levels, that flow is insufficient to maintain core tion. Steam starvation was not calculated to occur in cooling. At N4200 s, the fuel rods began to heat up the bundle. Cladding temperatures in node 4 exceeded as a result of the high pressure boiloff. At ~5100 s the oxidation threshold temperature of 1350'F, but did cladding temperatures were predicted to reach 1350 F, not reach the point of enhanced oxidation.

high enough to cause zircaloy oxidation of the clad- In the regions above the bundle, a significant amount ding and hydrogen production. Using extrapolated of heatup and subsequent meltdown was calculated.

boundary conditions from RELAP5, a SCDAP ealcula- Heated by core exit steam, the top guide, the core tion was initiated just before the onset of oxidation. shroud head, and standpipes were calculated to ex-The objective of the calculation was to determine the perience rapid oxidation and melting. The material in i extent of fuel damage, hydrogen production, and fis- the separator, dryer, and steam dome regions ex-sion product release during a high pressure tmiloff, perienced some oxidation, but no melting.

Details of the SCDAP code and its modeling of the SCDAP calculated that a total of 1.7 lb of hydrogen high pressure boiloff are presented in Appendix C. was produced during the transient by oxidation of the The SCDAP analysis began $100 s after transient fuel rod cladding, bundle canister, and material in the initiation, just prior to the onset of zircaloy oxidation upper vessel regions. Note that this value is for one of the cladding. RELAP5 results were used to provide 8 x 8 bundle only. A total of 32% of the Xc, I, Cs, the initial boundary conditions (until 5600 s). Because and Kr in the fuel was calculated to be released, as of the nearly steady state core inlet conditions, the a result of the dissolution of fuel by liquefied cladding.

l results were extrapolated to 11100 s. At that time the The value calculated for hydrogen production has rates of fuel damage and hydrogen production were some uncertainty associated with it because of certain nearly zero, and the calculation was terminated. modeling assumptions used in SCDAP. Because

( 1: was realized that the SCDAP core analysis could SCDAP performs a single bundle thermal-hydraulic affect the system response (such as hydrogen genera- analysis, no effect of bundle flow area reduction on tion increasing system pressure and changing core in- steam flow rate is considered. It did predict, however, let flow). In many cases, extrapolation of the required that the fuel bundle was blocked at 7155 s from I boundary conditions may not be valid and an iterative liquefied fuel and cladding which solidified near the approach between RELAP5 and SCDAP would be tap of the mixture level. A lower bound on hydrogen necessary. In this case, however, the core inlet flow production was estimated by assuming that no was small, and the enthalpy and pressure were chang- hydrogen production occurred after bhickage was I ing very little (pressure was held fairly constant due calculated. This reduced the amount of hydrogen to SRV actuation). Thus it was judged that the RELAP5 produced from 1.7 to 0.8 lb for a single bundle boundary conditions and the SCDAP calculated system analysis. SCDAP has no model to track the movement l conditions were sufficiently decoupled to provide a of liquefied material in the upper vessel regions. Thus, valid analysis. when this material begins to liquefy, the oxidation 35

5000 , ,

4000 - -

n a W n I' -

.g o 3000 - -

.~- )

m a

0 3 6.

I o a 2000 -

O AXlAL NODE 5 b ,-

! O AXlAL NODE 6 H ' a AXIAL NODE 7 y ,,

+ AXIAL NODE 8 1000 L [

X AX1AL NODE 9 -

o AX t AL NODE 10 0 ' '

510 0 7100 9100 1110 0 Time (s)

Figure 29. SCDAP fuel rod cladding temperature h: aries of the top six core axial nodes.

0.05 , ,

NODE 4

---

NODE 10 0.04 - -

^ '

N ', .

i:'-

f 0.03 -

y  ;  :: -

o I  ::  :: -

[ j  ::  ! :: .,

m m

0.02 -

I.i

!! 'I..

'l I o -

2  :

l,s:'  ::

I f 0.01 -

'l 1.! -

i 1

0.00 ,

510 0 7100 910 0 1110 0 l

Time (s)

Figure 30. SCDAP mass flow rate of steam in the bundle at nodes fou and ten.

1 36

7 I.

.?

i and heat transfer from the material is terminated. As rekication in the upper vessel region are required. Also, a result, the hydrogen production in the upper vessel an integrated analysis that accounts for bhickage and regions may be underpredicted, although the amount redistribution of coolant to other bundles would reduce cannot be quantified. the uncertainty in the hydrogen production in the core.

The numbers presented here should be taken as best Such an analysis capability is cu rently being developed estimate within the modeling limitations stated above. at the INEL, with the SCDAP/RELAP5/ TRAP-MELT To obtain more confidence in the results, models for computer code, ,

l 1 l

l l

l l

l l-(

l 1,

i 1

37 w

i I

l

7. ANALYSIS UNCERTAINTIES .

I The MSIV closure ATWS simulations analyzed in dimensional thermal-hydraulics formulation, the model j this report result in predicted plant conditions that ex- will accurately predict the mixing of boron for different i ceed the range of available data bases. Uncertainths flow conditions. If the boron is calculated to enter the I remain concerning plant systems reliability, effec- core symmetrically, a one-dimensional neutronics for-tiveness, and failure modes. Analytical models have mutation is then required to calculate core power l been applied in certain areas where data are not avail < levels. If entry is calculated to be asymmetric, then able for their proper checkout. This section outlines a three-dimensional neutronics formulation would be the significant uncertainties identified by this analysis, required.

and suggests methods for the reduction of the Ultimately, full scale plant data or properly scaled l uncertainties, experimental data are required for model verification. l In lieu of data, the General Electric model could be ]

7.1 SLCS Effectiveness i""Porated into the INEL version of TRAC-BDl/

MODI. The model could be exercised through the The SLCS is designed to inject sodium pentaborate range of conditions expected during a MSIV closure in the RPV for the purpose of achieving suberiticality ATWS. If boron is predicted to enter the core sym-in the event of control rod drive failure. The SLCS metrically, the one-dimensional neutronics formulation is a redundant system, and its reliability is not a major in TRAC-BD1/ MODI will accurately predict the concern. Rather, the question of SLCS effectiveness neutronics feedback. If the boron is predicted to enter centers around the mechanism of boron transport in the core asymmetrically, iterations between TRAC-the lower plenum, and the movemes of the boron solu- BDI/ MODI and a three-dimensional neutronics for-tion into the core. mulation would be required.

At BFNPI, injection of the SLCS is through a single sparcer h>cated in the lower plenum just below the core 7.2 Effectiveness of Level plate. This asynunetrical injection raises questions con-cerning the baron solution being transponed uniform-Control ly into the core. Because the boron solution has a During an ATWS, the reactor power is determined specific gravity of 1.1 and it is highly subcooled when by the energy and mass flow rate of the fluid injected it enters the RPV, the solution has the potential for set- into the RPV. If tSe iajected flow rate and energy are tling in the lower plenum particularly during low core known, the reactor power can be easily calculated by flow conditions. In addition, if power is not reduced the expression: Q = Wjp (hg- h)p), where Q is the early in a MSIV closure ATWS, HPCI may fail before reactor power, Wjp si the mass flow rate through the sufficient boron has been injected to shutdown the core. jet pump discharge, hgis the enthalpy of steam Dav-This condition becomes more probable if SLCS injec- ing the core, and hjp si the enthalpy of the jet pump tion is delayed. The ability of RCIC and CRD flows discharge fluid. The proposed EPGs (discussed in Sec- l alone to transport the boron solution into the core is tion 3 of this report) advocate level control for the I

unknown. purpose of reducing core power and the resulting The analyses presented in this report used a bound- steaming rate to the PSP. The reduction of level to TAF ing approach in modeling boron irjection. To deter- translates to reducing the core inlet flow, and conse-mine maximum effectiveness, the boron solution was quently core power. The uncertainties related to this calculated to move isotopically with the liquid in the maneuver, particularly the calculation of core power vessel. Ti:is provided results based on ideal conditions. are discussed below.

The other bounding solution was obtained by assum- After the recirculation pumps coast down, the RPV mg minimum effectiveness. This was done by calcu- reaches a stearly state natural rimtation mode of lating the boron worth to be zero until an amount . operation. Injected flow mh. ins a hydrostatic head sufficient for hot shutdown was injected. in the downcomer, wMeh is balanced by the pressure To better analyze the effectiveness of the SLCS, a loss in the core tiow path. Whea the dswncomer level mechanistic model is required which accounts for den- is lowacd to TAF, a riew momentum balance is sity differences between the boron solution and water. reached. The riowncomer hydrostatic head is balanced Such a model has been developed by the General Elec- by the combined hydrostatic, dynamic (flow), and ir-tric Company for use in the TRAC-BWR code.n recoverable (friction:) heads across the core. Since the flowever, coefficients needed for the model input are dynamic head is small, the hydrostatic head $n the proprietary. Presumably, when coupled with a three- downcomer is approdmately equal to the surr. of the 38

i

(,-

')

hydrostatic and irrecoverable heads taken across the Further complications arise when the changing core. thermal-hydraulic conditions feedback to the core-The downcomer hydrostatic head is easily computed. ~ neutronics. With saturated core inlet conditions, the However, sensitivities arise when trying to compute axial power profile shifts upward in the core compared the individual core pressure drops, even though the sum with subcooled inlet conditions. The core void profile -

of them is fixed. The hydrostatic head is linearly depen- also shifts as a result of the power profile shift, chang-dent on core void. An increase in core void will ing void reactivity. The end result is that as inlet sub-i decrease the hydrostatic head in the core. Because the cooling is lost, core power decreases. However, the overall pressure drop is fixed, this will tend to increase shifting axial power profile decreases core voids which the irrecoverable loss in the core. limits the power decrease.

Irrecoverable losses are dependent on both flow and Figure 31 illustrates the effects of some of these l

l void content. They are dependent on the flow, at least interactions. Plotted on the figure are predictions of quadratically, while they exhibit an . exponential the core thermal power (as a percent of rated) as a func-dependency on void content. Thus, an increase in core tion of downcomer level. The curve was generated with void increases flow losses not only by increasing the RELAP5 through a series of steady-state calculations, core flow, but also by exponentially increasing the and the points on the curve represent asymptotic steady-multipliers associated with wall friction and form state values. The model used for these calculations has losses. These losses are directly related to core power. been previously assessed with natural circulation data For example, an increase in power will increase the from the Full Integral Simulation Tr.st (FIST) facility, core void, which would tend to increase the core flow. An axial power profile typical of a saturated core inlet j However, a nonlinear feedback acts to restrain this in- condition (see Figure A-4) was used.

crease, as irrecoverable losses increase with the void. As the downcomer level is lowered by reducing The algorithm (drift flux formulations, interfacial makeup flow, there is a corresponding reduction in shear, etc.) by which the core void is calculated is core power. This continues until a plateau is reached pivotal in the estimation of power during leve? control, where further lowering of the downcomer level results  !

l 25 . . .

l l

b 20 - -

ic -

" Knee" "Plateou" -

I k -

u . .

o o

o 15 -

The effect of lowering the power -

S . profile is to lower the void .

E profile, which " moves" this curve C lo the right  :

e . ".

8 . .

$ 10 -

. Top of the .

Top of the *I ~ Top of the -

letpumps CC{I T upper plenum .

' ' ' ' ' I ' ' ' ' '

5 26 28 30 32 34 36 Downcomer Level (ft) un_aus-oi Figure 31. The effect on reactor power of lowering downcomer liquid level in steady-state operations.

39

in only a small reduction in core power. The plateau yields the following result. With saturated core inlet exists until the downcomer level reaches slightly below conditions, the core power with level at TAF and the TAF, at which point a cusp (or knee) is seen in the RPV at 1000 psia will be approximately 17-20%. This curve. Further lowering of the downcomer level again value agrees closely with previously published Elec-3 results in a corresponding power reduction. tric Power Research Institute (EPRI) results and with Both the plateau and knee in reactor power exist RAMONA-3B calculations" done at BNL as part of because of a delicate hydrodynamic balance occurring the SAS A program. Although these results have con-within the core, upper plenum, standpipes, and verged to a range of severalpreent, questions remain separator of the reactor. The plateau indicates that a on their accuracy. A paperl discussing the uncertain-stagnation of the fluid conditions within the core has ties in these power predictions claims that the actual occurred. In this stagnation condition, the steam exit- power has a 30-35% probability of being more than ing the core through the upper plenum and standpipes 5% (of rated power) higher than the predictions. In has insufficient velocity to entrain and transport liquid addition, the paper concludes that an uncertainty of out of the core, through the separators, and reflux it ~5% of rated power is not achievable with the cur-into the downcomer. That is, the liquid and steam flow- rent state-of-the-art.

ing from the core decouple, and the liquid fails back into the core. This differs from the state occurring dur' ing normal operation, and during full flow HPCI opera-7.3 Manual Rod insertion tion, wherein the phases are strongly coupled. As explained in Section 3, the proposed EpGs direct During normal operation, wherein the phases are the operator to attempt manual insertion of control nxis strongly coupled, the moderating coolant sweeps during an ATWS. The effect of this insertion on re-through the core at fairly high velocities. The shape ducing reactor power was not modeled in the pesent j of the void profile, which strongly influences both the analysis, because of the uncertainty involved. The in-core shape and total magnitude of the reactor power, worth of an individual control rod could v . y con-is strongly influenced by the coolant velocity. As this siderably depending on its position in the core. Also, vehicity is decreased, the liquid tends to settle towards the worth is dependent on how many other rods have the bottom of the core, thus modifying the in-core void been inserted and their location. In other words, it is profile and content, and the in-core power shape and unlikely that an average rod worth (negative reactiv-total reactor power. ity) could be used to model individual control rod On the reactor power plateau, changes in the total insentons.

void content of the core are compensated for by a re- To determine the effect on power, a detailed three-arrangement in the in-core shape of the void profile. dimensional neutronics calculation would be required.

At the knee, this compensation can no longer take Using a realistic nxl insertion pattern, the effect could place. Hence, the core void content increases, and the be calculated for the expected range of core inlet power again decreases rapidly with level. conditions. When level control is modeled in con-The physics dominating the shape of the dependency junction with manual rod insertion, the power reduc-of reactor power on downcomer level during ATWS tion from ux! insertion would have to feedback to the raitigation are thermal-hydraulic. However, the shape injection required to maintain level. This could be of the power profile is an important parameter actual- accomplished either by an integrated thermal ly projecting what power will occur with level at TAF. hydraulic /neutronic analysis or by iterations betwcen Varying the in-core shape of the power will shift the three-dimensional neutronics and thermal-hydraulics level control curve shown in Figure 31 across the calculations.

figure. Lowcring the power profile (typical of sub- The end produc of such an analysis would be a quan-cooled core inlet conditions) shifts the curve to the tification of the effect of individual rod insertions on right. Thus, a total power of ~8% can be realized if core power reduction. In addition, a valuable result the power profile is sufficiently bottom humped. Using from the analysis would be a table of control rod reac-a profile typical of saturated core inlet conditions tivity versus time. The table could be used in systems results in the curve in Figure 31, which indicates a total codes to model manual nxi insertion (through tabular power of ~18% with level at TAF. input of control reactivity) during ATWS simulations.

Other factors which are necessary for the prediction Because of the potential of manual rod insertion to j of power reduction from level control include core mitigate an ATWS, it is recommended that an effort  !

bypass flow characteristics, grid spacer loss cmffi- be undertaken to quantify its effectiveness. It is im-cients, reactivity coefficients, steam condensation on portant that individual control rod effects be separated subcooled makeup water, and xenon poisoning. Using from other power reducing mechanisms such as level best estimate values in a prototypically assessed model control, pressure control, and SLCS injection.

1 40

7.4 Containtnent Related progresses, then the assumption of a well mixed PSP is a g od n . In such a case, complete steam conden-Uncertainties sation could be expected until the pool bulk temperature The large amount of energy delivered to the PSP dur- was nearly saturated. The EPG action of activating the ing a MSIV closure ATWS causes it to heat up rapid- torus cooling mode of RHR when the PSP temperature ly. The PSP heatupjeopardizes HPCI operation within exceeds 95'F is conclusively appropriate, as it would 13 to 14 min, and it will eventually cause HPCI failure extend the time the operators would have before it if the plant is not shutdown. With HPCI unavailable, would be necessary to prevent an automatic ADS. The l vessel water level will depress below the LLLWL set- ability of the operator to activate torus cooling is ques- l point. Below LLLWL, in the presence of a HDWP tionable, however, as explained in Section 3, Were the signal, the ADS will automatically actuate. pool not well mixed, extrapolated commercial plant test There are some important concerns with respect to results* indicate that steam breakthrough could be ex- i the automatic ADS. The timing of the HDWP signal, pected at a pool bulk temperature of about 180*F. In and thus the automatic ADS timing, is uncertain dur- either case. mechanistic models to calculate the amount 1 I

ing a postulated ATWS. An unintentional ADS blow- of steam breakthrough to the wetwell atmosphere are down is undesirable, becaur it could lead to very large not available in todays containment codes.

core power cycles induced by RPV flooding by the low An automatic ADS blowdown to the PSP would oc-pressure injection systems. Our current estimates range cur at a time when the pool subcooling is significantly from 16 to 35 min for receipt of the HDWP signal, diminished. Predicted containment service during the depending on how well the PSP is mixed and on how unmitigated ATWS is outside the bounds of known effective the T-quenchers are in promoting steam con- stable condensation for the T-quencher devices." The densation. A large part of the uncertainty in time before structural loading of the PSP and drywell during the steam breakthrough in the PSP is dependent on what ADS blowdown was not examined in this work, the operators are able to do. If, early on, the operators However, the hydrodynamic loading during a are able to place the RHR system in the torus cooling blowdown with low subcooling in the PSP may be of mode and maintain this alignment as the transient concern, i

I

)

l 41

8. CONCLUSIONS AND RECOMMENDATIONS rate from 6 to 4*F/ min. Although several anal-A deterministic study was performed that analyzed yses have converged on the predicted power postulated MSIV closure ATWS scenarios at BFNPl. level range of 17-20%, uncertainties remain and Conclusions, results, and recommendations resuhing the actual power could be higher.

from the study are summarized in this section.

6. The action of depressurizing the RPV to avoid

1. Mitigative operator actions are required to pre-violation of the HCTL should be evaluated fur-vent containment failure and severe fuel damage ther. RPV depressurization should result in a during a MSIV closure ATWS. During the plant power reduction during an ATWS, and it may automatic simulation, primary containment fail-also preclude the need to blowdown the RPV ure was predicted from static overpressurization when the PSP has no condensing cr.pacity, 45 min after transient initiation. In less than four However, the potential for low pressure power hours, over half of the core was predicted to oscillations and unthrottled low pressure ECCS have liquefied and relocated. Significant hydro-injections could cause fuel damage.

gen production and fission product release from the fuel was predicted to occur. 7. Even with minimum SLCS effectiveness, it is predicted that PSP temperature would be less

2. It is predicted from these simulations that if all than 200 F at the time of shutdown. Using of the actions proposed by the EPGS are com-50 gpm of 13 % sodium pentaborate solution in pleted successfully, then a MSIV closure ATWS conjunction with level and pressure control would be brought under control. Some of the results in a PSP temperature of 195*F at shut-actions, however, may be very difficult to down. Assuming maximum SLCS effectiveness accomplish. results in a PSP temperature of 140 F.

3. Operator training should be an integral part of

8. Increasing SLCS capacity to 86 gpm reduces the the implementation of the EPGs. These sim-predicted PSP temperature at shutdown from ulations indicate that the operator must act 218 to 173*F for minimum effectiveness properly and promptly to reduce the risk from assumptions and no RPV depressurization. The a MSIV closure ATWS. The process of follow-reduction in temperature is significant in that ing the EPGs through their various parts could complete condensation in the PSP should ensure be very time consuming unless the operators that no steam breakthrough would occur, thus had a thorough understanding of them preventing drywell pressurization.

beforehand.

9. The uncertainty of assuming uniform control rod

4. More guidance should be given in the EPGs worth is currently too large to allow reliable relative to the injection of low pressure systems modeling. It is recommended that the effect of to the RPV. The operator is told to " slowly individual rod insertion be quantified in terms inject" when the RPV is depressurized. More of negative reactivity versus time.

definitive instruction is required, as are mea-sures to prevent automatic irjections. 10. Use of the torus cooling mode of the RHR system is not assured during a level control tran-

5. Level control as advocated by Contingency #7 sient. Once RPV level is raised back up, torus is an effective although limited means of re-cooling should be readily accomplished. The ducing reactor power. Lowering RPV water level to TAF at 1000 psia was calculated to importance of torus cooling lies not only in reducing PSP heatup rate but also helping en-reduce reactor power from 30% to approximate-sure that the torus remains wc!! mixed.

ly 17% of rated. This reduces the PSP heatup

)

12

9. REFERENCES

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

2. L. B. Claassen and E. C. Eckert, Studies ofBKR DesignsJbr Mitigation ofAnticipated Transients Hithout Scram, NEDO-20626,74 Ned 59, GE, October 1974.

3. R. M. Harrington and S. A. Hodge, ATHSat Browns Ferry Unit One, NUREG/CR 3470, ORNL/TM-8902, July 1984.

4. S. 2. Bruske and R. E. Wright, Severe Accident Sequence Analysis (SASA) Program Sequence Event Tree:

Boiling n'ater Reactor Anticipated Transient H1thout Scram, NUREG/CR-3596, EGG-2288, April 1984. ,

1

5. General Electric Company Prepublication Draft, Emergency Procedure Guidelines, BHR 1 through 6, Revi-sion 3, December 1982.

]

6. V. H. Ransom et al., REL4PS/ MOD 1.S: Models, Developmental Assessment, and User Information, (Modified for MODI.6), EGG-NSDM-6035, October 1982. ,

1. D. W. Hargroves and L. J. Metcalfe, CONTEMPT-LT/028-A Computer Programfor Predicting Contain-ment Pressure-Temperature Response to a Inss-of-Coolant Accident, NUREGICR-0255, TREE-1279, EG&G Idaho, Inc., March 1979.

l l 8. Nuclear Regulatory Commission, Reduction ofRiskfrom Anticipated Transients Hithout Scram (AIWS) Events j for Ught-H'ater-Cooled Nuclear Power Plants, Final Rule, July 1984. l l

l

9. L. J. SiefLen et al., FRAP-T6: A Computer Code for the Transient Analysis of Oxide Fuel Rods, NUREG/CR-2148, EGG-2104, May 1981.

10. G. A. Berna et al., SCDAP/MODl/VO: A Computer Codefor the Analysis ofLHR Vessel Behavior During I

Severe Accident Transients, IS-SAAM-84-002, EG&G Idaho, Inc., January 1984.

1

11. S. O. Peck, FRAII 6: A Fuel Rod Failure Subcode, WR-CD-025, EG&G Idaho, Inc., September 1980.

l l

12. D. D. Taylor et al., TRAC-BD1/ MOD 1: An Advanced Best Estimate Computer Programfor Boiling H'ater l l Reactor Transient Analysis, Volume 1: Model Description, NUREGICR-3633, EGG-2294, April 1984. i

13. B. Chenal et al., Reducing BHR Power by n'ater hvel Control During an ATWS, NSAC-70, August 1984.

I4. P. Saha et al., RAMONA-3B Calculationsfor Browns Ferry ATH'S Study, NUREG/CR-4739, BNL-52021, February 1987.

15. D. J. Diamond, " Uncertainty in BWR Power During ATWS Events," Proceedings ofthe Topical Meeting on Reactor Physics and Safety, Saratoga Springs, New York, September 17-19, 1986.

16. B. J. Patterson, Mark l Containment Program Monticello T-Quencher Thermal Mixing Test Final Report, NEDO-24542, August 1979.

17. Supression Pool Temperature Umits for BHR Containments, NUREG-0783, November 1981.

43

k

\

APPENDIX A RELAP5/ MOD 1.6 MODEL OF BFNP1 REACTOR PRESSURE VESSEL i

l l

I

( .

I l

l A-1 i

i APPENDlX A RELAP5/ MOD 1.6 MODEL OF BFNP1 REACTOR PRESSURE VESSEL l

This appendix to the report describes the RELAP5/ Table A-1. Plant modeling parameters MOD 1.6 compu'cr model of the reactor pressure vessel (RPV) of Browns Ferry Nuclear Power Plant Unit i Item Parameter (BFNPI), which was used in this analysis, and its qualification against plant transient data. Plant - Browns Ferry The BFNP1 model described here was originally - Nuclear Plant transmitted ^ 3 to the INEL by the Tennessee Valley One Unit Authority (TVA) in the form of a RETRAN-04 input deck. Table A-1 contains plant modeling parameter Plant type BWR/4 data. Converted to MODI.0 of RELAP5, the model 251 in, Belt line diameter was classified as interim and was used in a number of other analyses.A 2 m A-5 performed at the INEL. Number of bundles 7M Subsequently, the geometry of the model was quality P8 x 8R Type of bundles assured by tracing geometric data to plant drawings, and by comparing model performance to plant data, Warranted or rated core power 3293 M W so that the interim label was dropped. The quality assured BFNPI input deck described here was designed to run on RELAP5/ MODI.6. The steady state for the model described here was run un cycle 018 of RELAP5/ MODI.6. The recirculation loop has two associated special The model consists of the reactor pressure vessel of process models; the recirculation pumps and the BFNPI and related piping, such as the feedwater and jet pumps. Jet pump performance characteristics main steamlines, as shown in Figure A-1, and is con- were set according to available data as shown in sidered representative of the plant as reloaded through Figure A-2.A*^

cycle six. This model description segregates RPV regions according to their function. The main steamline The recirculation loop also includes the low is one such region. There are six such regions in the Pressure coolant injection (LPCI) point. Core in-model: let flow measurements are performed at the jet

1. Downcomer consisting of the annular region ex-tending downwards from the exterior of the 3. Lower plenum consisting of five components and dryers, dryer shroud, standpipes, upper plenum cells. It communicates thermally with the en-and core shroud, but excluding the jet pumps and vironment. The lower plenum is penetrated by .  ;

their discharge. The downcomer receives feed- both the control rod drive (CRD) tubes and CRD water and separator reflux, and provides makeup efflux and by the standby liquid control system .

to the recirculation loop. The downcomer com- (SLCS) nozzle through which the lower plenum municates thermally with the environment. The pressure measurements are performed. The lower downcomer consists of five components, num- plenum is numbered 290 through 294, and 350.

bered 415" through 475, comprising ten control The pressure tap is taken in volume 294.

volumes. RPV level measurements are performed in the downcomer. RPV steam dome pressure 4. Nuclear core includes the core inlet piece, chan-measurements are performed in the uppermost nels and fuel rods, bypass and reflector regions downcomer cell, inside the core shroud, and upper plenum. The core is numbered 300 through 380. The core

2. Recirculation loop combining the functional models P8 x 8R bundles of 12.5 ft heated aspects of both plant loops. The loop consists of length. The core model was developed to meet cight components, numbered 200 through 270 the following objectives:

and is comprised of 17 cells.

• Accurately predict hydraulic response of the

a. Referring to compment serial code CCC on RELAP5 cards core to various external perturbations, such cec (xxxxx). as a flow coastdown A-3

.\

- g , i O

c 5 t

O g [I

- 1W  ;

n 1 4> i,

-Rim n!

8 E 3l 5 Dl LJ Er o n s 5

5 SIP  %

• Sft -en $$t 4 g

- - - - - 2 g

e- 4

{ e--- 5 e- --- Q

$g

?

f R -R

=

- " ~ *~-

g j

- - 4 l

( a

• Sft M 999 +  % g W

> Git

~

k

~

__ -T 09E Di R x

~~

b -J oct c s -

_L. T

! a y ca --

@ rp  ;

$ p*

~ ~

a esa gg y, g g c

r $ "-

.c -

8e - a DSE sE E-S W E

@s 1-A-4

0.30 i i i i i y i j O DATA

-- RELAPS 0.25 -

o i E 0.20 - O -

I x

o )

L g 0.15 -

O e

U O 1 U

l 0.10 - 1 z o 0.05 -

0.00 O.50 0.75 1 1.25 1.50 1.75 2 2.25 2.50 f M (Flow) Ratio Figure A-2. Jet pump performance curve.

• Accurately predict the thermal response of ingA'8 is primarily 8 x 8 fuel Nearly all of the the fuel, especially in severe accident core has been reloaded. Only a few assemblics i l scenarios under boil-off conditions, and of the original loading remain in the cycle 4 core.

e re s ug to W repmsema@e of k

• Accurately predict the response of the core loadings of nearly all BWR/4 because of this age I reactor kinetics to various extemal perturba- j factor. Lattice modeling parameters are shown tions, and the resultant feedback to the in Table M I thermas hydrauhes. (

Data pertaining to the reactor kinetics are con-These technical objectives address SASA pro- tained in Table A-3. The moderator void reac- i gram needs. To meet the objectives, the core was tivity, fuel temperature reactivity, moderator divided both radially (across the core) and axial- temperature reactivity, boron worth, delayed ly (vertically). The resultant nodalization, as neutron fraction, and neutron generation time  ;

shown in Figure A-3, consists of three axially were derived from data presented in Refer-parallel flow channels. Two of these flow chan- ences A-9 through A-17.

nels represent the interior of fuel assemblies' The fuel loading represents EOC6 at BFNPl.

w hile the third models the unheated core bypass . .

Associa with this load.mg is a core axial power flow. Axially, the heated length of the fuel Profile. The axial power profile under assembly channels were divided into 1/ control n rmal power peration at EOC6 is bottom volumes of equal length. This level of detail was humped because the control rods are fully required axially to accurately model the non-withdrawn. Figure A-4 shows this effect in the .

linearity associated with the core hydraulics and P"**' P' ' The axial peakmg factor is 1.23.

reactor kinetics. ]

The power deh.vered to the fourth axial cell from  !

The core model was designed to model month 11 the bottom of the core receives 123% of the of cycle 4 at BFNPl. This time of cycle is very average cell pown delivery. Associated with the near the end of cycle (EOC), when the core is power profile is an importance profile, which was considered to be most reactive. The fuel load- generated by renormalizing the square of the A-5

.= D

VOLUME 370

-+1 I I I-*

HS 320 01 - AVERAGE POWER ASSEMBLIES

/////////////,///, ////////////// g g r '

-+ 5 - VOLUME 320 - * -

E -+

2 m F I 1 1 g

HS 360 01 - HIGH POWER ASSEMBLIES l///////////////////////////////////////E

-+ VOLUME 360 -+

/

Y/////////////////////////////////////////////////////////A i HS 360 02 HIGHEST POWERED ASSEMBLIES Figure A-3. Core nodalization.

Table A-2. Core lattice modeling parameters Item Parameter Fuel assembly type P8 x 8R Number of fuel assemblics 764 Total number of fuel rods 47,368 Active fuel length 150 in.

Total active rod heat transfer area 74,870.5 (ft 2) i Direct moderator heating 3 9c' A-6 L_._._______.________._.._.__ -

r~

Table A-3. Reactor kinetics data item Data Delayed neutron fraction 0.0054 Delayed fraction to generation time /A 123.7 s~l Void reactivity function P, = (-0.0979 + 0.01 Aa) Aa {

where a is the void fraction ,

1 Doppler reactivity function P d = -1.55 x 10-3 (% - %)

where Tp is fuel temperature (K) 1 Moderator temperature function Pm = -1.38 X 10" ATm where Tmis moderator temperature (K) {

Boron reactivity P 3 = (-3.75 x 10~3 - 5.96 x 10-4 Aa) RB where Rg is the mass of dissolved boron (lb)

Delayed Neutron Constants 1 2 3 4 5 6 Relative yield pi/p 0.0361 0.2343 0.1974 0.3734 0.1253 0.0335 Decay constant A (1/s) 0.0128 0.0314 0.1242 0.3213 1.3512 3.5802

a. Data obtained primarily from References A-11 and A-14.

1 I

)

1.4 , i , , , , , , , ,

^

1.2 -

c .

e 1 - _

s i o

l

$ 0.8 - _

= '

2 ,

e 2 0.6 - _

0.4 -

a SUBC00 LED $

o SATURATED

' ' ' ' ' ' ' ' ' i i 0.2 0 1 2 3 4 5 6 7 8 9 10 11 12 l

Core axial node i Figure A-4. Core axial power profiles.

A-7

power profile. The reactivity importance profile large hydraulic impedance in the hot channel.

hat, a relative peak in axial cell four of 140%. Assembly modeling data are presented in The importance profile is used to weigh the Table A-4.

relative worth of each axial location to the reac- . .

Corresponding to these core w.de i variations m tor kinetics.

power delivery (and exposure) are variations in Under normal power operation, the coolant sub- the thermal (and neutronic) properties of the fuel.

cooling at the core inlet is about 20 F. However, Fuel and gap thermal conductivities were under accident conditions, this is not always so, calculated for each the three prototypical fuel l Figure A-4 also shows a power profile, taken assemblies modeled. Average core void profiles from Reference A-19, which was used to simu- are shown in Figure A-5 for both power profiles late core conditions when the inlet of the core is under normal power operation.

near saturation. Such conditions exist when the

5. S,eparator region consisting of four components operator takes level control, and reduces the and cells (390,400,410, and 420). The separator i omer level to the top of the active fuel region includes the separator special process model which directs the liquid leaving the nuclear The core radial power profile is double humped, core to the downcomer (reflux) and the steam to the peak being located in the third annular ring the steam dome. The separation process was outwards from the core center. The fifth annular assumed to occur at the first pick off ring, ring, or ring 5, is the location of the secondary 6. Feedwater and main steamlines; model the feed-hump. According to data derived (see Refer-water and main steamlines. The feedwater line ence 8) from the core loading, ring 5 contains the i ts of two cells, while the steamline has ten.

eight highest powered fuel assemblics. The eight The steamline is numbered 500 through 525, and highest powered fuel assemblies are h)cated m the feedwater line is numbered 155 and 170.

an annular ring of relatively high power.

Associated with the feedwater line are the reac-It is felt that during a severe accident, wherein t r wata cleanup m, jection pomt and the high excessive fuel rod temperatures would be ex- .

pected to occur, these eight assemblics would be prenure c% ant mjdon (HM and readm c re is lati n c ! nt (RCIC) injection points.

most prone to fail. They were accordingly modeled with a single heat structure. This heat The steamline associates with the safety / relief structure was attached to the hydraulic channel valves (SRV), HPCI and RCIC turbines, and the representing ring 5. The thermal characteristics main turbine and turbine bypass. Turbine steam of the remaining fuel assemblies in ring 5 were can be interrupted by the main steam isolation separately modeled and also attached to the single valves (MSIV), or by the turbine stop/ control ring 5 hydraulic channel. The remainder of the valve. Safety / relief valve characteristics are heated core was modeled using a single channel presented in Table A-5.

and heat structure.

Tables A-6 and A-7 contain a summary of RPV geo-Using this approach, the eight hottest fuel assem- metric and elevational data as modeled, respectively.

l blies are associated with a relatively hot chan- Although much labor and care were extended in net. Thus, hot assembly thermal-hydraulics are development of this computer model of BFNPI, it was modeled while stabilizing the code with a relative realized that any plant transient behavior projected by l

l l

l l

Table A-4. Fuel assembly data I

1 Average Ring 5 Ring 5 Assembly Designator Core Average Hot Heat structure number 320-01 360-01 360-O2 j Number of fuel assemblies 652 104 8 Structure power factor 0.951 1.245 1.358 A-8 l

0.8 i i i i i i i i i i i l

0 SUBCOOLED PROFILE 0.7 - o SATURATED PROFILE ,  ; m

%)

g 0.6 -

z o

o 0.5 -

j

'- . i 0.4 -

0.3 3 -

E 0.2 -

0.1 -

l 0.0 6 i ' '

)

0 1 2 3 4 5 6 7 8 9 10 11 12 l Core oxial node I Figure A-5. Core axial void profiles at 20*F subcooling.

l Table A-5. Safety / relief valva characteristics Bank 1 2 3 4 Total Number of valves 4 4 3 2 13 Set pressure (psia)" 1120 1130 1140 1265 N.A.

2 Valve area (ft ) - - - -

1.1154

a. Within each bank, each valve's set pressure differs by 0.1 psi. All valves reclose 50 psi below their set pressure.

Valves relieve 3110 lb/s of ateam at normal reactor pressure.

1 A-9

Table A-6. RPV geometric data Volume Component Number Area, ft 2 Length, ft Volume, ft 3 AElevation, ft Lower Plenum 290 01 91.43 5.335 487.78 ~ 5.335 292 01 87.40 5.002 437.16 5.002 293 01 91.71 5.335 489.26 5.335 294 01 92.86 7.163 665.15 7.163 350 01 102.23 12.498 1277.70 12.498 3357.05 Core Region 300 01 80.43 0.526 42.30 0.526 320 01 80.43 2.083 167.55 2.083 320 02 80.43 2.083 167.55 2.083 320 03 80.43 4.167 335.I1 4.167 320 04 80.43 4.I67 335.I1 4.I67 320 05 80.43 1.413 113.64 1.413 370 01 65.68 4.672 3 % .86 4.672 370 02 65.68 4.056 266.40 4.056 370 03 65.68 5.711 375.09 5.711 380 01 264.67 2.363 625.41 2.363 390 01 304.61 1.519 462.86 1.519 3197.88 Separator Region 390 02 42.33 8.720 369,14 8.720 400 01 136.50 4.375 597.I8 4.375 410 01 351.34 12.667 4450.31 12.667 420 01 281.63 10.958 2395.78 10.958 7812.41 Downcomer 415 01 11.54 12.667 146.14 - 12.667 415 02 207,12 4.375 906.14 - 4.375 425 01 293.10 6.208 1819.66 - 6.208 435 01 293.11 3.084 903.86 - 3.084 435 02 87.86 6.916 607.67 - 6.916 435 03 97.04 1.956 189.82 -1.956 455 01 95.21 8.410 800.70 - 8.410 455 02 89.88 4.508 405.19 - 4.508 475 01 90.28 3.121 281.81 - 3.121 6060,99 A-10

i l

I 1

Table A-6. (continued)

Volume Component Number Area, ft 2 Length, ft Volume, ft 3 AElevation, ft l Recirculation Loopa I 200 01 7.80 20.998 163.70 - 12.458 210 01 7.80 5.784 45.14 ~ 5.784 220 01 7.80 23.136 180.59 - 21.333 220 02 7.80 13.566 105.89 2.534 230 01 7.80 19.561 152.69 2.000 240 01 7.80 31.248 243.91 19.463 250 01 7.80 2.750 21.46 2.750 250 02 16.38 13.729 224.86 1.667 250 03 7.53 16.773 126.30 13.250 250 04 5.47 4.017 21.99 0.833 250 05 5.47 11.879 65.05 11.879 250 06 5.47 2.703 14.80 1.883 260 01 7.30 4.417 32.24 - 4.417 270 01 7.30 4.417 32.24 - 4.417 270 02 12.53 2.402 30.11 - 2.402 1 270 03 22.28 2.402 53.53 - 2.402 {

270 04 34.87 2.402 83.76 - 2.402 1598.36 Feedwater Line l 155 01 5.07 129.250 655.61 - 18.170 170 01 4.24 99.48 423.870 45.370 1079.09 Main Steamline 500 01 12.20 23.867 291.15 - 22.516 500 02 12.20 23.867 291.15 - 22.516 507 01 6.09 28.338 172.85 0.00 510 01 12.20 30.462 371.63 - 4.394 510 02 12.20 18.878 230.31 - 15.500 510 03 12.20 28.719 350.36 0.00 520 01 10.14 33.940 344.31 0.00 520 02 10.14 33.940 344.31 0.00 520 03 10.14 37.960 385.09 35.7.10 525 01 10.14 13.712 139.20 12.000 2920.26

a. Referenced to left loop. This loop lumps both recirculation loops.

A-Il

Table A 7. RPV vessel elevations Table A-8. Steady-state conditions at 100% power a Elevation Model Designator Value Description (ft) (in.) Power, Energy, and Mass inside head lower plenum 0.00 0.00 Total reactor pown 3293.0 m Vessel ambient loss 0.134 MWt Recirculation suction 13.458 161.50 System total mass 687,656 lb Core support piece 17.500 210.00 Flows Feedwater flow 3692.1 lb/s Jet pump suction 26.377 316.52 Steamline flow 3710.0 lb/s Top of active fuel 30.526 366.31 Jet pump discharge 28,468 lb/s

1. "'" " " E"*E'

• Lo-lo-lo level 32.042 384.50 Core bypass 3383.0 lb/s Feedwater line 38.333 460.00 Bypass ratio 0.I19 : 1 penetration M-ratio 1.91 1 P m sures Lo-lo level 39.67 476.00 Lower plenum (292)b 1054.1 psia Normal water level 46.75 561.00 Lower plenum (294) 1050.0 psia HPCI trip level 48.50 582.00 Upper plenum (3801) 1024.6 psia Steam dome (41501) 1014.2 psia Main steamline 61.583 739.00 973.27 psia TCV (525) penetration Pressure Differences Recirculation pun'p head 233.68 psi Jet pump suction rise 22.2 psi the model could contain significant deviation from ac- Core inlet plate drop (294-370) 21.2 psi tual plant behavior. Without some means of verifying Core drop (294-380) 25.4 psi model behavior, verification of plant transient behavior Separator drop (380-41501) 10.2 psi could only be accompanied with extensive qualifica-N-ratio 0.11 1 tion because of model uncertainties. A need to verify Levels computer model behavior and to reduce and quantify Downcomer collapsed 46.75 ft I model uncertainties was identified. Bundle average void fraction 40.5 % i Towards this end, a series of benchmarks was per- Separator void fraction 81.0 %

formed. The benchmarks compared model behavior to Thermal Performance the best obtainable data. Plant operational transient Feedwater temperature 376.2"F comparisons were made to identify and reduce devia-tions from actual plant behavior.

Jet pump throat subc(x> ling 20.6'F )

Core inlet subcooling 22.3'F l The 100cf power steady state or rated conditions of Turbine inlet quality, X, 0.993 I l the nalel are shown in Table A-8. The steady states '

Peak centerline fuel temperatures presented are thought to be representative of plant con.

Average bundle 1556 F I ditions at 100% power. These steady-state conditions l Ring 5 bundle 1951*F  ;

were used as initial conditions for the two plant tran-sient data comparisons that are described here.

Ilot bundle 2181*F Two plant transient simulations were performed. The Recirculation Pumps first was used to compate model performance to plant Speed 1623.8 rpm l behavior during a pump coastdown. The transient ini- Head 707.1 ft I

tiator is the tripping of the recirculation pump motors Density 47.54 lb/ft 3 by interrupting the power to the motors from their associated motor-generator sets. Figure A-6 shows the a. Conditions obtained with RELAP5/ MODI.6/Cyc ensuing recirculation pump coastdown. As shown in 018, and with saturated power profile.

the figure, nulel behavior closely replicates plant data. b. Numbers in parenthesis pertain to RELAP5 volume l The tripping of the recirculation pumps induces a numbers. I rapid decline in the core inlet flow. Again, data A 12

2000 i i i i a PLANT DATA RELAPS L

m 1500 -

s Q.

b o

o

& 1000 m

O.

E 2

O- l 500 -

1

' ' ' ' i 0

0 5 10 15 20 25 Time (s)

Figure A-6. Pump trip test data comparisons - pump speed.

comparison is favorable, as shown in Figure A 7. The The second and final plant transient data comparison decreasing core inlet flow tends to increase the void performed compared the plant and model immediate-content of the core, causing a rapid decrease in the ly following a generator load rejection. The generator reactor power. Reactor power is shown in Figure A-8. load rejection transient sequentially clarifies the effect j l The core responds much more quickly than the inlet of several events: (a) the effect of increased steam de- 1 flow to the recirculation pump trip. The core flow mand on the vessel, (b) the effect of a reactor scram, )

gradually coast down to about 40% ofinitial over the and (c) the effect of simultaneous MSIV closure, recir-25 s of the transient simulated. The reactor power, culation pump trip, and HPC1/RCIC activation on the however, responds to the external perturbation in ap- reactorvessel.

proximately 10 s. The power transferred to the core Transient initiation is induced by the rejection of the I inventory vaporizes a portion of the inventory and electrical load impressed on the main generator (see causes the reactor to steam. The response of the vessel Reference A-1). This load rejection causes the turbines steaming, Figure A-9, and vessel pressure, Fig- to overspeed. The turbine control valves begin to close, q

ure A-10, indicate that the response of the main steam- and the turbine bypass valves open to reduce the turbine line and/or turbine typify the data reasonably well. overspeed. The overspeed continues, however, the re-Downcomer liquid level is shown compared to plant actor scrams at 1.6 s and the turbine stop valves close.

data in Figure A-11. When considered in the context This sequence of events results in downcomer level of the wide range instrumentation, the agreement with depression, and double low level is reached 6 s into plant data is somewhat poor. The collapsed downcomer the transient. The recirculation pumps trip, the main liquid level is measured in the annular region outside steamline isolation valves close, and the HPCI and of the dryer shroud. The level signal is used to actuate RCIC systems are activated.

many important systems such as the HPCI system. This Figures A-12 through A-16 show the response of the level is strongly affected by the model nodalization, plant and the model to this sequence of events. In Also shown is a pressure differential signal generated general, the correlation of plant data to the model is from RELAP5. There exists an uncertainty in this good, signal which influences transient timings related to in the load rejection transient, the vessel responds  !

level actuated signals, to steamline perturbations. Figure A-12 shows the A-13

100 r , , , ,  !

O PLANT DATA j 90 - 0 RELAP5 -

O R

v 80 -

0 3

o 70 - -

C D e 60 - -

.C_

e 50 -

O u

o

" U 40 -

O 30 - -

20 O 5 10 15 20 25 Time (s)

Figure A-7. Pump trip test data comparison - core inlet flow.

100 r , , , ,

O PLANT DATA l RELAP5 90 - -

7 u

80 - -

e 3 O o

Q-70 - -

e o

" O O

8 60 -

O O

50 -

u h u

~

O '

U

,0 Oa O 40 0 5 10 15 20 25 Time (s) l l

Figure A-8. Pump trip test data comparison - reactor power.

l A-14 l

110 i i i i O PLANT DATA RELAP5 100 t l 2  ;

" go _

t O ,

S 80 -

= i i

E

$ 70 -

M O O

60 u -

o 50 0 5 10 15 20 25 Time (s)

Figure A-9. Pump trip test data comparison - steamline flow.

l 1020 i i i i O PLANT DATA t a RELAPS n

. 1010 -

m S

e 1000

?n a -

L CL 990 - -

D e

E o O O a 980 _ _

c D b o e O

- O a ooa w 970 -

aao aO 960 O 5 10 15 20 25 Time (s)

Figure A-10. Pump trip test data comparison - steam dome pressure. I j

A 15 i

50 i i i i O PLANT DATA RELAP5 o WIDE RANGE AP 45 -

A COLLAPSED LIQUID LEVEL Z  !

O O 0 O U I

_ 0 0 0 g O e

$ 40 -

0 O ao 8 0 m OO cn O ~

C cE35 g

- ~

O o

TJ 3

30 - -

l t

25 O 5 10 15 20 25  !

Time (s) l Figure A-11. Pump trip test data comparison - wide range level.

I 150 i i O PLANT DATA RELAP5 I

^

O b OO 3 100 t -

O C

o C

'3 E

8 50 - -

s i m O j g

O U O a O O

O O C 0

O 5 10 15 Time (s)

Figure A-12. Generator load rejection data comparison - steamline now.

A-16 l

15 0 i i O DATA RELAP5 m

M 100 cR 8 -

t c0 3

O k

u O

e U

8 50 -

Q:

0 b

S * "

0 0 5 10 15 Time (s)

Figure A-13. Generator load rejection data comparison - reactor power.

1150 , ,

O DATA

^ RELAPS O

To 8

0 1100 -

D - I v

g O C l u

1 O O O O O E aO O

U 1050 - -

E O

O en i

O U ' '

1000 i

O 5 10 15 Time (s)

Figure A-14. Generator load rejection data comparison - steam dome pressure.

A-17

12 0 , ,

O DO O DATA D 0 RELAPS OO 0 0 CD e

b O R

v 100 e _

3 0 0 iZ O

% 80 - -

3 o e

L.

o O

60 - -

40 O 5 10 15 .

Time (s)

Figure A-15. Generator load rejection data comparison - core inlet flow.

1 l

40 au i i yp D PLANT DATA RELAPS i o WIDE RANGE AP i

^ a COLLAPSED LIQUID LEVEL z 20 - _

O f l i

s j

• g 0 -

o g E

O O O l o O '

Cr O U

$ 00 C a

O 5 -20 -

O OO _

OO

-40 O 5 10 15 Time (s)

Figure A-16. Generator load rejection data comparison - wide range level.

A-18 j l

1

sessel steam flow as measured at the main steamline trip w as specified according to plant data. Figure A-16 flow restrictors. The initial, sharp decline in steam shows the level response during this transient. For the flow is caused by the turbine stop valve closure. This purposes of this simulation, double lo level was decline is arrested by the action of the turbine bypass assumed to be reached 6.4 s into the transient.

valve. The flow rate is relatively constant until 7 s, The results of these plant transient data comparisons  ;

when the closing action of the MSIVs becomes pro- are mixed. In general, recirculation pump speed, core nounced. Upon their closure, a no-flow condition is inlet flow, and reactor power and pressure show good  !

reached. It is thought that the steamline flow described data agreement. The model is expected to be able to by the data is erroneous after 10 s. Flow cannot reach reliably predict these parameters. On the other hand, l the flow measuring devices once the MSIVs have vessel steaming and downcomer level response are not closed, as well predicted by the model during the transients The response of the reactor power to the lead rejec- presented here. The exact cause of the deficiencies in tion is straightforward, as shown in Figure A-13. At these signals is not well underst<xid, but the significance 1.6 s, the reactor scrams and power decreases nearly of the deficiencies to ATWS simulation can be monotonically. estimated. '

The vessel or steam dome pressure responds to both It is thought that the deficit and uncertainty accom-steam flow and power. The turbine stop valve closure panying the vessel and/or core steam are of marginal causes a rapid vessel pressurization (Figure A-14). significance when interpretated in terms of ATWS This pressurization is limited by the relieving action simulation. This deficiency is induced by the short-term of the turbine bypass valve, and by the reactor scram. transient behavior. Over the typical duration of an Vessel pressure declines; but again rises due to the ATWS simulation of many hundreds of seconds, this

! MSIV closure. The rate of subsequent pressure in- short-term transient behavior will subside. Short-term creases is influenced by the relieving action of the behavior is governed by the momentum equation. Over HPCI and RCIC steam driven turbines No safety / relief a long term, momentum effects will be dominated by l

valves were actuated during the simulation. the mass and energy balances of the vessel, which are l

l Core inlet flow response shows reasonably good data usually independent of the momentum balance.

I agreement, as is evidenced in Figure A-15. The rapid The level signal is regarded as having large dynamic vessel pressurization induces a sharp increase in core uncertainties. The significance attached to this signal inlet flow. The flow increase is limited by the response is attached to operator action. During the normal course of the core thermal-hydraulics. The core thermal- of a transient, it is expected that the operator will hydraulics respond to the steamline flow perturbation monitor the downcomer level signal. Thus, simulation l much more rapidly than the pressure of the vessel does. timings contingent on proper downcomer simulation i Flow coastdown is induced by the tripping of the recir- will be somewhat in error. However, in the long term, culation pumps on double lo level. vessel mass and energy balances will dominate momen-Because of the poor downcomer level response tum induced effects, so that the error in the level signal shown by the model, the timing of the double lo level will not be divergent. l 1

i l

I l

I A-19 I

_]

REFERENCES A- 1. Interoffice correspondence, J. A. Raulston (TVA) to R. R. Schultz (INEL), October 24,1980.

A-2. R. R. Schultz and S. R. Wagoner, The Station Blackout Transient at the Browns Ferry Unit One Plant:

A Severe Accident Sequence Analysis, EGG-NTAP-6002, September 1982.

A-3. 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.

A-4. W. C Jouse and R. R. Schultz, REIAP5/ MODI 6BHR/4 A7WS Demonstration Calculation on the Browns Ferry Unit 1 Nuclear Plant, EGG-SAAM-6397, November 1983.

A-5. W. C. Jouse, Sequence Matrixfor the Analysis ofan ATWSin a BWR/4; Phenomena, Systems, and Opera-tion of Browns Ferry Nuclear Plant Unit 1, a draft report for comment, May 1984.

A-6. Design and Performance ofGeneral Electric Boiling Water ReactorJet Pumps, APED-5460, GE, September 1968. l 1

I A-7. G. E. Wilson, INEL One-Sixth Scale letpump Data Analysis, EGG-CAAD-5357, February 1981.

A-8. Interoffice correspondence, C. J. Prone (GE) to J. T. Robert (TVA), " Browns Ferry 1, Cycle 6 Reference Loading Map", CJP: 83-047, March 1983.

A-9. R. C. Stirn, Generation of Void and Doppler Reactivity Feedback for Application to BUR Design, NEDO-20964, GE, December 1975.

A- 10. R. J. Grossenbacker, et al., BUR /4 and BHR/5 Fuel Design, Licensing Topical Report, NEDO-20944, GE, October 1976.

A-11. D. J. Diamond, Generation ofBHR Point Reactivity Famctions, BNL-NUREG-32412, September 1982. j A-12. D. J. Diamond and H. S. Cheng, " Higher Order Effects in Calculating Boiling Water Reactor Doppler ,

and Void Reactivity Feedback," Nuc. Tech.,46,439, December 1979.

A-13. InterofGee correspondence, G. C. Slovik (BNL) to W. C. Jouse (INEL), December 28,1984.

l A-14. S. L. Forkner et al., BHR Transient Analysis Model Utilizing the REIRAN Program, TVA-TR81-01, l December 1981.

A-15. W. Frisch et al., "The Signincance of Fast Moderator Feedback Effects in a Boiling Water Reactor Dur-ing Severe Pressure Transients," Nuclear Science and Engineering, 64,1977, pp. 843-848.

A 16. Browns Ferry Nuclear Plant Final Safety Analysis Report (FSAR), Tennessee Valley Authority.

A-17. General Electric BHR Technology, NEDO-10260.

l A-18. Interof fice correspondence, W. D. Driskell to E. T. Laats, " Power Distribution for the Browns Ferry Cycle 4 Core", WED-I-83, August,1983.

A-I9. B. Chexal et al., Reducing BWR Power by Water inel Control During an ATWS, NSAC-70, August 1984 A-20

APPENDIX B DESCRIPTION OF BFNP1 PRIMARY CONTAINMENT AND CONTEMPT /LT-028 MODEL I

j t

B1

APPENDIX B DESCRIPTION OF BFNP1 PRIMARY CONTAINMENT AND CONTEMPT /LT-028 MODEL

1. BROWNS FERRY CONTAINMENT DESCRIPTION in this section, the Browns Ferry Nuclear Plant additional resistance to deformation. The drywell Unit I (BFNPI) primary containment is described as houses the nuclear steam supply system (NSSS), in-an introduction to discussion of the CONTEMPT / cluding the reactor pressure vessel, recirculation loops, LT-028 code model of the containment. For a detailed steamlines, safety / relief valves, feedwater and emer-description of the containment geometry, refer to ' gency core coolant systems piping. The drywell is con-References B-1 and B-2. nected to the wetwell by eight large ventilation pipes, The BFNP1 primary containment is the Mark I each 81 in. in diameter and arranged symmetrically General Electric Design. A Mark I containment is about the periphery of the drywell sphere lower half.

composed of essentially two major parts; the drywcIl These vent pipes protrude into the wetwell and con-and the wetwell. A representation of the lightbulb- nect to a 57-in. diameter ring header which itself forms -

shaped drywell/ torus wetwell configuration is shown a torus within the wetwell proper. Ninety-six down-in Figure B-1. The Mark I containment is used in the comer pipes protrude from the ring header, each two majority of the operating BWRs in the United States. feet in diameter, and normally submerged at least three The drywell is a steel pressure vessel consisting of feet into the torus pool (see Figure B-2).

a spherical lower section 67 ft in diameter and a cylin- The torus has an inside minor diameter of 31 ft and drical upper section 38 1/2 ft in diameter. The drywell a major diameter of 111. 5 ft. Nominally, it contains is enclosed in reinforced concrete for shielding and about one million gallons of demineralized water and j

t * * *] ' .....

Recetor pressure vessel ,

)

'l b l l

t / Drywell

/ s ,

j L

t

C,s suam

" i.

[ '.

Moin sisom line /-

4' / l,, .... - Voeuum relief volve i

\ l*' R h =

[

. ., , , 't.,

}

Moln ,-

vent /

\ ewncomer 0 O.tV / rs _ m/ .

m L226-KW253-02 Figure B-1. Mark I drywell/ torus configuration.

B-3

. . . ..l

DRYWELL >

VENT FEADER BELLOWS MAIN VENT

\

.........-::::)

~~'....

I m S/RV Lite

( #

POOL SURFACE

/ \

q w-DOWNCOhCR PRESSURE SUPPRESSION POOL b VENT HEADER SUPPORT D_, '

ECCS HEADER N5100374 Figure B-2. Cross section of the forus.

is maintained within the normal operating water level containment safety sprays. It serves as a secondary range from 14.6 ft to 15.1 ft, nearly half the torus water source for the high pressure coolant injection diameter. The torus is designed to contain the initial (HPCI) and reactor core isolation cooling (RCIC) energy release during a double-ended large break in systems. The toms also accepts steam exhaust from the 28-in. reactor recirculation pump line, sustaining the HPCI and RCIC turbines.

a 50*F temperature rise in the pmcess. Maximum torus The torus hs an external design pressure of 2 psid initial water temperature is 95'F per technical and is protected from collapse by two 18-in. vacuum specifications constraint. relief valves, u hich equalize pressure between the reac- l The torus is one of the distinguishing features of the for building and the wetwell within one second after f Mark I containment design, and it performs a number a 0.5 psi pressure differential is sensed. Internal design of important functions. It will accept vessel effluent pressure of both the torus and the drywell is 56 psig during a loss-of-coolant accident (LOCA) through the at 281 F.

drywell/wetwell vent system. The torus condenses During normal operation, the drywell is inerted with steam released from the vessel through the safety / relief nitrogen and positively pressurized with respect to the valves during off normal plant operation or an acci- wetwell by 1.1 to 1.3 psid. System pressure is nor- 3 dent. Vessel steam diverted in this manner is dis- mally maintained by a compressor and a differential j charged to the pressure suppression pool through pressure control system between the wetwell and the j sparging devices known as the T-quenchers. One pur- drywell. The positive pressurization maintains minimal l pose of the T-quencher is to maintain k> cal vek> cities liquid column height inside the downcomer pipes (the j at the discharge point sufficiently high to avoid drywell side), which would reduce dynamic loads in  ;

hydrodynamically induced pressure oscillations in the the torus during an accident in which the drywell is {

pool. The T-quencher is also designed to ensure that pressurized early and the vents clear. Should the complete steam condensation occurs in the torus pressure differential reverse and exceed 0.5 psid, there preventing pressurization of the wetwell and drywell. are twelve 18-in. swing check vacuum relief valves The torus is the primary source of water for the low (VRV) k>cated at the ends of the 81-in. vent pipes pressure coolant injection (LPCI) and low pressure which would relieve pressure from the wetwell to the ,

core spray (LPCS) emergency systems as well as the drywell. One purpose is to prevent a vacuum in the i I

IM ,

l l

l drywell when the drywell sprays are actuated during systems) which exchange heat with the reactor building a large break LOCA, preventing liquid backflow into closed cooling water (RBCCW) system. Eight of these :

the drywell. Another purpose of these VRVs is to units are normally operational with two spares. The alleviste the effects of wetwell pressurization during' operating load capacity of the drywell cooling system extended discharge through the SRVs. is about 5.2 million Btu /hr. Normal dry,vell tempera-The prenure suppression pool temperature can vary ture averages about 125'F. Bulk temperatures up to considerably depending on the time of year and the 135'F are acceptable according to design specifica-status of plant operatior.. The technical specifications tions, restrict the water temperature to less than 95'F. The The drywell and torus together form the primary l water is maintained within plant specifications by the safety barrier between the NSSS and the reactor residual heat removal system, whose four heat ex- building, the reactor building serving as the secondary changers combined have a design capacity of containment and the final safeguard between power 280 million Btu /hr, about 2-1/2% of rated power. The plant and environment Detailed discussion'of the secondary containment is beyond the scope of this  !

cooling capacity is 9 function of pool temperature, in-creasing with increasing temperatures. report. However, the reactor building does house a Drywell temperature is maintained below 135 F by number of systems which couid be important to acci-ten fan cooling units (arranged in two large ab handling dent management in different situations. .l

2. MODELING THE BROWNS FERRY CONTAINMENT l

In this section, the CONTEMPT /LT-028 computer either single or double sided. A number of heat transfer code and modd of the BFNPI containment are dis- options are available for slab boundary conditions, cussed. Section 2.1 gives a brief description of the code Mass and energy additions to a compartment can also and its capabilities. The BFNPI containment input be input as boundary conditions in a tabular format at model is next discussed, and the calculated initial con- the user's option, ditions are presented in Section 2.2. CONTEMPT /LT-028 has sufficient latitude to ,

U enable modeling of fan coolers and various ECC and s fety spray sys'. cms. There is a vent model which 2.1 Code Description calculates two-phase, two-component flow through the  ;

The code used for the analysw documented herein was vent system between the drywell and wetwell. This  ;

CONTEMIT/LT-028 (Reference B-3). CONTEMIT is model is basically a series of flow resistances for the a containment thermal hydraulics code originally purpose of fluid transport between the drywell and developed at the Idaho National Engineering Labora- wetwell. There 'is no volume or mass associated with  ;

tory (INEL). The code has been used for licensing pur- the vent model and it is most suitable for forced con- )

poses and was chosen for the ATWS studies because vective flow per the LOCA analysis, original design it has been assessed to a greater degree than other con- intentions of the code, tainment codes available at the onset of the ATWS Increased tabular data input capacities have been calculations (see References B-4 and B-5). ' programmed into CONTEMPT /LT-028 to facilitate CONTEMPT /LT-028 is a lumped parameter code interfaces with driver codes such as TRAC-BD1 or l which was originally designed for LOCA analyses. The RELAP5. Such data transfers are done with computer l code numerics are explicit. Current modeling capability software to minimize the chances of error. Safety / relief includes BWR Mark I, II, and 111 plus PWR dry con- valve blowdown to the wetwell pool is treated as a tainments. The code has a generalized, lumped volume boundary condition in the BFNPI ATWS modeling, compartment model that allows liquid and vapor to and a tabular input option has been coded to enable exist in nonequilibrium. The state of each phase is this condition. SRV mass flow rate and energy addi-uniform throughout that phase. Heat and mass transfer tion rate versus time can be input, with up to 2000 data provisions exist between the phases, and leakage to and points allowable for any one problem.

from a compartment can be modeled. Energy additions CONTEMIT/LT-028 is an inexpensive, fast running to a compartment can occur from heat structures,20 code. Typical one-hour BFNPI ATWS problems run of which are permittect. The heat structure model (for on a Control Data Corporation CYBER 176 computer stored energy or heat sink) is distributed, allowing up for under $30, at a rate approximately 14 times faster to 101 mesh points per structure. A heat slab can be than real time.

B-5.

As p rt f the quality assurance program for the input l 2.2 Model Description model, steady state calculations were run to generate '

The BFNPI containtnent model used for the reported initial conditions of drywell and wetwell pressure and calculations was based on infortnation from several temperature for the ATWS transients. This procedure sources, the Tennessee Valley Authority, the BFNPI was helpful in the isolation of input errors and often FSAR, etc. It has been constructed with best estimate suggested modeling improvements. Importantly, it methods whenever possible. In cases where informa- assured that the initial conditions were stable values tion was not available, conservative engineering judg- free of significant numerically induced drifts.

ment was used. For example, the heat structures be- The drywell (DW) and wetwell (WW) initial condi-tween the drywell and reactor building are insulated tions calculated by CONTEMFr are compared to plant in the model and no credit is taken for heat loss to the operating specifications in Table B-2. The steady state reactor building, comparisons are generally good. The wetwell pressure The CONTEMPT /LT-028 model for the BFNPI is is slightly high, resulting in a calculated DW to WW shown in Figure B-3. One compartment each is used differential pressure that is too low compared to plant for the drywell and wetwell. For the analyses docu- specifications. Because the compressor and pressure mented herein, it has not yet proved necessary to model control system are not modeled, the drywell and wet-the reactor building. Instead, heat structures represent- well are driven closer to an equilibrium state at the start ing the reactor building provide boundary conditions of transient calculations than would be the actual case, for heat transfer from the torus. There are 11 heat However, the lower drywell pressure results in a slight-structures in the model, as detailed in Table B-1. This ly longer time to high drywell pressure (HDWP). The tabulation lists the heat structures, the compartments initial DW pressure is within the normal operating to w hich the structures are connected, and the surface range of plant conditions, and the slightly higher than boundary conditions of the heat slabs. Vessel heat slab normal wetwell pressure has a negligible effect on the initial conditions were taken from RELAP5 calcula- ATWS.

tions. Drywell and wetwell stab initial conditions were Drywell temperature is controlled by the drywell based on as erage operating conditions where possibie, coolers. These are modeled by a constant heat removal but these sometimes required engineering judgments rate from the drywell for these calculations, as where detailed data were unavailable. CONTEMirr has no mechanistic fan cooler model. For C3

.g .. . .

.i y O c' sJ

7 04  ! % + c7

! s ,

ii

$ i

/

.' d oe o3

!(N r= '. :i

! s

.ce: >

.. . . . . . , t j

;  % /F% h M.

T ' r

~ '.,

2

)

L226-KW263-C3 10 Figure B-3. CONTEMPT Browns Ferry containment matel.

IM

S n a

t o

s t

l u

0 i f

a 1 t S S e l l d l l n

o o i o

t t t p

n n n o o o o A

i i i t t t a ta a h i i D

ig d d id a A d ed e DA I

a a A A d d e I R r r r e I n D D D t t C

& H i lua la t

o & & l l

t a l a

l l

l I l u U lu i

i l i

d t

C C C s u C C C C s s n n s

h e

o n N N U I / / / /

I n n U N U U I /

/ / /

I

/ /

I T

C WWWWWWWWWWW r y

r a

W W D D D WD O D W W i a

o t

d r n

u o p

o a B

! v n 9 7 1 2 r o 5 S S S e i

t c

1 1 I I

1 1 H t a

e n t o

t o o o t t f

w n C n n o o C n C C T TI it o o o i

t i

t i o

t T

I I

I I a i ia a t a

g f I 8 .

i d d r e A t t

. C d a a s n L t s A A A s a r r r s .

n D no D D D on N . + + + a t i

a s le t s

I I

C ll l l

l l C b mix

- d n I i i i e e o o C 5 C C C 5 r u C C C ht t m C U Rs U U U R

/ / / T N N N / / n n o

e / W se W W W W

/ / /

e s

s

/ /

WWWW t o o r

u W WV D D D V D D W W nd e n

_ t c dni o et u

r pa s t

e n s d e y ed t d e) 0 0 7 5 5 5 5 5 5 2 4 u n a aF la o t

a 9 9 2 3 3 3 2 3 3 8 e t e t

• S( 5 1 1 1 5 1 1 01 v c h S r y

mo a f r e s r s n t

e gi o F t s d inis d n a e

n n o e

w h n d c nt o m o oc r o it a

c e r B t t

d a r n o o n d is c l

8 o b l u ae 2 a 0

i t

& w o t e

- ip r f h nh e

f g e r t

f T

L r

c l

a l

a d d

e e l

le l d

le n v i

icf s h h a r h a f i

e n e i d o he i i

/

r S t

s h fe n T D r o ly h p m e s l iu m o oi o P te p p c s e ly. d e l ct p c a a o t b i w v a s c o r d M t l le l

le y r o

r e n le e ig o l s l f E s s e w wy t s s T u r

ur s s y t

n s

s s

s lo c a

u r

n c o o e r e o a y e i Re oT N r e r r O T T V D D V V V B t t

a ad C e n h u o

1.- Ab B e Dno l

l r i i

t u Cc e le t c 1 2 3 4 5 6 7 8 9 0 1 b u 1 1 Uv a r n t

. o T S a c

Table B-2. ATWS calculation initial conditions compared to plant specifications CONTEMPT /LT-028 Calculated BFNP1 Specifications Wetwell Drywell Wetwell Drywell Pressure 4 ? 14 psig + 1.38 psig 0.5 psig max., 1.8 psig max.a

-0.5 psig min., 0.6 psig min.

Temperature 90 F 134.9 F 33'F-95

• F 135'F Humidity 0.97 0.20 - -

Torus water level 175.8 in. 0.0 175-1/4 in. -

to 181-1/4 in,

a. BFNPI specification on DW to WW AP is 1.1 to 1.3 psid. See discussion in text, this model, a heat removal rate of 3.75 million Btu /hr valves are fully open simultaneously and instantaneous-(1.1 MW) was used, approximately equivalent to the ly (with no opening delay) when the required differen-load capacity of eight drywell cooling units. tial pressure is reached. This is a limitation of the code, j The wetwell to drywell vacuum relief system was not user specified, but should have very little effect modeled as twelve 18-in. relief valves, each with a llow on the results. The CONTEMPT /LT-028 BFNPI con-area of 1.67 square feet. The relief valve model was tainment model has been reviewed, including a detailed programmed to open at a pressure differential (WW engineering review of the input deck. The model is to DW) of 0.5 psid. The code assumes that all 12 believed to be an accurate representation of the plant.

REFERENCES i

l B - 1. General Electric Company, " Containment Data", Customer Interface Data Document No. 22A5737, Rev.1, April 1979.

B-2. General Physics Corporation, BilR Simulator Training Afanual, Second Edition, Volume 1,1979.

i B-3. D. W. Hargroves and L,. 3. Metcalfe, CONTEMPT-LT/028-A Computer Programfor Predicting Con-tainment Pressure-Temperature Response to a Loss-of-Coolant Accident, NUREG/CR-0255, TREE-l279, EG&G Idaho, Inc., March 1979.

B4. G.1. E. Willeutt, Jr. and R. G. Gido, Comparison of CONTEMPILLT Containment Code Calculations with Marviken, LOFTand Battelle-Frankfurt Blowdown Tests, NUREG/CR-l564, LA-8423-MS Informal Report, July 1980.

B-5. B. S. Anderson, Comparison ofCONIEMP1;LTPredictions to Marviken ErperimentalData,1NEL lnterim Report 1-347-75-02, Aerojet Nuclear Company, June 1975.

B8

APPENDIX C DESCRIPTION OF THE SCDAP CODE AND THE BFNP1 SCDAP MODEL l

1 l

C-1

i APPENDIX C $

DESCRIPTION OF THE SCDAP CODE AND i THE BFNP1 SCDAP MODEL During the high pressure boiloff portion of the plant

• Rates and chemical forms of released fission gases automatic transient, cladding temperatures were pre-

• Rates and characteristics of debris formation dicted to reach the threshold point for cladding J oxidation. To obtain information about hydrogen pro-

• Coolability of the disrupted core geometry and/or duction and fission product release, a SCDAPN debris calculation was performed. Results from that analysis

• Vessel internal state up to loss of the core sup-were presented in Section 6.2 of this report. p rt structure integrity, The SCDAP analysis was initiated a short time before the onset of zirealoy oxidation (>1340'F) of SCDAP/ MOD 1/VO is a modular computer code and the cladding. The calculation was terminated as the rate has been designed so that either all of the code or in-of hydrogen generation approached zero. SCDAP was dividual subcodes and models can be readily used to used because it was specifically developed to predict perform calculations. This modularity allows the user core behavior from the onset af zircaloy oxidation and code developer (a) to tailor calculations for par-through severe core damage. RELAP5 results were ticular problems or experiments, (b) to test individual used to provide the initial boundary conditions for the parts or all of the code, and (c) to test alternative SCDAP case and were then extrapolated to provide the models or subcodes to establish the sensitivity of dif-boundary conditions for the entire SCDAP analysis, ferent modeling assumptions. For this case the sub-This was possible due to the almost steady-state con- code, SCDVESSEL, was used. SCDVESSEL will dition of the plant during the postulated core damage compute the behavior of a single fuel rod, control rod, i scenario. shroud, an individual bundle, or a reactor vessel con-The version of the SCDAP code used for the analysis sisting of core, upper and lower plena, and layered  !

was SCDAP/MODl/VO subcode SCDVESSEL. outer shroud. The simplified boiloff thermal-hydraulics SCDVESSEL has simplified boiloff thermal hydraulic model, developed for the fast running version of models which are fast running. This version also had SCDAP/ MODO (designated SCDSIMP) was used.

an upper vessel model to provide information on SCDVESSEL is a pilot capability designed to test the damage progression in that region of the vessel. models which treat plenum behavior anr1 radiation heat The discussion in this appendix is divided into three transfer in arbitrary geometries. SCDVESSEL can be parts. A description of the SCDAP code is given in used for analysis up to the point of debris formation Section 1 Section 2 discusses the model used in the through quench-induced fragmentation. SCDVESSEL study. Section 3 reviews the boundary conditions used consists of SCDSIMP, and pilot plenum and radiation in this part of the study. models.

SCDAP/MODl/VO simulates disruption within the wssel through a detailed analysis of representative

1. Code Description vessel components. The code can analyze a boiling The SCDAP/MODl/VO computer code models the water reactor (BWR) vessel and components, a pres-progression of a postulated event or experiment up to surized water reactor (PWR) vessel and components, and including core geometry changes and raaterial and experimental bundles through a user defined relocation due to severe overheating and fragmenta- description of the components, bundle, and vessel. The i tion of embrittled materials. Important phenomena code consists of detailed physical models for uranium occurring in both the upper and lower plena are dioxide (UO2)-zircaloy fuel rods, thermal models for modeled. Cladding and stainless steel oxidation, one-dimensional cylinders and slabs, plenum structure aydrogen generation, and fission product release are models, and general one-dimensional thermal-modeled. Cladding ballooning and rupture, material hydraulics models for both the core and plena. Models 1 liquefaction and relocation, component fragmentation are provided to describe the formation and thermal-during reflood, debris famiation and thermal-hydraulic hydraulic behavior of the plena and any core debris behavior are also modeled, Major outputs include: which may form as the vessel components are disrupted through either a material liquefaction and redistribu-

• Rates of hydrogen generation due to oxidation tion process or a quench induced material fragmenta-reactions tion process.

C-3

l'

2. SCDAP Model Description Table C-1. SCDAP input parameters The SCDAP model used for the analysis consisted Parameter Units Value of a core model and an upper vessel model. The core inlet conditions were provided by the RELAPS studies. Active fuel length ft ' 12.5 The core outlet flow provided the inlet condition of Fuel pellet radius ft 0.0171 the upper vessel regions.

The core model used for this analysis consisted of Cladding outside radius ft 0.02012 one fuel bundle including the canister, typical of those Cladding inside radius ft 0.01746 m the BFNPI plant. The fuel rods were in an 8 x 8 configuration. The bundle was divided into five Axial node heights ft 1.25 regions, called components. The fuel rods were divided Pitch ft 0.0535 into three rod components, describing the radial power distribution. One component was used to model the two Coolant pressure psia 1040 water rods in the bundle and one component was used (transient initiation) to model the canister.

Information for the model was obtained from the Final Safety Analysis Report (FSAR),C2 or was provided by ,

the Tennessee Wiley Authority (TVA).UUU In '

accordance with the available data, the bundle was conditions from the core and are able to describe the assumed to be in the eleventh month of the sixth cycle. processes of oxidation and melting of upper vessel The hot bundle from ring five was modeled. The structures. The structures were scaled by the number bundle burnup was 10064 MWD /MTU. The control of bundles in the core, rods were assumed to be completely withdrawn at the time of the transient and were not modeled.

'Ihe fuel was divided into 10 equally spaced axial

3. Boundary Conditions regions of 1.25 ft. Ten regions is the maximum The SCDAP study was an extension of the allowed by SCDAP. Radially each rod group was RELAP5/ MODI,6 analysis. The SCDAP calculation modeled with six nodes, four in the fuel and two in was initiated a short time before the onset of zircaloy the cladding. Other input variables are listed in oxidation. The required boundary conditions taken Table C-1. from the RELAPS analysis were enthalpy and mass The vessel above the core was modeled to describe flow of the fluid entering the core, coolant pressure seven distinct regions associated with the reactor at the inlet to the core, and core power. Enthalpy, design. The seven regions are: pressure and flow boundary conditions were required during the entire SCDAP calculation. The boundary

1. The area from the top of the active fuel length conditions from RELAP5, were in a quasi steady-state to the top of the camster, condition during the time that the SCDAP calculation

2. Upper core grid, was conducted. Therefore these conditions were ex-

3. The region from the top of the canister to the top ""E '" # * # " # # "## E" # " "#"""'Y of the core shroud head, put or A was na at t DM com analysis could affect the system response (such as

4. The region from the top of the core shroud head hydrogen generation increasirg system pressure and to the top of the stand pipe, changing inlet flow) and in many cases extrapolation

. of the required boundary conditions may not be valid

5. The separator region, and an iterative approach between RELAP5 and

6. The dryer region, and SCDAP would be necessary. In this case, however, the core inlet flow was very small and the enthalpy and

7. The steam dome reg. ion.

pressure were changing s'owly (pressure was held fair-In the upper vessel region, SCDAP only models zir- ly constant due to SRV actuation) and consistently.

caloy and stainless steel. Since the first two regions Thus it wasjudged that the RELAP5 boundary condi-were mostly zircaloy and the other five mostly stainless tions and the SCDAP calculated system conditions steel, the upper vessel region could be adequately were sufficiently decoupled to provide a valid analysis, modeled. The upper vessel models used inlet coolant as described. d C-4

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

REFERENCES C- 1. G. A. Berna et al., SCDAP/MODUVO: A Computer Codefor the Analysis ofDVR VesselBehavior During Severe Accident Transients, IS-SAAM-4-002, January 1984.

C-2. Browns Ferry Nuclear Plant 1-inal Safety Analysis Report (FSAR), Tennessee Valley Authority.

1 C-3. R. J. Grossenbacker et al., BWR/4 and BWR/5 fuel Design, Licensing Topical Report, NEDO-20944 Rev.1, October 1976. .

l C-4. Systems Manual, Boiling Water Reactor, USNRC Inspection and Enforcement Reactor Training Center, i Volume 1, Chapter 2, February 1980.

C-5. Supplemental Reload Licensing Submitral for Browns Ferry Nuclear Plant Unit i Reload No. 4 Y 1003J1 A19 Rev. O, March 1981.

i C-5

I

)

f i

f

' APPENDIX D FRAP-T6 CODE AND MODEL DESCRIPTION I

I l

l D-1 4

w____

APPENDIX D .1 FRAP-T6 CODE AND MODEL DESCRIPTION During the plant automatic simulation described in the rod was divided into 1i radial nodes,9 in the fuel i Section 5.1, large power excursions were predicted. and 2 in the cladding. The radial power profile was [

These excursions were caused by unthrottled flooding obtained from the FLXDPa2 subcode and the axial of the RPV by low pressure injection systems. In all, power profile was obtained from Reference D-3. The seven power spikes were predicted between 1275 and fuel was UO2 , with a 95% theoretical density. The fill 2600 s in the transient. To determine the probability gas was helium. The rod was pre-pressurized to and extent of fuel damage during the power spikes, 44.0 psia (see Reference D 3).

the FRAP T6 43 code was used. Results of the code For this calculation, the FRAP-T6 code was used analysis were presented in Section 6.1. with the FRACAS I cladding deformation model. With -

This appendix presents a description of the FRAP- FRACAS I the stress induced deformation of the fuel T6 code and input model. Also described are the pellet is ignored. This feature is more conservative than boundary conditions used in the FRAP-T6 analysis. the FRACAS 11 model which allows for stress induced deformation of the cladding. FRACAS I was used due j

• * "VerSence Problems found in FRACAS II, when

1. FRAP-T6 Code Description the cladding temperature exceeds 1000 F.

The FRAP-T6 computer code calculates the thermal The following assumptions are found in the and mechanical behavior of light water reactor fuel FRACAS I model:

rods du;ing a wide range of operational transients and

• Increment theory of plasticity hypothetical accidents. Models m the code calculate fuel and cladding temperature distributions, rod inter-

• Prandtl-Reuss flow rule )

nal pressure, and clastic and plastic fuel and cladding

• Isotropic work-hardening deformation. These rod behavior areas are treated both as interdependent parameters and as functions of a No low temperature creep deformation of cladding power, rod design, burnup gap and surface heat

• Thin wall cladding (stress, strain, and temperature transfer, and changes m material properties. In addi-uniform through cladding thickness) tion, such parameters as cladding oxidation, CHF, and conditions at cladding failure are calculated.

• No axial slippage occurs at fuel cladding interface The user must input the initial rod geometry, coolant when fuel and cladding are in contact channel geometry and boundary conditions, and power

• Bending strains and stresses in cladding are history. If the rod has experienced prior burnup, mi-negligible tial conditions for a transient may be supplied through a data file from FRAPCON-2, the companion steady

• Axisymmetric loading and deformation of state fuel rod analysis program. cladding.

Several options are available to the user. The FRAP-l T6 code is dimensioned to handle rod arrays oflimited Deformation and stresses in the cladding in the open

! size, but currently no feedback is provided to account gap regime are calculated using a model which con-  !

for subchannel interactions occurring as a result of siders the cladding to be a thin cylindrical shell with I coolant flow redistribution, cladding deformation, or specified internal and external pressures and a pre-fuel rod failure. An option is available by which uncer- scribed uniform temperature. Fuel cladding mechanical tainties in fuel rod behavior are calculated due to uncer- response is important since this mechanism effects heat l tainties in fuel rod design, power, coolant, and material transfer across the gap as well as providing informa-properties. In addition, choices can be made among tion on cladding ballooning and failure. FRAPCON-2 several deformation, film boiling, and critical heat flux (Reference D-4) provided the fuel conditions at the l correlations. Also available is an option that calculates start of the transient. To be consistent with the SCDAP fission gas release, analysis, burnup was assumed to be one year.

The Carthcart oxidation model was used to predict cladding xidati n when cladding temperatures ex-

2. FRAP-T6 Model eceded 1340'F. The no balloon and relocation options The FRAP-T6 model consisted of a single fuel rod, were used. The no balloon option based cladding divided into 12 evenly spaced axial regions. Radially failure on effective cladding plastic strain exceeding D-3

instability strain. The relocation model used the Table D-1. Additional input for the FRACAS 11 fuel reh> cation and fuel thermal conduc- FRAP-T6 analysis tivity models with FRACAS 1.

Additional input variables are shown on Table D-1. Value Parameter

3. Boundary Conditions Peuei o.D., ft 0.03417 Power and thermal hydraulic boundary conditions Clad thickness, ft 2.1 x 10-3 are reqaired as input to FRAP-T6. All of these bound- Clad O.D., ft 0.0402 ary conditions were supplied by the RELAP5 analysis.

The specific thermal hydraulic conditions required Pellet height, ft 0.03417 were; heat transfer coefficient, system pressure, and llelium Fill gas coolant temperatures along the axial length of the rod.

These conditions were available to the FRAP-T6 Burnup, MWD /MTU 10G64 analysis from a data file at every RELAP5 time step.

Plenum spring data:

Since rod failure was predicted during the first power 90 Number of coils excursion, only the first excursion was analyzed. The 0.78 Spring height, ft emphasis of this analysis was on qualitative rather than Coil diameter, ft 3.7 x 10-3 quantitative results, i.e., did these excursions cause rod failure and roughly to what extent. The first excursion Rod height, ft 12.5 was also the lowest power excursion, and if this excursion would cause failure any of the subsequent excursions would. The power input used in FRAP-T6 is shown in Figure D 1.

l l 15 0 i , i l FRAP-T6 INPUT l l 9 b

U l r .

2 10 0 - -

u 8

e V

$ 50 - -

.=

l z

O 1200 1250 1300 1350 1400 Time (s)

Figure D.I. FRAP-T6 input power.

D-4

REFERENCES D- 1. L.1. Siciten et al., FRAP-T6: A Cornputer Code for the Transient Analysis of Oxide Fuel Rods, NUREG/CR-2148. EGG-2104, May 1981.

D-2. 1. A.

Dearien et al.,

FRAP-S: A Computer Codefor Steady State Analysis of Oxide Fuel Rods, Vol.1 - l FRAP-S2 Analytical Models and input Manual, SRD-S2-76, January 1976.

1 D-3. R.1. Grossenbacker et al., BWR/4 and BWR/S Fuel Design, NEDO-20944, Rev.1, October 1976. I

{

D-4. G. A. Berna et al., FRAPCON-2: A Computer Codefor the Ccdcidation ofSteady State lhennal-Mechanical ;

Behavior of Otide Fuel Rods, NUREG/CR-1845, Decernber 1980.

l l

l l

l l

l I

l l

l l

l D-5

.. - - _ - - -_ - - __- _ ___ _ _ _-__ a

ia..oa,av a... r,oc v.... w,

v. .ucua. u ut*, ,oo c..

NUREG/CR-4165

'E BIBLIOGRAPHIC DATA SHEET EGG-2379 o r.. .vi.n

u. ,,,si.ucv.o a a....L. .

l

, ,,, a . o .v. .. a l

Severe Accident Sequence Analysis Program - Anticipated . ..u n, oar cou,uno ,

Transient Without Scram Sirtulations For Browns Ferry woar* naa Nuclear Plant Unit 1 l May 1987

. avi.ousi ' "'" "*' av'a Wayne C. Jouse " "'" "'"

R. Jack Dallman Steve R. Wagoner l 1987 Richard C. Gottula May rdward F. Hnlcnmh Philin D Wheatlev e emoJgCT(TASE/WOma what seues tR i PE73PORWeNG OftMNt2at40s. maug amo maegioso aoomass #sasema la coes Idaho National Engineering Laboratory . ... oa o=*= r =wuna EG&G Idaho, Inc.

P.O. Box 1625 A6354 Idaho falls, ID 83415

- ' ge (s.es als TvPt Of aEPOmf 10 SPON50Meh0 0m0Aastdat aON NAWB AND MAsLle#G AoDA&M f:

Division of Reactor Accident Analysis Office of Nuclear Regulatory Research *""'"'""*"-a***-'

U.S. Nuclear Regulatory Commission Washington, D.C. 20555 t 2 SUPPLtwth1 ARv NOf 88 SJ A05f A4Cl iJGf ewWs er wt An analysis of five anticipated transients without scram ( ATWS) was conducted at the Idaho National Engineering Laboratory (INEL). The five detailed deterministic simula-tions of postulated ATWS sequences were initiated from a main steamline isolation valve (MSIV) closure. The subject of the analysis was the Browns Ferry Nuclear Plant Unit 1, a boiling water reactor (BWR) of the BWR/4 product line with a Mark I containment.

The simulations yielded insights to the possible consequences resulting from a MSIV closure ATWS. An evaluation of the effects of plant safety systems and operator actions on acci-dent progression and mitigation is presented.

.....,,,s.,,

.. oocu i i .. 03.s . . ..onos ouc.,,v os.. STATEWGNf UNLIMITED .

't SECUntivCLA88s8tCAf ton trag es,w

. o o n. UNCLASSIFIED

. ,oiar.. ..=m f F#4 esperrt i,flNC1 avunaAS o. ..oesSIFIED it PRIC1

\

i

-)

l

, ., n- s. , . c n o 7 1 1AN1"1 r31y ,6;F0 - I'nU:.c "- : ; 'l

- * ;, ', T 2 P'~

=F f boticy e. tU" 2:' u

, _ r. 3 7 _

c .

,3:s,,,T:t 1

,, I

]

1 1

1

{

l I

l l

EGEtG /daho P.O. Box 1625 Idaho Falls, Idaho 83415 L- _ - ---_- . .

.