ML20214T485
| ML20214T485 | |
| Person / Time | |
|---|---|
| Site: | Bellefonte  |
| Issue date: | 09/30/1986 |
| From: | Bieniarz P, Gasser R, Tills J SANDIA NATIONAL LABORATORIES |
| To: | NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| References | |
| CON-FIN-A-1258 NUREG-CR-4563, SAND86-0576, SAND86-576, NUDOCS 8612080603 | |
| Download: ML20214T485 (71) | |
Text
{{#Wiki_filter:. - NUREG/CR-4563 SAND 86-0576 RG Printed September 1986 Analysis of Station Blackout Accidents for the Bellefonte Prossurized Water Reactor R. D. Gasser, P. P. Bieniarz, J. L. Tills Sa a onal Laboratores for th Un ed S a e D r en of E under Contract DE-AC04 76DP00789 i l . t Prepared for U. S. NUCLEAR REGULATORY COMMISSION sr290cata an 8612090603 860930 8 PDR ADOCK 0500
s NOTICE This report was prepared as an account of work sponsored by an agency of the United States Covernment. Neither the United States Government norany agency thereof,or any of their employ-ees, makes any warranty, empressed or it splied, 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 process disclosed in this report, or represents that its use by such third party would not infringe privately owned rights. 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 l 4 I
0 NUREG/CR-4563 SAND 86-0576
'RG
- ANALYSIS OF STATION BLACKOUT ACCIDENTS FOR THE BELLEFONTE PRESSURIZED WATER REACTOR 4 R. D. Gasser P. P. Bienlagg* J. L. Tills Printed: September 1986 Sandia National Laboratories Albuquerque, New Mexico 87185 Operated by Sandia Corporation for the U. S. Department of Energy Prepared for-Division of Accident Evaluation Office of Nuclear Regulatory Research U. S. Nuclear Regulatory Commission Washington, DC 20555 - Under Memorandum of Understanding DOE 40-550-75 i NRC FIN No. A-1258 4 i Risk Management Associates, Albuquerque, NM Tills and Associates, Albuquerque, NM
I ABSTRACT An analysis has been performed for the Bellefonte PWR Unit 1 to determine the containment loading and the radiological releases into the environment from a station blackout accident. A number of issues have been addressed in this analysis which include the effects of direct heating on containment loading, and the effects of fission product heating and natural convection on releases from the primary system. The results indicate that direct heating which involves more than about 50% of the core can fail the Bellefonte containment, but natural convection in the RCS may lead to overheating and failure of the primary sys-tem piping before core slump, thus, eliminating or mitigating direct heating. Releases from the primary system are signifi-cantly increased before vessel breach due to natural circulation and after vessel breach due to reevolution of retained fission products by fission product heating of RCS structures. iii/iv
L l { \ CONTENTS vaae ABSTRACT................................................ iii ACKNOWLEDGEMENTS........................................ ix 1.0 EXECUTIVE
SUMMARY
................................. 1 1.1 Current Issues.............................. 1 1.2 Codes and Methodology....................... 3 1.3 Accident Sequences.......................... 3 1.4 Results and Conclusions..................... 4 1.4.1 Direct Heating...................... 4 1.4.2 Primary System Releases............. 5 1.4.3 Containment Releases................ 6 1.5 Modeling Assumptions and Limitations.. ..... 8 l r
2.0 INTRODUCTION
...................................... 10 3.0 CODES AND METHODOLOGY............................. 12 4.0 TMLB' WITH HIGH PRESSURE EJECTION................. 15 4.1 Containment Loading.....'.................... 16 4.2 Radiological Releases from High Pressure Ejection........................... 20 4.3 Primary System Releases..................... . 21 4.3.1 TMLB' (HPE) Without Natural Convection (Type A)................... 25 4.3.2 TMLB' (HPE) With Natural Convection (Type B)................... 27 4.3.3 TMLB' (HPE) With Late Fission Product Reevolution (Type C).......... 28 l 4.4 Containment Aerosol Transport and Releases.................................... 33 4.4.1 Single-Volume Calculations............ 33 4.4.2 Multi-Volume Calculations............. 36 5.0 TMLB' PUMP SEAL LOCA.............................. 38 5.1 Containment Loading......................... 39 5.2 Primary System Releases..................... 39 5.2.1 Pump Seal LOCA Analysis Without . Natural Convection (Type A)........... .40 ( 5.2.2 Pump Seal LOCA Analysis With
- Natural Convection (Type B)........... 41 l
5.3 Containment Aerosol Transport and Releases.................................... 43
6.0 CONCLUSION
S....................................... 45
7.0 REFERENCES
......................................... 47 v
I TABLES Pace 1 Sequence of Events for the TMLB'-HPE................ 49 2 Case Matrix for TMLB'-HPE........................... 50 3 Assumptions for Direct Heating Calculations......... 51 4 Results of Direct Heating Calculations.............. 52 5 Initial Fission Product Group Inventories........... 53 6 Direct Heating Fission Product Releases (BMI-2104).. 55 ) 7 Direct Heating Fission Product Releases (MCT2)...... 56 ; 8 Summary of Primary System Releases-TMLB'-HPE........ 57 9 Particle Size Distribution for RCS Releases......... 58 10 Summary of Containment Releases-TMLB'-HPE........... 59 11 Comparison of Multi-Cell Releases to Single-Cell Releases.................................... 60 12 Sequence of Events for the TMLB'-PSL................ 61 13 Case Matrix for TMLB'-PSL........................... 62 14 Summary of Primary System Releases-TMLB'-PSL........ 63 15 Summary of Containment Releases-TMLB'-PSL........... 64 O e f vi
I i FIGURES
- Pace 1 Diagram of Code Linkages............................. 65 2 Containment Pressure TMLB' Without Direct Heating (Case-000)........................................... 66 3 Containment Temperature TMLB' Without Direct Heating (Case-000)........................................... 67 4 Primary System Flow Path-TMLB'-High Pressure Ejection............................................. 68 5 . Description of Natural Convection Model.............. 69 6 Nusselt Number Correlations for Natural Convection... 70 7 Primary System Pressure-TMLB'-High Pressure Ejection. 71
- 8 Average Core Temperature-TMLB'-High Pressure Ejection............................................. 72 9 Core Melt and Clad Oxidation Fractions-TMLB" High Pressure Ejection.................................... 73 l 10 Hydrogen Generation Rate-TMLB' High Pressure l Ejection............................................. 74 j 11 Outlet Gas Flow-TMLB' High Pressure Ejection......... 75 12 Outlet Gas Temperature-TMLB' High Pressure Ejection.. 76
- 13 Primary System Gas Temperature-TMLB' High Pressure Ejection (Type A).................................... 77 14 Primary System Structure Temperatures-TMLB' High Pressure Ejection (Type A)........................... 78 15 Fission Product Heating-TMLB' High Pressure Ejection (Type A)............................................. 79 16 CsI Retention Factor-TMLB' High Pressure Ejection
- (Type A)............................................. 80 17 Te Retention Factor-TMLB' High Pressure Ejection (Type A)............................................. 81 18 CsI Mass Retained-TMLB' High Pressure Ejection (Type A)............................................. 82 19 Te Mass Retained-TMLB' High Pressure Ejection (Type A)............................................. 83 20 CsI Mass Released-TMLB' High Pressure Ejection (Type A)............................................. 84 21 Primary System Gas Temperatures-TMLB' High Pressure Ejection (Type B).................................... 85 22 Primary System Structure Temperatures-TMLB' High Pressure Ej ection (Type B) . . . . . . . . . . . . . . . . . . . . . . . . . . . 86 23 Fission Product Heating-TMLB' High Pressure Ejection (Type B)............................................. 97 24 CsI Retention Factor-TMLB' High Preusure Ejection-(Type 3)............................................. 88 25 CsI. Mass Retained-TMLB' High Pressure Ejection (Type B)............................................. 89 26 CsI Mass Released-TMLB' High Pressure Ejection (Type B)............................................. 90 27 Te Retention Factor-TMLB' High Pressure Ejection (Type B)............................................. 91 28 Te Mass Released-TMLB' High Pressure Ejection ,
(Type B)............................................. 92 vil _ __ _ _ _ _ . _ _ _ ~ _ _ _ _ _
Figures (Cont.) Page 29 Primary System Gas Temperature-TMLB' High Pressure Ejection (Type C).................................... 93 30 Primary System Structure Temperatures-TMLB' High Pressure Ejection (Type C)........................... 94 31 CsI Retention Factor-TMLB' High Pressure Ejection (Type C)............................................. 95 32 CsI Mass Retained-TMLB' High Pressure Ejection (Type C)............................................. 96 33 Airborne CsI-TMLB' High Pressure Ejection Case-0llB)........................................... 97 34 Leaked CsI-TMLB' High Pressure Ejection (Case-OllB).. 98 35 Leaked CsI-TMLB' High Pressure Ejection (Case-012A).. 99 36 Leaked Aerosol Size Distribution-TMLB' High Pressure Ejection............................................. 100 37 Aerosol Disposition-TMLB' High Pressure Ejection (Case-011B).......................................... 101 i 38 Aerosol Disposition-TMLB' High Pressure Ejection ! (Case-012A).......................................... 102 l 39 Diagram of Multi-Cell CONTAIN code Model............. 103 40 Containment Pressure-TMLB' Pump Seal LOCA (Case-030). 104 41 Containment Temperaure-TMLB' Pump Seal LOCA (Case-030)........................................... 105 L 42 Primary System Flow Path-TMLB' Pump Seal LOCA........ 106 43 Outlet Gas Flow-TMLB' Pump Seal LOCA................. 107 44 Primary System Structure Temperatures-TMLB' Pump Seal LOCA (Type A)................................... 108 45 Fission Product Heating-TMLB' Pump Seal LOCA , (Type A)............................................. 109 46 CsI Retention Factor-TMLB' Pump Seal LOCA (Type A)... 110
- 47 Te Retention Factor-TMLB' Pump Seal LOCA (Type A).... 111 l 48 Te Mass Released-TMLB' Pump Seal LOCA (Type A)....... 112 49 CsI Mass Released-TMLB' Pump Seal LOCA (Type A)...... 113 50 Primary System Gas Temperature-TMLB' Pump Seal LOCA (Type B)............................................. 114
- 51 Primary System Structure Temperatures-TMLB' Pump LOCA (Type B)........................................ 115 52 CsI Retention Factor-TMLB' Pump Seal LOCA (Type B)... 116 53 Te Retention Factor-TMLB' Pump Seal LOCA (Type B) . . . . 117 54 Te Mass Retained-TMLB' Pump Seal LOCA (Type B)....... 118 55 Airborne CsI-TMLB' Pump Seal LOCA (Case-030).........
- 119 56 Leaked CsI-TMLB' Pump Seal Loca (Case-030)........... 120 57 Airborne CsI-TMLB' Pump Seal LOCA (Cast-331)......... 121 ;
i 58 Airborne CsOH-TMLB' Pump Seal LOCA (Case-031)........ 122 i ) 59 Airborne CsI-TMLB' PUMP Seal LOCA (Case-131)......... 123 l i , viii i
l l ACKNOWLEDGEMENTS Special thanks must go to Dave Williams who offered much useful advice regarding the presentation of the analytical results and who also performed some of the direct containment heating cal-l culations using the DHEAT Code. We would also like to express our gratitude to Thomas Walker of the NRC Containment Systems Research Branch for his careful editing of this document and his cogent editorial comments. 1 i s l ix/x
- i 1.0 EXECUTIVE
SUMMARY
H An early attempt at an integrated "best estimate" of containment loading and radiological releases for specific accident sequen-4 ces was carried out at Battelle Columbus Laboratories and dis-l' seminated in the BMI-2104 report. In that study five reactors 1 were selected in order to characterize different major reactor ' types, and analyses - were performed for a few of the accident sequences that were thought to dominate the risk or present unique challenges to the containment.
~
The BMI-2104 analysis employed computer codes that were run independently and in large part were in an early stage of development. A number of issues have been raised in the, interim, some of which are related to phenomenology not modeled in the BMI-2104 suite of codes. It is ' 4 the objective of this study to address some of these issues with 4 a view to improving the calculated estimate of containment load-ing and fission product releases to the environment. The pres-ent study has been limited.to large, dry containments in general and to the TVA Bellefonte Unit 1 Pressurized Water Reactor (PWR) in particular. Since it is currently thought _that transient accident sequences
- tend to dominate the risk for large, dry PWRs in general, this
! analysis has been concentrated on two variations of the TMLB'
; accident sequences. The TMLB' sequence consists of the situa-4 tion in which there has been a complete loss of offsite and
- onsite AC power together with a loss of auxiliary feedwater and 1 a failure to recover power before a core-damaged state has 1 evolved. The two variations involve,-in one case, an intact primary system boundary until the core slumps and melts through j the lower vessel head, and in the other case, an induced LOCA due to the failure of primary coolant pump seals.
, It is the specific purpose of this analysis to estimate the i primary system fission product transport and releases into con-tainment, the containment pressure / temperature response, and the l i transport of fission products through the containment and ulti-mate radiological releases to the environment. The~ calculations have been performed for a number of cases and configurations in 4 which several important issues have been specifically addressed. These issues are discussed below. 1.1 CURRENT ISSUES An issue of considerable importance.for PWRs, since it supplies a mechanism for early containment failure, is that of direct containment heating due to the high pressure ejection of molten core debris into the upper containment atmosphere. The ejected , core debris, concisting of molten-fuel, cladding, and structural i materials, can form a dispersed aerosol capable of rapidly transferring its heat to the containment atmosphere. Because the calculated loads on the containment structure due to direct heating may challenge its integrity. simultaneously with near
l maximum concentrations of radiological. sources in the contain- , ment atmosphere, this scenario has the potential for large releases and severe effects on public health. There are, however,, a number of uncertainties attached to the direct heating . question .which range from the magnitude of its l effect to whether it will, in fact, occur. Large uncertainties
- in the core meltdown phase of the accident, for example, yield a wide range of possible initial conditions for core debris at the time of vessel breach. Much of this uncertainty may be removed when the more advanced melt-progression codes such as MELPROG become available. In view of the current lack of knowledge ;
regarding direct heating phenomenology, the present study has j taken a parametric approach to direct heating. The BMI-2104 calculations for primary system fission product transport employed a set of codes (MARCH, CORSOR, MERGE, and TRAP /M r ' T) which were run successively (without feedback) and
- which . id not contain models for fission product-heatup of the ;
} primary system, nor did they contain models for treatment of ! l natural convective processes. Transport of volatile fission
- products through the primary system and releases of these i
materials into the containment are extremely sensitive to the temperature history of the primary system, so that these two i processes should not be neglected. As a consequence of.the-l absence of models for the aforementioned p ocesses, the BMI-2104 i calculations for the TMLB' sequence may underestimate primary- ! system releases during the meltdown phase of the accident. Fur- ! ther, in that study it was not possible to calculate long term ! reevolution (revaporization) subsequent to vessel failure of I those volatile fission products that had been retained on ! primary system surfaces during the meltdown phase. Certain ! containment conditions, such as a gradual depressurization due I to leakage over the time period in which the volatiles are being reevolved, can supply a mechanism for moving these sources out of the primary system and ultimately into the environment. In the present study an attempt has been made to treat fission product heating and natural convection in the primary system heat and mass transport calculations. An additional-benefit of this modeling capability has been a somewhat improved charac-terization of the primary system temperature history, and as a result, the ability to gain some insight into another important issue: that of potential temperature induced failure of the hot-leg piping. Enhanced heat transfer due to natural.circula-tion processes could generate high enough. temperatures at the hot-leg nozzles or in downstream piping to weaken the structure
- and cause a breach. This may be particularly likely for the TMLB' station blackout sequence due to the high system pressure j (2000-2400 psia). If primary system failure in the hot-leg
! occurred before core slump, direct heating might be effectively j precluded. The magnitude and. characteristics of the counter-l current flow regime in the pipes leading from the vessel to the
-_ . _ _ _ ._m
pressurizer and steam generators have not been established, and the present. analysis probably yields heat transfer rates that are somewhat high. Multidimensional codes may be required to achieve a definitive answer regarding natural convection in the complicated geometry of the primary system. I 1 i- 1.2 CODES AND METHODOLOGX The primary strategy in this analysis was to use the best calculational . tools available. to treat each phase of the , accident. The RELAP5. code was used to calculate the thermo-
- hydraulic sources during the pre-core-damage phase, and the MARCON code was employed during the period from incipient core meltdown out to and including the core / concrete interaction I phase (the CORCON code which treats the core / concrete inter-action has been combined with the MARCH code in the MARCON code). A interactive version of the CORSOR, MERGE, and TRAP /
MELT code package was utilized to calculate primary system j fission product releases. This code, called the MCT code, contains new models for fission product heating and natural i circulation. It must be noted that these models were developed by P. P. Bieniarz and have not been extensively evaluated. The , i natural circulation flows .are one-dimensional and driven-by temperature differences between connecting volumes. Radiolog-ical releases from core / concrete interactions were calculated using the VANESA code. The thermohydraulic sources.due to direct heating were calculated using a small stand-alone code i backed up by hand calculations. Both the code and the hand ! calculations used an adiabatic model in which the core debris ! was brought into thermal equilibrium with the atmosphere. The l aerosol sources for direct heating were obtained from results of ! the SPITS experiments conducted at Sandia National Laborato-ries. Heatup and degassing of the concrete floors in the con- , tainment after deposition of the debris from direct heating were ) also calculated using a stand-alone code. The various sources, both thermohydraulic and radiological, were supplied to the
- CONTAIN code which calculated the containment loading, the 4 transport and deposition of fission products in containment, and 1
the release of radiological sources to the environment. 1.3 ACCIDENT SEQUENCES Two basic scenarios were examined in the study. The first scenario consisted of a TMLB' sequence in which temperature , induced failure of the primary system did not occur prior to core slump. The primary system eventually failed at instrumen-tation penetrations in the lower vessel head due to the thermal attack of molten core debris. Since a previcus breach in the primary system boundary did not occur (except for normal opera-tion of. the pressure operated relief valves, PORVs), the system
- was at high pressure (PORV set point, about 2400 psia) and the l potential existed for direct heating. For this scenario a
~ -- - ~ . - - . - _ .. .. - .. ..
) matrix, of cases- was analyzed in which selected direct heating parameters were varied. The primary' system ~ radiological releases were also varied by using MCT code predicted releases and selected combinations of the models in that. code.(i.e.
- fission . product heating, natural circulation, .reevolution of volatiles). Some cases were also analyzed using the _RCS releases calculated in the BMI-2104 study of the Zion plant.
The Zion releases were adjusted for differences in~ core power between the Zion and Bellefonte plants. The second scenario assumed.a TMLB' with failure of-the seals in , the primary system coolant pumps (pump seal. loss of coolant j accident, LOCA). . In this scenario all four pump seals,were l assumed to fail at the outset of the accident, so that the
; system was at a somewhat lower pressure at the time of vessel
{ breach. Thus, the direct heating scenario, if it occurred, would be .of a smaller magnitude. For this analysis it was +- assumed that direct heating did not occur in the pump' seal LOCA scenario, 1.4 RESULTS AND CONCLUSIONS f 1 1.4.1 DIRECT HEATING
! For the direct heating calculations a number of parameters.were ; varied including the fraction of core injected into the contain-i ment, the quantities of hydrogen and molten steel burned during l the event (all the injected zirconium was assumed to be oxi-
! dized), and the amount of reactor cavity water in"olved in the process. i. i For the case in which 90% of the core debris was involved, the peak containment pressures were in excess of the estimated failure pressure range for the Bellefonte containment (144.7 to ~ 153.7 psia). With 50% core debris involvement the peak pres-sures were marginally close to the estimated containment failure criterion. Without the involvement of' hydrogen and steel oxida-tion, pressures were somewhat below the failure criterion while j inclusion of hydrogen and steel oxidation resulted-in pressures ! at or slightly above the failure criterion. . However, for the more probable scenarios in which cavity water .was involved j
- (primarily additional effective heat capacity), complete j oxidation of the steel and the hydrogen (from the in-vessel
! oxidation of the cladding) was required te-yield a peak pressure high enough to challence the containment. i l The results of the direct heating calculations were used to postulate the containment failure mode for the radiological' source calculations. The 90% injection case was assumed to fail containment (a 7 ft2 hole) at the time of direct heating. The i 50% injection case, since it was marginally close to-failure, was assumed to stress the containment sufficiently to induce.a j relatively small leakage pathway. ; l l i i 1 - -.
i 1.4.2 PRIMARY SYSTEM RADIOLOGICAL SOURCES Three release " types" have been calculated in the present study. The release types correspond to MCT code calculations that were performed exercising combinations of the two major modeling
- changes that were added to the code. Tvoe "A" releases are the RCS (Reactor Coolant System) releases obtained for the meltdown phase of .the accident employing only the fission product heatup model. Releases from the RCS during the meltdown phase which utilize both the fission product heatup and the natural circu-
) lation models are referred to as tvoe "B" releases. Finally, ! releases from the RCS during the reevolution phase, that is the period subsequent to vessel breach, are calculated using only the MCT fission product heating model and are referred to as tvoe "C" releases. I The primary system releases for Bellefonte as calculated by the i MCT code package are presented here in juxtaposition with the BMI-2104 Zion plant calculations. As discussed below the MCT i type "A" calculations for Bellefonte produce releases that are j comparable to those calculated for the Zion plant. However,- the ! BMI-2104 study also analyzed the Surry and Sequoyah reactors and for those plants calculated relatively higher releases when com-1 pared to Zion results. The reason for the differences in the releases between the Zion, Surry and Sequoyah plants in the i BMI-2104 study is not clear. There may be sensitivities in these calculations to different modeling assumptions in the code package or to actual differences in the primary system configurations. In any case, the comparison between the present
- results and the BMI-2104 Zion calculations should not be interpreted as being particularly significant. Rather, the selection of the Zion results for purposes of comparison should be viewed as an arbitrary one, and the Zion releases merely a
" data base" against which the present results were compared.
l The important conclusion regarding primary system releases is to i be derived from comparisons between the three types of releases (A, B and C in Tables 8 and 14). For the high pressure cases using only the fission product heating model the meltdown phase releases frem the primary j system (up to and including the releases at vessel failure, ; Table 8, type 'A' releases) of CsI and CsOH were slightly higher (by a factor of 1.5) than those predicted by BMI-2104 for the l Zion plant. The Te releases were higher by a factor of 6, ' partly due to differences in the MARCH code options that l affected the oxidation of zirconium in the meltdown phase and 1 resulted in higher Te releases from the core. Although the ! fission product heatup model did not produce significant changes in predicted releases during the meltdown phase, the long term heatup effects subsequent to vessel failure produced significant late reevolution of volatile fission products (Table 8, type 'C' releases). For the leakage cases involving a gradual con-tainment depressurization, a mechanism existed for moving
kl t i-I revolatilized fission products out of the breached vessel, and ! with .these additional releases included, the total releases of- ! CsI. and CsOH from the RCS exceeded.the BMI-2104 estimates by a j factor _of 10. Implementation of the natural convection model *
; for. the meltdown phase releases -(without reevolution, type 'B' ,
j releases) produced RCS releases higher than BMI-2104 Zion ' ! releases by a factor of 15 for CsI and CsOH and a factor of 4.5 for Te (approx. 30% for CsI and CsOH and 17%-for Te of the
- original inventories of these species). .Although performance of the- rather expensive computer calculations for long. term reevo-i 'lution using the' natural circulation models after vessel breach
- was beyond the scope of this analysis, the releases could be still higher than those quoted above.
There were no comparable Zion BMI-2104 calculations against i which to compare the pump seal lDOCA calculated' releases. l Without natural convection the primary system volatile releases j for this scenario were on the . order of 10% of the original I inventories (Table 14, type 'A' releases). With natural I convection the predicted RCS releases of CsI and CsOH were very high, on 'the order of 80% or 90% and Te was about 13% (type 'B' i releases). However, since direct heating was not involved in this sequence, there was no early containment failure and very i small containment releases. Also of significance was the predicted primary system tempera-l ture history. Hot-leg piping temperatures on the order of 3 17000 F were calculated prior to core slump:when the natural circulation model was used. It should be emphasized that the natural circulation calculations employed a rather simplistic one-dimensional model which represents a first cut at qualita-tively accounting for this-phenomenon. The complicated geometry
- of the primary system may not be amenable-to so simplistic a
] treatment, and it seems likely that the primary system tempera-i tures predicted by this model are somewhat high. However, the
- calculated temperature at the outlet nozzle would be within the regime in which failure could be induced, even if the actual temperatures were several hundred degrees lower.. If the hot-leg
- piping failed in this mode soon enough to depressurize the pri-4 mary system prior to vessel breach, the direct heating scenario
! could be precluded. 1 l.4.3 QQN_TALNENT RELEASES
- A general description of the containmer.t release calculations is i
given in Tables 2 and 13 and the releases for these cases are , summarized in Tables 10 and 15. Radiological releases from the containment are heavily dependent on the containment failure ! mode. Direct heating calculations indicate that for core debris ! involvement greater than about 50%, gross containment failure is likely in the Bellefonte containment. The 90% injection direct j heating casgs were, therefore, assumed to rupture the contain-l ment (7 ft hole), and the releases of the volatiles, CsI and-i
CsOH were an the order of 1.5% of the initial inventories (Table 10, Cases-001 to 011A). Since the primary system releases of these materials were a factor of 10 higher when natural convec-tion was modeled, the releases of these materials from the con-tainment were also proportionally higher for this case, about 15% (Case-011B). Releases of Te were on the order of 20% to 30% while Ru was about 40%. Releases for the latter two fission product groups (Te, Ru) were strongly enhanced by an oxidation reaction that was assumed to occur during direct heating. Spe-cifically, the species Te, Ru, and Mo were assumed to form vola-tile oxides during direct heating which were condensed as aero-sols. The oxidation fraction (the fraction of the species that was converted to the oxide) applied to these species were taken from the Reactor Safety Study (WASH-1400). The WASH-1400 oxi-dation fractions were estimated and did not represent actual experimentally determined values. Containment failure at the precise point of maximum fiEJion product Concentration made these cases the most severe in terms of releases to the environ-ment. For the 50% direct heating cases (Table 10, Cases 002 to 112A), it was assumed that containment leakagg pathways gpen at the time of vessel failure. Areas of 5 in and 12 in gere used to estimate releases for these scenarios. The 5 in leakage area was obtained from the Containment Performance Working Groups estimate of pressure induced leak area for the Zion plant. Estimated leakage characteristics are not presently available for the Bellefonte plant. For both of the leak areas used, the containment was gradually depressurized and, since much more time was available for aerosol removal, the releases were significantly lower. which reevolution of volatiles For was the worst case (Case-ll2A), ig calculated and a 12 in leak area was used, releases of the volatiles and Ru were on the order of 2% of inventory. These releases were about a factor of 12 higher than those calculated by assuming BMI-2104 Zion releases frcn the primary system (Case-102A). The assumed, mechanism responsible for driving reevolved fission products out of the RCS was the depressurization of the containment. For the TMLB' scenarios without direct heating (Cases 000,100) thg induced leakage area was varied between 1 in 2 and 2.4 in , and the releases for these cases were very small, on the order of 1.0E-4 of the original inventories of CsI and CsOH. The pump seal LOCA scenarios, since they are assumed not to involve direct heating, also yielded relatively small releases, on the order of 1.0E-4 for CsI and CsOH (Table 15, Cases 030,130). Prinary system releases calculated with natural convection, however, increased the containment releases to about 0.2% for CsI and CsOH and to about 0.8% for Te (Table 15, Cases 031,131). There are several conclusions that can be drawn from this study. One of the most significant is that, if direct heating
occurs in which. more -than about 50% of the core debr'is is'
~
involved, containment failure is likely. _In addition, this scenario fails. the containment.at a time when the fission prod-uct concentration in the containment atmosphere is maximized and leads Eto-.high releases outside the containment. Introducing simple modeling for natural. convection and. decay heating in the primary _ system fission product. transport codes has significantly increased the predicted releases into the containment due to the . added releases from reevolved volatiles and to natural convec-l tion heating of the primary system. However, the same modeling changes have lead to an indication-that primary system struc-i tural temperatures. may be. high enough to induce failure =and depressurization of the primary system before core slump, thus
- eliminating direct heating and with it the primary mechanism for-i early containment failure for this accident sequence.
1.4.4 MODELING ASSUMPTIONS AND LIMITATIONS l There are a number of modeling assumptions and limitations inherent in the analysis. Some of the limitations are related to the capabilities of the code package and others are related l to engineering assumptions necessary to perform.the analysis i within budget. The major assumptions will merely be identified L here while the more significant ones-will be discussed in some l detail in the appropriate sections of the report. Due to the prohibitive expense associated with multi-volume i ! containment code calculations, the bulk of the present analysis , i was performed using a single volume to model the containment. j one multi-volume calculation was . performed Eto baseline the , single volume analysis. The new models that have_been supplied j to the primary system fission' product transport code package i have not been rigorously evaluated. The primary system natural
- convection model has two limitations that'have been identified.
l- The model appears to somewhat overpredict heat transfer between l control volumes, and the heat that is transferred from the core j by natural convection does not cool the core. To remove the latter limitation it would be necessary to link the meltdown
- progression code with the RCS fission product transport code-i package. Parallel or multiple flow: pathways have not been l modeled in the MCT code calculations. The radioactive. decay of
[ fission product species to their daughter species has not-been treated. Due to a lack of phenomenological models for direct
- heating at the inception of the analysis, it was necessary to J
employ a number of limiting assumptions and a highly parameter-ized approach to the direct heating calculations. The decompo-sition of CsI due to high temperature and a high radiation field has not been treated. Lumped parameter models are employed in the MCT code package to estimate heat structure thermal j response. Lumped parameter models can significantly underpre-= i dict structure surface temperatures and, thus, overpredict i fission product retention on RCS surfaces. Finally, the assump-tion has been made that core / concrete interactions do not l 8-
commence until all the water has been vaporized in the reactor cavity, i r i
-9
2.0 INTRODUCTION
The Severe Accident Sequence Analysis program at Sandia National Laboratories (SNL SASA) has as part of its objectives the task of performing detailed best-estimate analysis of reactor systems behavior during severe accidents. The analysis presented here is a study of the TMLB' (WASH-1400 nomenclature) station black-out accident sequence for the TVA Bellefonte (PWR) reactor. The TMLB' accident sequence involves a complete loss of offsite and onsite AC power together with a loss of auxiliary feedwater and a failure to recover power before core damage is sustained. This report addresses in some detail two major scenarios of the TMLB' type accident sequence; namely, the low pressure and high pressure primary system failure scenarios. The hich oressure scenario occurs when the primary coolant system boundary remains intact until vessel melt-tnrough. Breach of the RCS boundary due to temperature-induced pump seal failure or piping failure which results in complete or partial depressurization of the RCS before vessel melt-through results in a low pressure TMLB' scenario. A probabilistic risk assessments is not currently available for B&W p gpts with the Bellefonte configuration. It is thought 1, however, that the TMLB' sequence represents a dominant sequence for large dry PWR's in general. Based on that observation, together with the fact that many of the scenarios developing out of external events closely resemble the TMLB' accident sequence, the present analysis has been concentrated in that area. To approach as closely as possible a best-estimate analysis, an i attempt has been made to incorporate the most recent thinking in ' terms of the type of phenomenology deemed to be important and appropriate for the sequences under examination. In particular, for the present study, this includes calculating the effects of direct heating of the containment atmosphere due to the high pressure ejection of core material at f v****1 breach. In addition,recentinformationyge'3'kf*e*mergingfrom Sandia's experimental programs currently being conducted to assess the effects of high pressure ejection have also been utilized in this study. This information includes estimates of the fraction of the ejccted debriu that is injected into the containment atmosphere, the fraction of the injected debris that remains in the atmosphere as aerosols, and the particle size distribution of the aerocols generated by direct heating. The primary thrust of the Bellefonte analysis is to determine both the thermal-hydraulic loadings en the containment system and radiological releases to the environment. The methodology employed in the analysis will be discussed in detail in Section
- 3. In general the most recent versions of the containment phenomenological codes were employed where appropriate, while
4 i stand-alone codes or hand calculations were used for phenome-nology not contained in the major codes. Examples of the latter , include the direct heating calculations and the calculation for the dehydration of the containment concrete floor subsequent to the deposit on it of core debris from direct heating. Large uncertainties continue to limit the degree of accuracy
- ' that can be obtained in terms of arriving at best-estimate num-bars for containment loading. It is generally agreed that in i the event of vessel failure at elevated pressure, direct heating i may pose a threat to containment integrity at precisely the time l that radiological are exacerbated by the same pheno-l menon. Experiments sou{gys have indicated that large fractions of j molten core debris may be ejected out of the cavity at high
- velocity. Intervening structures and surfaces may not offer sufficient holdup to prevent most of this material from being i injected into the upper containment atmosphere where it may j deposit both its sensible energy and the chemical energy that will be evolved upon interaction with the atmo1phere. The
, experiments, it must be noted, have been performed using molten
- ejecta. It is by no means certain that all or most of the core
- debris will be molten at the time.of vessel breach. One of the 1
sources of uncertainty relates to the in-vessel meltdown pro-gression. It is not known with any certainty what will be the ] composition and state of the core debris at the time of vessel . breach. Large chunks of core debris, for example, would cer-tainly affect the dynamics of the high pressure ejection sca-i nario, and perhaps inhibit it to some degree. It is hoped that
- the meltdown progression codes currently-under development (i.e.
i MELPROG) will help reduce the uncertainty and better quantify j the physical state of the core debris. i j Similarly, the timing and location of primary system failure is j uncertain. It has been postulated that, for the TMLB' sequence, the transport and deposition of fission products along the flow l path leading to the relief valves, together with energy trans-port by natural convective processes, will heat these structures
- and perhaps result in primary system failure at a higher eleva-j tion and at a time prior to lower vessel head meltthrough. The result would be that the primary system pressure may be relieved j in advance of vessel melt-through, thus eliminating the threat of high p ejection. A more recent version of the TRAP /
l MELT codetggssure I which models both fission product heatup and natural convection has been used in this analysis and predicts
- significantly higher primary system temperatures than have heretofore been reported. With a better estimate of primary system temperatures it should be feasible to perform creep / fail-ure calculations to estimate the time-to-failure for hot primary system structures at pressure.
i Since there exists considerable uncertainty in the parameters that affect the high pressure ejection phenomenon, and argu-ments can be given to justify core debris ejections ranging
~
4 anywhere 'from no ejection to _nearly- complete ejection, the l
- ' fraction of the core participating in direct heating has been j treated parametrically. The range of direct heating participa-tion has been covered by including cases with no injection, 50%
!= injection, and 90%. injection._ The term " injection" as apposed i to ~" ejection" .is used_. here to differentiate between ejection from the vessel and injection into the containment atmosphere. The latter 90 experiment gagy,us%
- injection, ing molten ismaterials. the_ case suggested by noted, It should be recent however, that .the experiments were designed to model the. flow paths and . obstructions in the Zion plant reactor cavity region and do not necessarily characterize the Bellefonte configura-i-
tion. In the present study it has been assumed that the ejected core debris is entirely molten at the time of vessel failure and . that failure occurs due to the thermal attack of the molten j debris on the lower reactor vessel head. , 4 Radiological releases from containment have been estimated based on two The primary system ! releases sources from for primary systqg) the BMI-2104 L releases. Zion plant analysis were L { adjusted up on the basis of reactor fuel inventory and used for 1 the primary system releases. As stated in the executive summary, the releases for the Surry or Sequoyah plants could also have i been used. The Zion releases were.used becsuse they were com-
; parable to the Type "A" releases (Section 4.3) and were con- ! venient for purposes of comparison. In addition a_ composite 1
(hard-linked) version of-the CORSOR, MERGE and TRAP / MELT codes, called the MCT code, was also employed to calculate the primary 1 system fission product releases. j , The overall methodology employed in-this study together with a ,
! discussion of the codes used and the method of linkage is de- l tailed in Section 3.
~
A discussion of the TMLB' high pressure' 3 failure scenario and results are given in Section 4., while the
- low pressure case (TMLB' with induced pump seal LOCA) is treated j in Section 5. The conclusions are given in Section 6.
f i 3.0 CODES AND METHODOLOGY I
- The treatment of phenomena that are not modeled in any of the
- i. currently available codes has somewhat complicated the present
- analysis. Mechanistic- models for direct heating and concrete degassing due to -fission product heating are. not presently _
. available in containment codes, for example. Small stand-alone ! codes together with hand calculations have been used in conjunc- { tion with a suite of codes which treat various aspects of severe j accident phenomenology. This section will discuss the phenome-j nology and the analytical tools employed. i ! A diagram outlining the flow of information-between codes.for i the high pressure ejection scenario is given in. Figure 1. The
- early phase of the reactor blowdown was calculated with the
{ RELAPS code and this calculation was run to_ incipient core k [ i l
3 degradation when the RELAP calculations became suspect. The
- criterion for terminating the RELAP calculations was a 1000 K 0 temperature in any 1RELAP core node. The RELAP5 analyses for both the TMLB' high pressure and pump seal LOCA case w performed. and the results supplied .to Sandia by EG&G 7'gye .
i The- thermohydraulic sources thus obtained were input to the f MARCON code as source tables. Figure 1 shows a direct link
- between RELAP and CONTAIN, but in fact, the effective linkage is between MARCON and CONTAIN. The MARCON code was modified.to 1
write an output tape that'.contains all the.thermohydraulic sources from the primary system including those from RELAP. The phase of the accident that commenced with the start of core degradation and, for some i i concrete interactions, has been situationsincluded.ex-vesselcoyg{ analyzed using the MARCONt computer code. The MARCON code consists of a hard link of the MARCH 2.0 containment code and the CORCON/ MOD 2 core / concrete ' interaction code. As indicated in Figure 1 linkages have been established between the MARCON code-and both the CONTAIN and the
- MCT codes. MARCON' supplies primary system conditions and j thermohydraulic sources to the MCT code.
The primary system radiological s9gpces are calculated with the MCT (MERGE,CORSOR, TRAP / MELT) codei 1 Information regarding i the core and primary system conditions were supplied to.the MCT code which calculated fission product releases from the degraded core, transport of fission products through the system, plateout ! and settling of fission products within the system, and esti-i mated fission product releases into the containment building. The MCT component codes, especially TRAP / MELT, have been modi- . fled and improved to- treat phenomenology not modeled in the l original codes. These include the heatup of primary system structures due to the presence of fission products and due to heat transported from the core region by natural circulation. The code numerics have also been improved as have the linkage interfaces. i The MCT code has been supplied with an output / input linkage with the CONTAIN code which is similar to the MARCON/CONTAIN inter-4 face. This interface passes information regarding fission-product species source rates (including vapors and aerosols) j into the containment atmosphere. ,
- Currently available containment codes do not presently contain mode}g0) for direct heating phenomenology. A small computer
! code . that calculates end conditions after direct heating l was used in this analysis to determine the peak temperature and ' pressure in the containment as well as the energy delivered to the atmosphere. The code performed an adiabatic calculation and i assumed that the ejected debris came into thermal equilibrium with the atmosphere. Chemical energy from the oxidation of zirconium, steel, and hydrogen was also accounted for in the code. Hand calculations were performed for several of'the
< i,<' .
1 4 1 r direct heating cases and good- agreement .with the c code was , l observed. . .The. net energy delivered to the. atmosphere as calcu-
"l lated by the direct heating code was. input into the CONTAIN-code as a mass. source with an artificial specific enthalpy which .
!' yielded the- cerrect integrated energy source.- The core debris which .was injected into the. containment during high pressure } , vessel failure was not available on the reactor cavity floor to 6 participate in a core / concrete interaction. It was, however,_
- distributed on the containment floor in a rather shallow (per-haps 1/2 to 2 inches, if evenly. distributed) debris bed. The initial temperature of the debris was the temperature at~which,' <
- thermal equilibrium between the debris and the atmosphere was '
j
- calcy{gged to occur. A small finite difference computer i code was employed .and modified to ' calculate transient' thermal conduction in the containment concrete floor and to'
. estimate .the quantity of steam evolved from concrete dehydra-tion due to the layer of hot core debris depositod in ) containment by the high pressure ejection; process. The degassing
' l model Igps similar to the one used in the PWR . version of '
MARCON except that the free and bound water were, assumed to , come out of .the concrete over two separate tamperature ranges. ; l< The free water was evolved over' a 20 K tenperature range ! centered at 3700 K and the bound water emerges over a 20 K l range centered at 510 0K. Although the debris layer on the containment floor, even for 90% injection was not sufficiently i deep to melt the underlying concrete, it was found that i considerable dehydration of the concrete ; occurred in thi's , i situation. In fact, the rate of heat transfer from the debring I
- to the atmosphere (ranging from 800 MW at the time of DCH tob l about 6 MW 25 minutes after DCH for the 90% injection case) was' l
! dominated by the flow of steam' out of the concrete surface f } rather than by radiation. The steam sources due/to concrete i dehydration, as obtained from thase calculations, were irput to i the CONTAIN code in the form of source tables. " ' ~ It has already been mentioned that the CORCON code, which is included as a subroutine of the MARCON code,chas bee.n utiliz'ad-to calculate the thermohydraulic sources from the<co're/ concrete interaction. The CORCON code was also used sto supriy the required information to the VANESA code to estimate.the radio-logica p 12,eleases from core / concrete interactions. The'VANESA code was designed to detail the chemical reactionsiin the molten pool and to estimate the releases of vaporstand aerosols from- the surface of the molten pool into the containment atmos-phere. The releases, both thermohydraulic andfradiological, as calculated by CORCON/VANESA were used :.as source terms to the CONTAIN code. Containment conditions, including temperature and pressure, and the transport of gases and aerosol into and out ofsthe contain s ment atmosphere have been calculated for this analysis using the. !. CONTAIN code. A detailed description of the CONTAIN code ~can;be! l found in Reference (13). The thermohydraulic and radiological l . > s
) i,p f
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . . _ _ - _ - - __ _ . - _ . _ - _ . _ _ _ _ _ -. i _ i. - -_
5 l l sources obtained from'the other codes used in the analysis were all input to CONTAIN which then calculated the response and estimated the aerosol concentrations in the containment atmo-sphere as well as radiological releases to the environment. 4.0 TMLB' with Hich Pressure Eiection J The TMLB' accident sequence involves a loss of both offsite and onsite power together.with a loss of auxiliary feedwater. Table 1 shows the sequence of events that occurred'in the high pres-
, sure TMLB' case- that was studied in this analysis. The steam generator dried out at about 5.8 minutes and incipient core melting occurred at 53.4 minutes. The lower reactor vessel head was calculated to fail at about 82 minutes at which time direct j heating took place. Containment failure or leakage was assumed to occur concurrently with direct heating .
Table 2 gives a description of the case matrix for the TMLB' high pressure ejection sequences. The primary parameter in this matrix is the fraction of core debris injected into the upper containment during direct heating. This fraction was assigned values of 0, 0.5, and 0.9. The second parameter indicated in Table 2 is the participation of cavity water in the direct heat-l ing event. The MARCH code predicts that about 75,000 lbs of coolant will be residing in the cavity at the time of vessel i breach. If this water is present in that location it will l almost certainly be swept out with or ahead of the core debris
; during core ejection. To determine the effects of cavity water, cases were run both with and without_the involvement of cavity water. For the cases in which cavity water was involved, it was assumed that the injected debris and the water were in good l thermal contact so that the water was completely vaporized. In j a real direct heating situation it is not clear what fraction of i
the debris and the water would be in intimate contact. In order to keep containment failure thenumbero{gpsestractable, criterion the estimated was employed as a bifur-cation point. The direct heating calculations were used to j estimate the peak pressure- and this, compared to the ultimate loading estimates, determined which branch the analysis followed , for each case. Thus, if the direct heating calculation seemed I to indicate a high likelihood of exceeding the ultimate loading capacity of the containment structure, the containment was I assumed to fail by opening a large hole in the boundary at a time coincident with direct- heating. If the pressure did not equal or exceed the failure criterion, a leakage commensurate with the magnitude of the peak pressure was employed. For those i cases in which the direct heating loads failed con $ainment (all' the 90% injection sequences) a break area of 7 ft' was assumed to open in the containment boundary at the time of direct heat-ing. For those cases which were marginally close to the esti-mated ultimate loading capacity of the containment (all the 50% ejection sequences) a leakage pathway was assumed to open coin-cident with direct heating. The leak area for these cases was
.y # J trgated parametrically, two values being used,. 5'in2 and 12 in The "best estimate" leakage of 5 in2 was obtained from the estimated leak area suggested in
. Working Group final report for the theContainment({gpformance Zion plant. 1 This value wasselectedintheabsenceo{correspondingestimatesfor-the Bellefonte plant. A 12 in leak area was;also used to
- determine the sensitivity of-containment releases to thg assumed ~ r
' leakage parameter. It should be noted that'the 5 in leakage'
- area estimated for the Zion containment was obtalned assuming a'
- -gradual pressurization 'of. the containment.
- There are no esti- i 1
mates for containment ' leak areas induced by rapid: transient , j events such as direct heating, steam spikes, or' hydrogen burns,. 3 ) The remaining parameters in Table 2'are related to the treatment 4 of fission product releases and transport. The primary system releases were estimated in two different. ways. The first method involved using' the primary system releases calculated in the BMI-2104 study for the same accident sequence in the Zion plant. These releases were merely 4 adjusted up according to the oper-ating power of> the two plants. The second method involved 1 actually running the suite of codes that Battelle used to calcu-late primary system. releases. An interactively run version of- l these codes, called the MCT code (for HERGE, CORSOR, TRAP / MELT) was employed for that purpose. Combinations of the modeling i
! capabilities in MCT vare treated parametrically as indicated in- !
4 Table 2. Finally, for two cases (Cases-012A, and 112A) a core / l 1 concrete interaction occurred late in the sequence (at about 20 j hours) and the VANESA code was used to estimate the radiological
- releases into containment due to this interaction.
4.1 Containment loadinc~ i
- For the accident sequences that include high pressure ejection j of core debris, direct heating imposes more severe loads on the '
j containment than occur at any_other time during the course of
~
the accident. A detailed examination of direct heating, there-fore, will yield the peak loading conditions. i Uncertainties in many aspects 9 of the direct heating phenomenol-i ogy required that it be treated parametrically. . Table 3 out- ! lines the basic assumptions for the direct' heating calculations.
- These assumptions- include ejection from the vessel of 100% of
! the fuel' and cladding materials and about 50% (100,000 lb) of l the potentially available ; steel. This quantity >of steel is-l about twice the amount of steel available in the core region and , accounts for partial melting of steel in the core support struc-
.tures and lower' plenum. .The initial temperature of the debris ,
d t as it exits from the vessel was'taken to be the corium eutectic temperature (approx. 2550 K). , , The quantity of Zircaloy reacted in-vessel-during:the core. melt-down phase was forced to be 50% by adjusting _the core slump
parameters in the MARCH code. Selection of a 50% metal / water reaction reflects the uncertainty in the extent of in-vessel clad reaction. This fraction of in-vessel oxidation was achieved by slumping the core when 65% of the core was molten. The hydrogen generated by in-vessel oxidation of zircaloy resulted in an average hydrogen concentration in containment at i vessel failure of about 3 volume percent. The fraction of the remaining zircaloy that was injected into the containment was i assumed to be completely oxidized during direct heating. l The direct ua ns were performed with a stand alone code heatinY1ofa ~ (DHEAT ) that calculated the end conditions of the event assuming that it was so rapid that it occurred essentially adiabatically and that the injected debris was, nevertheless, in contact with the atmosphere long enough for thermal equilibrium to be achieved. There were, at the time that this analysis was being performed, no phenomenological codes available that treated the direct heating process mechanistically. Holdup of debris on surfaces due to impact, heat transfer to surfaces, changes in particle size' distribution as debris passes through lower containment volumes on its way to the upper containment, interaction with water in the cavity, heat transfer and chemical reactions between the ejected debris and the atmosphere. and radiation between the hot debris, water droplets in the atmosphere, and the passive heat structures could not then be quantified. In the interim,. work has been proceeding in developed that thisareaandatleastonecomputercodeggg'g9pn can treat g of these phenomena .
- Experiments done at Sandialgag) '
seem to indicate that minimal j holdup will occur and about 90% of the ejected core debris could 4 be transported into the upper levels of the containment. How-
, ever, the experiments were performed for a cavity configuration
- similar to that in the Zion plant and may not be completely f characteristic of the Bellefonte configuration. Heat transfer and the combustion processes as well as the aerodynamics of
~
4 direct heating are involved and have not yet been studied in detail analytically. Potentially, the most important parameters that affect the , magnitude of the direct heating loads'on containment include: the fraction of the core debris injected into containment, the degree of completion of the combustion process for the various combustible components in the debris, and the heat absorbing capacity of the containment atmosphere. Table 4 shows'the , matrix of parameters selected and the calculated containment l conditions for each case. The fraction of core injected into i the atmosphere was varied from 0% to 90% of original core i inventory. The 0% case is not shown since it contributed no additional loading. The significant combustibles include zir-conium, steel and hydrogen. At the temperatures involved it is likely that the injected zirconium will be completely oxidized. The steel oxidation was varied by assuming no involvement or 100% involvement. Although hydrogen at a concentration of 3% is l 1
A not within the normal combustion limit, it is believed that much of it may be consumed in a catalytic process at debris particle surfaces. The hydrogen combustion contribution was assessed by 1 calculating cases with no hydrogen combustion and complete + hydrogen combustion. With regard to the heat absorbing capacity of the atmosphere, although the mass of non-condensible gases in the containment does not change significantly in the time frame of the direct heating process, the quantity of steam and water contained in the atmosphere depends largely on the accident sequence and on the timing of direct heating. MARCH code calcu-lations predicted that about 75,000 lbs of water were present in - the reactor cavity at vessel breach. This represents water that . was condensed in the containment building during the blowdown l phase, filled the containment sump, and overflowed back into the i reactor cavity. This quantity of water, when its heat of vaporization is included, represents a large effective heat
- capacity. If this water is actually located in the reactor cavity, it is probable that most or all of it would be swept out with the debris and should be included in the direct heating calculation. The effect of this additional heat capacity was considered by performing the calculation both with and without the cavity water in the atmosphere. For the cases in which the cavity water was assumed to be swept into the containment together with the core debris it was also assumed that the water and the debris were in good thermal contact resulting in complete vaporization of the water.
The results of the direct heating calculations indicate (see Table 4) that all the cases involving 90% injection resulted in containment pressure loadings that were either in or above the estimated containment failure range (144.7 to 153.7 psia). The j peak pressure, for example, in the case of 90% debris injection with hydrogen combustion, no steel oxidation, and complete - injection of the cavity water resulted in a peak pressure of 169 psia. Since the optimistic assumptions for the 90% ejection cases resulted in loadings that exceeded the estimated failure
; pressure, it was not necessary to perform the cases with more conservative combinations of parameters.
i The 50% injection cases were marginally close to the containment failure criterion and quite sensitive to the assumptions regard-ing the oxidation of hydrogen and steel. For the cases in which 1 cavity water was not involved (Cases-002), it required the com-plete combustion of the steel to exceed the failure pressure. Direct heating with hydrogen combustion alone produced a pres-sure of 139 psia while direct heating with neither H DOE steel combustion yielded a peak pressure of only 117 psia,2 well below the estimated failure range. With the involvement of cavity water (Cases-002A) only the most pessimistic assumptions i regarding combustion put the peak pressures into the containment failure range. ) l l . .- .__ -. . - -. . . -.
Although there are clearly a number of uncertainties regarding
- the direct heating scenario, it can generally be concluded from.
. these parametric calculations that for core debris inje'ction of ,
90%, containment failure is essentially certain, and that for ! injections much in excess of about 50% containment failure is i likely. The contribution due to steel combustion appears to be the largest single factor increasing the pressure by about 40 ) psi. Cavity water and hydrogen combustion have about equal but opposing effects, the cavity water decreasing the pressure by about 20 psi and the hydrogen combustion increasing the pressure by about the same amount. 4 The energies transferred to the atmosphere during direct heating as calculated by the DHEAT code were input to the CONTAIN code over an assumed event duration of 30 seconds. The last column in Table 4 gives the net energy transferred to the atmosphere during direct heating. Although the modeling assumptions in the DHEAT code (and in the hand calculations) assumed adiabatic-conditions (i.e. no heat transfer to structures), the CONTAIN code calculated the heat transfer processes during the direct heating transient. A comparison of the peak pressures for the adiabatic calculations and the CONTAIN code calculations revealed no significant differences. However, the version of the CONTAIN code that was used in this analysis did not have a model for radiation from the atmosphere to the walls and heat structures (The most recent version of the code now has a i radiation model). Recent calculations performed he Univer-sity of Wisconsin for the Surry reactor plant {7 utilizing newly developed mechanistic direct heating models have suggested that radiation to the walls during the direct heating transient can somewhat reduce containment pressure and temperature. For
- example, a case calculated without water droplets in the atmo-
- sphere showed a reduction in the peak pressure from 103 psia to about 85 psia when radiation was modeled. The effect, however, i may be overstated in the University of Wisconsin analysis. In ,
those calculations t j taken to be 20 seconds. thedurationofthedLgegy,heatingeventwas Experiments b ' however, seem to indicate a much shorter duration, on th; order of 5 to 10 l seconds. A shorter event duration would, of course, reduce the quantity of heat that could be transferred to the walls, { i The conclusions derived from the direct heating calculations have been used in the remainder of this study to define the containment failure modes. Thus, for cases involving 90% , ejection the containment was assumed to fail in a gross manner blowing down to atmospheric pressure within minutes. Those i scenarios which involved 50% core ejection or less were assumed l to develop containment leakage pathways. As already mentioned the peak loading conditions for the sequen-ces that involve direct heating were in each case achieved'at the time of direct heating, and those conditions are detailed in Table 4. Two 0% ejection cases were also calculated (Cases-000
and 100) with the assumption that, although the system was at pressure at the time of vessel breach, no debris was ejected out of the cavity region. In the absence of an energetic ejection of debris into containment, .the loading on containment at the time of vessel breach consisted entirely of the " steam spike" that results from core slump, vessel blowdown at failure, and the interaction of core debris with coolant in the reactor , cavity. The peak pressure at vessel breach for these cases (see Figure 2) was about 62 psia while the containment temperature l was about 260 F (Figure 3). Because the core debris was not ejected out of the reactor cavity for the non direct heating cases, the reactor cavity water was completely boiled off. at about- 450 minutes and a i core / concrete interaction ensued which began generating hydro- i gen. At about. 17.5 hours, a flammable mixture . occurred in containment and a hydrogen burn resulted which raised the containment0 pressure to about 100 psia and the temperature to about 1000 F (Figures 2 and 3). Hydrogen burns were not predicted to occur within the first 20 hours for any of the direct heating scenarios. 4.2 RADIOLOGICAL RELEASES FROM HIGH PRESSURE EJECTION Direct heating, besides its apparent capacity for initiating early containment failure, also represents a serious threat in terms of its potential for generating large quantities of radio-active aerosols and depositing them directly in the contain-ment atmosphere. Although most of the ejected core debris will i fall out onto the containment floor and hort immediately after direct heating, experimentsLgogyal~ ' surfaces have~ indi-cated that perhaps 2 or 3% of the ejected debris will subse-i quently remain in the atmosphere as aerosols for an extended period. In addition, several of the important radiological ! materials such as ruthenium may be heavily concentrated in this l aerosolized component due to. their possible involvement in. oxidation reactions during the direct heating sequence. Currently, the basis for a best estimate of the aerosol source. i term the high pressure injection sequence is the SPIT-19comingtest{Jom. In this test, 10.3 kgs of thermitically generated melt was involved, 9.8 kgs of which were ejected at high pressure. Based on measurements at various locations'in the experimental chamber, the total aerosolized fraction of , i particles less than 10 microns in diameter was in the range of 1% to 3% of the injected mass, and the component believed to be generated from the condensation of vapor was in the range of .6 to 1.8%. Thus, a bimodal particle size distribution was appar-ent with one mode centered at a mean aerodynamic equivalent diameter of 0.7 microns and the other at about 30 microns. The very small particle size mode appears to have been generated t
1-from the condensation of vapor components evolved from combus-tion reactions. The larger size distribution was generated mechanically . (i.e. atomization during the ejection,etc. ) .
- For the purpose of this analysis a total aerosol mass of 2% of the ejected debris mass was used, 1% consisting of the 0.7 l
' micron = size distribution and the other 1% consisting of the 30 micron distribution group. The fission product components for i j the 30 micron size aerosol were calculated by taking 1% of.each . of the fission product. group masses that were present in the injection. The fission product groups were the same as those
- used in BMI-2104 (see- Table 5). One percent of the inert i materials in the core debris were also included in the large l size aerosol component. The'small size (.7 micron) component, since it consisted mostly of re-condensed combustion products,
! was made up of only those . fission' products that undergo a ., significant oxidation reaction- together with the inert oxides (oxides of iron, zirconium, etc.) formed in the direct heating event. Thus .the 0.7 micron aerosol distribution' component was heavily enriched in the three fission products that may undergo-
- an oxidation reaction, namely tellurium, ruthenium and molyb-
! denum. The mass of each of these three. aerosol components was l calculated according to the following equation: 1 Mi=fo f i f ox Moi, (1) ! where Moi is the original inventory of species i, f is the 4 fraction of the core debris ejected, f1 is the fraction, of the species present in the core debris at the time of ejection, and i f ox is the oxidation release fraction (i.e., the fraction of species i present in the atmosphere that actually becomes oxidized). Theestima{ggfortheoxidationreleasefractionare taken from WASH-1400 , 0.6 for tellurium, and 0.9 for l ruthenium and molybdenum. There is some controversy regarding what values should be used for the oxidation release fractions
! and some indication that 0.9 may be too high for the ruthenium
~
i release. Tables 6 and 7 show the direct heating fission product i releases for cases in which BMI-2104 primary system releases
, were used and for the cases in which the MCT code was used to calculate primary system releases, respectively.
l The direct heating fission product releases were supplied to the i CONTAIN code along with the primary system releases for calcu-
- lating the transport and retention in the containment system and for estimating releases to the environment.
! 4.3 PRIMARY SYSTEM RADIOLOGICAL RELEASES {' f Fission product transport in the primary system and releases into the containment system have been estimated using the lhtt-telle suite of codes, CORSOR, MERGE and TRAP / MELT. These three j codes have been combined into a single. code package called MCT (MERGE, CORSOR, IRAP/ MELT). A description of the code together l l
- ?
1
. . -- . - _ - - - . = . - . . . . - .
4 with the modeling and numerical modifications can be found-in Ref. 5. Three configurations were run for the TMLB' high pressure 4 scenario employing the MCT code package. The first case (Type- , A) consisted of the configuration in which. fission product. decay heat- was -allowed to heat up the passive heat structures in each
- control volume of the primary system, but natural convection i between control- volumes was not accounted.for. In the second i i case (Type B) a model which estimates the heat transfer between J connected control volumes due to natural convective processes was activated. In the third case-(Type C) the same modeling options used in the first case were utilized-(i.e. fission
} product heating with no natural convection) and the case was run-out to about 10 hours after vessel failure in order to-calculate the late reevolution of fission products in the primary system. 1 Figure 4 shows the transport pathway for. gas flow and fission product relocation through the primary system. This diagram . also specifies the control volume sizes, ' flow areas, heat- l transfer areas and the heat capacities-for passive heat struc-
< tures in each control volume. The flowrate of gases and vapors i from the top of the core-together with the exit gas temperature :
was supplied to MCT by the MARCH 2.0 code. Gases exiting from ) the core region entered the upper plenum, passed through the outlet plenuu, hot-leg piping, surge line and pressurizer and ! exited through the PORVs into the containment. This pathway was the same for the first two cases. For the third case, however, j the vessel has failed and the flow path was somewhat reversed. j For that configuration the PORVs were assumed to be closed and ! the flow that exited- from the vessel had to do so fron'the breach in the bottom head of the vessel.
- There are several limiting assumptions that are inherent in the
! MCT code modeling and that may introduce some inaccuracies in ! the results. The TRAP / MELT fission product modeling does not i incorporate chemistry for the decomposition of CsI due to a high temperature and radiation environment and does not account for l radioactiv$32 decay of ggssion products into' daughter isotopes
- such as Te into I. There is considerable- evidence
! that CsI decomposition is an important aspect and may lead to ! higher releases into containment. ~For those cases in which
- natural circulation has been modeled there is a pathway for-natural circulation from the core region into the upper plenum.
However, there is no provision for the cooler downflow from the upper plenum to cool the core region. This is an artifact of the non-interactive interface between the thermohydraulic code ! (MARCH 2.0) that calculates the core condition and the primary system fission product transport code (MCT). The heat absorbing-structures have been treated using a lumped parameter model. , Although the heat structures are composed of steel ~which has-a
- fairly high heat transfer coefficient, the Biot number can be in l excess of 0.5 for the thicker walled structures (up to about 7 l
~ inches. thick). -This Biot number indicates that there will be a significant temperature gradient in the heat structures and the .)
-surface temperatures will- be somewhat higher than those calcu- '
lated by the- lumped parameter model. The higher _ surface tem- l peratures will primarily affect the timing of fission product i relocation tending- to' -move the volatiles through the system l somewhat faster than has been estimated here.
- The MCT code modeling for' natural convection has already been
, identified as a rather simplistic one-dimensional approxima-tion. The lumped parameter modeling that is characteristic of the MCT code package in general makes this type of approach necessary in order to evaluate the effects of natural convection on the RCS releases. The question remains, however, as to the accuracy of the model used to calculate natural circulation in i , view of the rather complex geometry in the primary system. Figure 5 describes the model that is currently in the MCT code. The model assumes that the Bernoulli equation , v= [2d P/p0]* , (2) ! I can be ' applied to describe the flow field between two control i volumes located one above the other. The driving mechanism that sustains flow is the buoyancy resulting from the lower control j volume being at a higher temperature than the upper volume, P= (py-p0)gH. (3) The Bernoulli equation (Eqn 2) together with the buoyancy term j (Eqn 3) and the ideal gas law yields the natural convection velocity, v = [2gH(T0 /T1-1) ] .5, (4) The further essumption is made that there is no pressure dif-j ference between the two volumes so that the upward flow of gases must be balanced by an approximately equal downward flow. To implement this requirement the flow opening is divided in half with half of the area devoted to upward flow and the other half ! to downward flow, thus, forming a single-roll flow cell. j A search of the literature has revealed a very limited treatment of this type of heat transfer problem. The approach that has i ggg used in the MCT code was described in the literature However, the approach as _ described in the literature i , was applied to the case in which the control volumes were l l located side by side with an opening of height, H, in'the ver- , tical separating partition. In this configuration, H,=can be l thought of as the vertical length of the flow path between the l control volumes. For this configuration a steady flow field is i expected and the Bernoulli equation can be employed to approxi-mate the flow velocities. The configuration in which the con-trol volumes are one above the other with the lower cell having . I
i-4 i. i the higher temperature is basically unstable and a steady flow field .cannot be assumed. Thus, the Bernoulli equation cannot be used to the flow for this situation. An experimental { 4 correlation desy{gge does, however, exist for this configuration
- and in-terms-of the Nusselt number the expression is as follows
Nu = 0.0546Gr.55Pr(L/H).33, (5) where L is the characteristic dimension of the opening and H is , the vertical length of the opening, while Gr and Pr are the-Grashof and Prandtl numbers, respectively.
- In order to estimate the accuracy of the model used in the MCT code a comparison was made between the MCT model and the above correlation. The velocity equation (Eqn 2) used in the MCT model can easily be cast into a heat transfer correlation of the type shown above and when this is done an expression for the Nusselt number in terms of the Grashof and Prandtl numbers is obtained
- .
Nu = 0.33C Gr.5Pr. (6) i Here C is the flow discharge coefficient. Choosing two control volumes such that L/H is about.l.0 and setting the discharge coefficient equal to 1.0 (as is assumed in the MCT model), cal-
- culations of the Nusselt number using the two correlations were made. Figure 6 shows the Nusselt number as a function of the temperature difference between control volumes for the two cor-2 relations. In figure 6 the curve labeled Nu(1) is the correla-tion in Equation 6 which is the one used in the MCT code. The curve labeled Nu(2) is the one expressed by Equation 5. For this geometry the MCT code overpredicts the heat transfer
- between control volumes by a factor of about 4 at low tempera- l ture differences and a factor of 3 at higher temperature dif- i
- farences. Clearly, for the configuration in which control
- volumes are one over the other, the MCT code overpredicts 4
natural convection heat transfer. However, for volumes that are connected in a horizontal arrangement such as between the out-let plenum and the hot-leg piping, assuming that H'is defined as i the height of the opening and not the differential elevation
- between the two volumes, the heat transfer rates between volumes may be a fair approximation.
Lumped parameter treatment of heat structure thermal response
- represents yet another source for error in the calculations.
Calculations show that for steel structures with thicknesses on . the order of 4 inches, the surface temperature of the structure l as calculated by one-dimensional transient finite difference ! methods can be as much as 30% higher, and on the average 15% higher than the temperatures calculated using a lumped parameter treatment. Since lumped parameter calculations predict that surface temperatures are lower than they actually are,_the l effect is to hold up fission products on structures and move l __ _ . _ _ _ _ _ . _ _ _ _ . _ - . _ .. _ _ _ . _ _. _ _ ._ _ _. _ _ ._ _ _ _ . _ _
~ .. -
. ?.' them 'through the system more slowly than they would actually move. 1 The errors associated with overpredicting natural circulation ! heat transfer and underpredicting heat structure temperatures i l are competing and tend to reduce the - net _ error in terms of fission product Multi-dimensional code calculations _ and experiments yggpsport. ! I have verified the importance of natural convection in -the primary system with regard to primary system heatup and the associated cooling effects on the core. Turning to the results of the primary system fission product transport calculations for the TMLB' high pressure scenario, Figures 7 to 10- show the conditions in the primary system as 4 calculated by the MARCH code. The blowdown phase of the acci-- dent (until -incipient core meltdown) was calculated by the RELAP5-MOD 1.6 code and does not appear on these plots. Figure 7 4 shows the primary system pressure increasing up to the PORV set point and- remaining at that level until vessel breach. The average ' core temperature is given in Figure - 8. indicating a ]
- gradual temperature increase until core slump when the tempera-ture was about 4700 0F. As indicated in Figure 9, the core i slumped at 65% meltdown and the ' fraction of clad-oxidized in-vessel was about- 50%. The in-vessel hydrogen production, 1
Figure 10, for 50% clad reaction was about 1150 lbs and produced a hydrogen concentration at vessel failure of about 3 vol 4 in containment. 4.3.1 TMLB'(HPE) WITHOUT NATURAL CONVECTION (TYPE A) j The core outlet gas flow and outlet gas temperature for the high pressure ejection cases are shown in Figure 11 and 12. The sud - l den increase in flow at about 4600 seconds results from core slump which produces large steaming rates as the core debris ] quenches in the coolant on the lower vessel head. The outlet
- gas temperature follows the same trend as the core temperature increasing to a peak of about 3800 F at vessel failure. The
. thermal response of the primary system is shown in Figures 13-I and 14. As expected the gas temperatures decrease with distance from the . core region, but very large temperature differentials l exist between adjacent control volumes ~.which under.the rela-tively quiescent conditions in this accident. sequence would result in buoyancy driven flows. The upper plenum, as expected, is the hottest volume in the reactor system because~it is nearest to the hot core exit gases.and it also receives radia-j tion heat transferred from the top of the core. The surge line i is calculated to heat up quickly because of the relatively low i thermal mass of the structure. The outlet plenum, the hot-leg and the pressurizer heat up somewhat more-slowly, because of
- their large thermal masses.
I Figure 15 shows the fission product heat deposited in each control volume. Clearly, the. upper plenum region of the reactor-system contains the largest portion of the fission products. At
i approximately 4600 seconds into the accident, the core collapsed and produced a' relatively large flow of steam out of the core . region which swept, the vaporized volatile fission products out of the upper. plenum region. This accounts for the observed. decrease in the fission product ' heat at that time. Corres-
- ponding. to this decrease, there is an increase in the fission l product heat in the pressurizer and the surge line. This is due I to the retention of the volatile vapors by those cooler volumes. '
Figure 16 shows the retention factor of cesium iodide. The { i retention factor as used in this report is defined as the frac-l tion of the total fission product species released out of the l fuel matrix which is retained by either vapor condensation on structures, vapor condensation on aerosols with subsequent set-tling of the aerosols, or chemisorption. The sudden increase in. j gas flow out of the core at the time of core collapse drives the
- gas-borne volatiles away from the upper plenum region to the i cooler volumes- which include the hot-leg up to the surge line, l
the surge line and the pressurizer. There.the vapors condense on the structures as well as the aerosols. This accounts for the i increase in the overall retention factor for both cesium iodide 4 and cesium hydroxide, at approximately 4600 seconds. i The behavior of tellurium, shown in Figure 17, is similar to that of cesium iodide and cesium hydroxide. The only difference of interest occurs at the time of revolatilization of the con-densed tellurium in the upper plenum. This revolatilization is
- shown as a decrease in the upper plenum tellurium retention l- factor at approximately 4600 seconds. At that time, tellurium revolatilizes and is entrained in the gas flow. The. residence i time of the revolatilized tellurium in the upper plenum region I is insufficient to result in significant chemisorp*cion and therefore, the tellurium vapors continue flowing to the down-stream volumes where the vapors are readily removed by chemi-j sorption and condensation on structures and aerosols.
i It is of interest to note that the principle removal means for l i both cesium iodide and cesium hydroxide is by vapor condensation l on aerosols and their subsequent deposition. This is illus- l trated in Figure 18 which shows the mass of retained CsI. For
- both CsI and Cs0H vapor condensation on structures plays only a j minor role. In the case of cesium hydroxide, chemisorption did not play a significant role in its removal. Tellurium, however, l shows significant removal by chemisorption (Figure 19). As indicated in Figure 20 the total CsI released to containment was about 0.8 Kg primarily in the form of particulate.
r l The results of this analysis indicate that for all fission i product species, greater than 90% of the fission products released from the fuel-during the core melt process remained in the primary system. The majority of the fission products remaining in the primary system were located in the upper plenum region of the reactor vessel with small quantities distributed throughout the remainder of the primary system. As will be seen in a subsequent section, following reactor vessel failure the
~26-
- - ~ . _ _ , - - . _ - . - . . . -- .- . -_. . - - - _
fission products deposited on structures in the primary system , will- slowly heatup the structures and be revolatilized. The ! revolatilized fission products under certain conditions can be readily released through the opening in the lower reactor vessel head. 1 4.3.2 TMLB' (HPE) WITH NATURAL CONVECTION (TYPE B) Figures 21 and 22 show the thermal response of the primary i system during the TMLB' sequence when the effects of natural + convection are included in the analysis. It is readily observed
- from the gas temperatures that the gases-in the upper plenum and i
the hot-leg up to the surge line are quite well mixed because of
- the recirculating flows between volumes. Similarly the gaseslLn the surge line and the pressurizer are also well mixed. When
- compared to Type A (Figure 13) discussed previously, it is seen
- that the gas temperatures in the primary system volumes are significantly higher than in the case where natural convection a was neglected. This is due to the presence of recirculating i flow between the upper plenum volume and the core itself, l resulting in a continuous exchange of gases between the top of l the core and the upper plenum region.
I The heat structure temperatures calculated in this analysis are shown in Figure 22. These results indicate that the-hot-leg i temperature increases to a level at which loss of strength of j the hot-leg piping would be expected. This level is approxi-mately 1300 0 F and it occurs at about 4400-4500 seconds, i before the calculated time of lower reactor vessel head fail-I ure. The effect of a failure in the RCS before core slump could j be to negate the high pressure ejection of core debris and the
; dispersal of significant fractions of the core into the upper i containment regions.
Figure 23 shows the fission product energy. in each primary i system volume. It is seen that the fission products are some-I what more spread out through the primary system than in the case { without natural convecti~on. Figure 24 shows the cesium iodide ! retention factor. The retention factors for the volatile spe-i cies~ are considerably lower than those calculated for the case , without natural convection. The reason for this is that the i natural convective flows increase the gas temperaturcs in the j RCS, so that the primary means of removal of the volatile spe-
- cies is no longer through condensation on aerosols and subse-
- quant deposition, ~but rather condensation on the structural I
walls, as seen in Figure 25. 'This is due to the inherent i assumption in the MCT2 code that the aerosols exist at the
- carrier gas temperature. Because the temperatures are much j higher in this case than in the case without natural convection, the volatiles no longer condense on the aerosols, but rather, condense on the heat structure walls since the structure walls are cooler than the gas. Due to the existing inter-volume flows, the residence time in a given volume is insufficient to allow significant condensation on heat structures. Eventually
- - . - - - n . -. - -
,o 4 . .the heat structure temperatures increase sufficiently to com-plately terminate further . vapor- condensation. At the' time of ! core slump, starting at approximately 4600 seconds, the . majority l- of the volatile material that has not been deposited on the heat structures in the_ primary system is in vapor form rather than } condensed onto gas-borne aerosol surfaces. The. increased flow at i core slump sweeps out the vapors from the primary system ~into i the containment with;a small portion of the vapors being trapped } in the surge line and pressurizer regions. This small. trapping i accounts for 'the increase in the overall. retention . factor observed in Figure 24 at approximately 4700 seconds. I Figure 26 shows the accumulated mass of cesium iodide which was
- released to the containment during the TMLB sequence. During
! the early_ stages of the _ sequence up to approximately 4400 4 seconds, the primary form of the. volatiles released to the l containment was particulate, i.e. vapors condensed on aerosols.- { When compared to the corresponding figure in the previous case (Case A), the amounts released to the containment in particulate ! form appear to be considerably larger. This is due to the effect of the recirculating flows between the surge line and i pressurizer which have the effect of reducing the deposition I rate of the suspended aerosols which carry the condensed vola- ! tiles. At approximately 4400 seconds the flow rate of gases out l of the core begins to increase significantly moving the vapors i from the reactor vessel volumes through the surge line and j pressurizer and out into the containment. This~ gas flow also i sweeps- the existing gas-borne aerosols out of the surge line'and i pressurizer. Therefore the arriving volatile vapors are pri-marily removed by the structural walls, as there are no gas- ! borne cool aerosols to provide large condensation surfaces. The removal process by condensation on the walls is slower than that ! by condensation onto aerosols and therefore a substantial ~por-tion of the vapors continue on through the pressurizer and out ! to the containment. It is only after the gas temperature in the ! pressurizer region falls at about 4700 seconds that condensation l onto the aerosols takes place and an increase in the retention j of volatiles by the pressurizar is observed. The behavior of tellurium shown in Figure 27 does not differ i significantly from the other volatiles. Once again, the high j gas temperatures in the primary system result in tellurium j existing' as a vapor rather than condensed onto aerosol sur-I faces. Consequently the primary means of removal of tellurium i is through chemisorption and the rate of removal depends on the
- residence times of tellurium in each primary system volume. At t the time of increased gas flows out of the core, i.e. approxi-mately 4400 seconds, the vaporized tellurium in the primary systen is swept out into the containment as shown in Figure 28.
4.3.3 TMLB' (HPE) WITH LATE FISSION PRODUCT REEVOLUTION (TYPE C) The results of the TMLB' analysis with no natural convection indicate that a great majority of the fission products released from the fuel matrix during the meltdown are retained in the
- . - . ~ . -- - - . . - _. - -. . - - - - - . . _ _ ._ _
- ~
I i primary system. Moreover, most'of the retained material remains - in the upper plenum region. Following the reactor vessel. fail- ! ure, these retained materials will heat up the structures on
- j. which they are . deposited to a level at which revolatilization l back into the gas phase will take place. The pressure rise-due to heatup will induce a small flow out of the reactor vessel, but the primary mechanisms for removing reevolved fission prod-
- i. ucts from the RCS are a gradual containment depressurization due l to leakage, or a late containment failure. In order to estimate.
! the amount of fission products that would be carried out during . the long-term heatup of.the reactor system, a simplified analy- ! sis was performed, one which examined the heatup rate of the i gases and structures and-calculated the revolatilization of the i deposited fission products. A more detailed analysis, although
- possible with McT2, was not performed due to resource con-
! straints. The actual releases of reevolved volatiles was esti-
, mated from the reevolution rate calculated by MCT together with j the leakage induced containment depressurization rate calculated
! by the CONTAIN code. Late reevolution was calculated only for i the cases with containment. leakage. Early gross containment. l failure depressurizes the containment before significant i reevolution can occur. Primary system releases after contain-ment failure are thought to be small because no effective
- mechanism exists for getting the reevolved fission products out l of the primary system.
The thermal response of the primary system for Type "C" is shown j in Figures 29 and 30. Vessel failure occurs at approximately l 5000 seconds. After vessel failure, the heat structure tempera-4 tures begin to increase which causes-the deposited materials to revolatilize. The upper plenum region contained the largest q quantity of deposited material at the time of vessel failure, { and therefore, that region is observed to heat up most rapidly. ' 1 Figure 31 shows the cesium iodide retention factor and it is i observed that at approximately 6000 seconds, the revolatiliza-tion of cesium iodide commences and by 8000~ seconds, all of the !_ deposited cesium iodide in the upper plenum has revolatilized. ! l Because of the simplifying assumption of the long-term' analysis, ! (that specified that the post vessel failure flow of gases con- ] tinued in the same manner as the pre-vessel-failure flow, i.e. 1 from upper plenum to the pressurizer) , the revolatilized ' i material is carried into the downstream volumes. In reality, j this would not be the case and, in fact, the revolatilized ! material would be carried from the upper plenum back through the j core region, out the breach in the lower vessel head and into ! the containment. Moreover, any revolatilized materials from the i remainder of the primary system would not flow into the pres-l surizer but would flow out through the: ruptured lower head of
- the reactor vessel. Therefore, for the purposes of estimating j the release rate of volatiles from the primary system in the
- long-term following vessel failure, the rate'of revolatilization I was assumed to be the rate of the loss of retained mass, as i extrapolated down to zero retention, as shown in Figure 32.
l Thus, since the majority of the retained material was located in. the upper plenum region and the total mass retained is shown by
d 4 Lthe peak in Figures 32, the revolatilization rate can be ob-- tained from these figures by calculating the rate at which the retained mass decreases.- - At- the time the extrapolated curves , intersect the -abscissa, the volatiles initially present in the reactor- system are completely in vapor form and available for release to containment. The rates of release of the revola-tilized fission products from the RCS to the containment were
- based on the depressurization rate of the containment system.
i The analysis of the revolatilization case did not account for natural convective flows because of resource constraints. Qualitatively, however, it is expected that convective flows would exist in the RCS because of the existing large thermal gradients. These flows would have the effect-of carrying the l vapors released in the upper plenum as a result of revolatili-j zation into thel cooler regions of the primary system where the
- vapors could recondense. The location of the break in the lower j vessel head together with the gradual leakage induced contain-ment depressurization would lead to bulk flow in the opposite direction to natural convective flows and would tend to limit
, upward relocation of fission products. The net effect of i accounting for natural circulation flows in the long term ! analysis would be to somewhat extend the time during which the j revolatilized fission products would be released to the contain-i ment atmosphere and reduce the rate at which these releases take I place. Since the heat generated by radioactive decay accounts for only about 10% of the heat absorbed by the structures during the meltdown phase of the accident, retention of fission products in the primary system prior to reactor vessel failure does not depend significantly on the inclusion of fission product heating in the calculation. However, the effect of fission product heating becomes very important after vessel-failure when exten-sive revolatilization takes place solely as a result of fission
- product heating of structures. This can have a marked effect on i the source term if this revolatilization takes place in conjunc-l tion with a containment undergoing deprassurization or if con-l tainment depressurization occurs shortly after the revolatiliza-tion in the primary system. Moreover, the calculations show I that the inclusion of the natural convection phenomena plays a l significant role in the retention capability of the primary I system. The net effect of the natural convection phenomenon is to increase the primary system gas temperatures and, thus, structure temperatures which causes a reduced rate of fission product removal from the gas stream. This is primarily due to the inability of the volatiles to condense onto the aerosol surfaces, that are at the gas temperature. Since the aerosol-area to volume ratio is much larger than the structural area to volume ratio, the removal rate of volatiles is markedly reduced since the only mechanism of volatile removal other than chemi-sorption is through condensation on structural walls. In addi-tion, at the time of core collapse, a large gas flow is induced through the primary system which sweeps the vapors out into the containment. This occurs at such a rate that the residence times
of the vapors,.as they pass through:the cooler volumes on their , } way- out to the containment, are insufficient to result in any l l significant removal .of these vapors from the gas stream. The : l net effect is a considerably reduced retention of_ volatile l fission products in the primary system. i i Table 8 summarizes the primary system releases compared to-those calculated the BMI-2104foraTMpgj.sequenceintheZionplantasreportedin j study It was stated previously and should be
- reiterated here .that the selection of the BMI-2104 Zion plant '
l calculations for purposes of comparison with the present results i was somewhat arbitrary. In terms of primary system releases, ! the Surry or Sequoyah BMI-2104 results could equally well have been used, since both are PWR plants. The primary system , releases calculated in the BMI-2104 study were significantly . 1 higher for Surry and Sequoyah than for Zion and the differences i are 'not entirely clear. Because of the differences in RCS j releases between these three plants, the comparisons made ! between the present results for Bellefonte and the results for ' j the BMI-2104 Zion plant should not be viewed as being particu- { larly significant. The important conclusions emerging from this j study are related to the relative effects on the RCS releases due to the incorporation of new phenomenological models in the i primary system fission product transport code package. 1 I For the meltdown phase without natural convection the MCT code
- predicted CsI and CsOH releases are only slightly higher than
) the BMI-2104 releases (factor of 1.5). Tellurium releases are higher by a factor of 6. The MARCH code modeling for the melt-progression phase and in particular, the clad oxidation parameters, strongly affect the release of tellurium from the fuel. Consequently the high primary system Te releases are partly due to melt phase fuel releases that are more than twice as large as those in the BMI-2104 Zion analysis.
- The net effect on the releases of volatiles up to the t'ine of l vessel failure due to the modeling of' fission product heatup is
! seen to be relatively insignificant. There is not sufficient j time before vessel failure for the fission products to heat up j the primary system heat structures to the extent that would i allow reevolution of volatiles in that time frame. Fission product heating as it affects the reevolution of volatiles subsequent to vessel failure is, however, seen to be qu.r.e significant. Ultimate releases of CsI and CsOH in the reevo-lution phase are nearly an order of magnitude high than what was released in' the core meltdown phase. Total releases, including , , pre and post vessel failure releases are a factor of 10 higher i than BMI-2104 estimates for Zion. ' The present analysis which represents an initial attempt at : estimating the effects of natural convection on the melt-phase releases indicates that these may be higher by a factor'of 10 for CsI and CsOH. Assuming that the reevolution phase with-natural convection produces reevolved releases similar to those shown in Table 8 for reevolution without natural convection, the I t .-.-,---.-m.--- .-..,--,,,,...---.,.,--..,_-s.-- ._-m_----- - - _.,- - ,,._- -~ - - ,.. - ,
1 i total releases of CsI and Cs0H could well be in excess of 50% of l the initial inventory of these species for this accident sequences. It should be reiterated here that the natural circulation model is quite simplistic and represents an early attempt at treating natural convective processes. The TRAP / MELT code employs lumped parameter modeling throughout the code and would have required massive alterations in order to incorporate the 2 or 3 dimen- ; sions which would be necessary to treat natural convection in a j completely rigorous manner. have some precedence in the Thepresentgggrggyh,whichdoes literature '
, employs the l Bernoulli equation (ideal flow assumption) together with a buoy-ancy term to calculate the natural convection flows between adjoining control volumes and appears to somewhat overpredict natural convective processes.
The particle size distribution of the aerosols emerging from the primary system varied only slightly through the accident.- The initial aerosol releases came out of the system at about 1.5 microns increasing quickly to about 3.5 microns within the first 300 seconds and then gradually to about 5.5 microns after the release levels became significant. The preponderant fraction of the aerosols, however, were released in the 5.5 micron range (see Table 9) with a standard deviation of about 3 microns. l l l 1
1 [ -4.4 CONTAINMENT AEROSOL TRANSPORT AND RELEASES , l . l l The thermohydraulic sources from the various codes used, the- ) ! radiological sources .from the primary system releases, direct ' i heating releases, and' core / concrete interaction releases were
- all input to .the CONTAIN code. The CONTAIN code was used to j ~ calculato the transport and retention of aerosols and fission i products in the containment system and to estimate the radiolog-ical releases to the environment.
l Except for one multiple-volume case, the CONTAIN code analysis ! of the TMLB' high pressure sequences was performed using a ~ single containment volume connected to a very large " receiver" volume representing the environment. The " receiver" volume-was necessary for accumulating the various fission products.that were released from the containment either by leakage or by gross .
- containment failure.
The CONTAIN code models aerosol particle size distributions ! through a user specified number of discrete particle size j groups or " bins". The input requires the mass median particle size together with the geometric standard deviation of the par-
- ticle size distribution for each aerosol species. From this d
information the code sets up the group particle size ranges and i the total masses of aerosols in each size group. For the single volume cases the aerosol size distribution was characterized i using 20 particle size groups. The particle size distributions j are given in section 4.2 and 4.3. 1
- 4.4.1 SINGLE VOLUME CALCULATIONS i
! A matrix consisting of 15 cases was analyzed for the high pres- ! sure ejection scenario. The case matrix is described in section ! 4.0 (See Table 1). The matrix of cases utilized to assess con- ! tainment releases for this accident sequence employs combina-j tions of assumptions, parameters, and modeling capabilities-3 applied across the entire set of codes and calculations used in i this analysis for the various thermohydraulic and radiological ! source terms. Direct heating parameters, containment failure j mode assumptions and primary system release modeling parameters i are combined in such a way as to cover a reasonable range of possible source terms. l Table 5 shows the constitaents of each of the fission product groups and their initial total core inventories. Table 10 is a summary table that gives the integrated fission product releases ! to the environment at 20 hours for each of the fission product l groups for the cases in the high pressure ejection matrix. Case-000 represents the most optimistic set et assumptions for the high pressure ejection scenario. Thic : case combines BMI-2104 primary system releases with no ejection of' core debris. into the containment atagsphere (no direct heating component) and a very small (1 in ) induced containment leakage. The maximum airborne masses of CsI and CsOH for this case were about i l
..-. . - - . _ - - - - ~_~ - - , - - - . . . .
I-l 1 l- 1.4 I ! x 10~gsand9lbsandtherelgasesfromcontainmentwereabout4-lbs and 4.4 x 10~ lbs at 20 hours, respectively. [ These are extremely small releases. The releases from the'tel-t lurium, . strontium and lanthanum groups result primarily-from j core / concrete . interactions that commence at about 8 hours-into- 1 the accident sequence when the cavity water was-boiled'away. l !- The effect of leakage area on containment releases for this ! scenario can be seen by comparing Cases-000 and 100 in Table i 10. These cases are identical'except for the leakage area which i is 2.4 times larger in Case-100. As expected, the releases are j approximately a factor of 2 larger for Case-loo. i The most pessimistic case in the matrix is Case-011B. This case { includes a 90%' injection of coredebrisintothecontainmeng' i atmosphere which results in grosa containment failure (7 ft
- hole). The primary system release calculations for this case
! included a model for natural convection that contributed to j quite high releases of volatiles into containment. These are j the type "B" releases (see Section 4.3.2). Comparison of-the i airborne CsI mass (Figure 33) to the mass of CsI leaked.to the environment (Figure 34) reveals that a large fraction (about , 2/3) of the aerosols released into the containment during the meltdown and vessel failure phases are released into the envi-j ronment. Table 10- shows that the total releases of the vola-l tiles CsI, CsOH, and Te at 20 hours to be 16%, 12% and 21% i respectively. Containment failure coincident with direct heat-ing which was assumed to generate high concentrations of Te,=Ru, and Mo aerosols leads to very high releases in these groups. The Ru group, for example, which contains both Ru and Mo'has a 41% release fraction for this case. l There are several comparisons that can be made across the set of high pressure injection cases (Cases-001 through OllB.in 90% 1 Table 10). Comparing cases that used BMI-2104 Zion ~RCS releases I l with those calculated from the MCT code without natural circu- ! lation (type "A" release) reveals essentially no differences in I ! the CsI and CsOH releases. The tellurium releasesLare about j twice as large in the MCT prediction due primarily to a dif- { ference in the quantity of Te released' from the fuel. The
- ruthenium group releases are about the same because both are j dominated by the assumed direct heating releases _rather than the I RCS releases. The MCT code apparently predicts somewhat' lower
! releases from the strontium group and about the'same releases in ! the lanthanum group. The major differences for-the 90% high pressure injection cases results from the use of the MCT Type "B" RCS releases. These are the RCS releases obtained from using j the natural convection models in MCT. A comparison'between i Case-001A and 011B reveals that the CsI and CsOH releases from !' containment using type "B" RCS release are a factor of 10 larger l than those using the BMI-2104 Zion RCS releases. Againi the Te and Ru groups are similar, because, these are dominated by the direct heating releases rather than the RCS releases. , The 50% debris injection direct heating scenario is probably j best represented by case-012A or ' Case-112A. The difference i between the two cases was the size of the induced containment l 1.~____-.._.__...___.__._~_____..____._ _ . _ _ _ _ . _
leak area. These cases involve a 50% injection of core debris
; into -the contaigment atmosphage, which induces a containment
} leakage. (5 in and 12 in ) rather than a catastrophic
- failure of containment. The primary system releases for these cases consists of type "A" releases during the core melt down
- phase (without natural circulation) and type "C" releases for j late reevolution of volatile fission products-(see Section 4.3).
i j Begause the -leak area for Case-012A was relatively small (5 , in ) compared to containment failure scenarios, the contain-i ment depressurized rather gradually with adequate time in the
- containment system for aerosol removal . mechanism to have an j effect. The total releases of the volatiles CsI, CsOH and Te as j well as the ruthenium group at 20 hours were about 1% of the initial inventories for this case. The ultimate releases would have been somewhat higher since, as the graph of leaked CsI (Figure 35) indicates, the release rates had not yet leveled out at 20 hours.
Comparisons made between the various 50% injection cases show very strong enhancement in containment releases due to the j reevolution of fission products in the RCS. A comparison of Cases-002A and 012A, for example, shows CsI and CsOH releases to
;' be a factor of 13 or 14 higher with reevolved RCS releases than with the BMI-2104 Zion RCS releases.
t l A rather surprising result is seen in comparing Cases-112 and , ll2A. The only differences between these two cases is that case i 112A assumes that the water in the reactor cavity at the time of
! vessel failure was swept out with the core debris and completely
{ vaporized in the direct heating event, while in Case 112 it is j assumed not to participate. A comparison of the CsI and CsOH
- containment releases for these two cases shows that the case i with the additional steam component due to'the inclusion of the
! cavity water (Case-112A) has releases of these volatiles that is j higher by a factor of about 1.7. There are two competing l effects that explain the results. The additional water in the i atmosphere would tend to condense on the aerosols and more ' ! readily* " wash" them out of the atmosphere. However, the con-densation of the additional steam on structures in the contain- ! ment reduces the containment pressure more rapidly in Case-112A ! than in Case-112. Also, the containment pressure was higher in ] Case-112 because the water was not swept out of the cavity and i L remained as a steam source as it was heated by the debris j remaining in the cavity. The more rapid depressurization for Case-ll2A results in about twice the quantity of revolatilized i fission products being released from the RCS. Apparently, the large additional release of revolatilized fission products dominates over aerosol removal by condensation resulting'in higher releases for Case-112A. The particle size distribution of the aerosols which were leaked to the environment for Cases-011B and 012A are compared in Fig-ure 36. The abscissa on this plot represents the fraction of the total mass of aerosols that are in that particle size inter-val bounded halfway between adjoining points. The effects of i
aging of the aerosols is clearly illustrated in this figure. The aerosols that were released from containment at the time of containment failure (Case-OllB) show the distinct bimodal dis-tribution characteristic of direct heating releases. The i aerosols in Case-012A have been released gradually over a 20 hour period and display a size distribution that indicates aerosol aging. The quantity of aerosols in the lower size bins are now a somewhat higher fraction of the total as they are removed by settling less rapidly than larger particles. The larger size particles in the distribution have been depleted for l the same reason, and agglomeration has shifted the peak in the l size distribution to the right. The bimodal distribution is barely noticeable in the aged aerosols for this case. The disposition of aerosols between the atmosphere, the various surfaces within the containment and the quantity leaked to the environment is illustrated in Figure 37 and 38. The large fraction of aerosols ejected out of the containment due to the direct heating induced failure is evident in Figure 37 which shows results for Case-OllB. The fraction of material ejected from the containment amounts to about 63% immediately after direct heating with the remainder being settled out on the floor (27%) or plated out on the walls (10%). The leakage scenario represented by case-012A demonstrated a much different disposi-tion of aerosols. In this case (Figure 38), only a very small fraction of the aerosols were leaked from the containment, about 2%, with the largest fraction ending up on the containment floor (72%) and the remainder (26%) was deposited on the vertical walls. 4.4.2 MULTI-VOLUME CALCULATIONS The calculations that have been presented to this point were performed using a single-cell representation of the Bellefonte containment system. In order to achieve an estimate of the attenuation of fission product releases due to a more realistic representation of the the actual containment configuration, a multi-cell calculation was performed for Case-012A. In this calculation the containment was sectioned into 7 physical vol-umes with associated flow paths (see Figure 39) The main physi-cal boundaries were chosen according to floor levels within the containment. The volumes for the calculations as numbered in Figure 39 were: !
- 1. Reactor Cavity
- 2. Steam Generator / Pressurizer Compartment
- 3. Steam Generator Compartment
- 4. Lower Level Rooms
- 5. Mid-Level Rooms l 6. Upper Level Rooms
- 7. Containment Dome.
The Case-012A accident scenario has a complex sequence of hydraulic and aerosol sources which can be broken down into early and late sources. The volumes into which the sources were released were as follows:
Hydraulic Sources--- _ . Reactor Cavity (vessel failure-{early), concrete interaction (late}} Lower Level Rooms (pressure relief valve (early)) Dome (concrete floor degassing due to injected' core debris (early and late}} Aerosol Sources--- Reactor Cavity (vessel failure (early)) Lower Level Rooms (pressure relief valves (early}}. Containment Dome (direct heating aerosols (early}}. Results from the calculations which compara single and multi-cell leaked masses are given in Table 11. The two multi-cell calculations show a comparison of results with- and without natural circulation included in the inter-cell gas flows. The numbers shown in Table 11 are the ratio of the multi-cell releases to the corresponding single-cell releases. These results indicate that the differences between the single and multi-cell releases 'are minor for this scenario. The m differences that have been observed in previous studiesyjp which compared multi-cell to single-cell calculations were the result of different aerosol agglomeration rates that can occur when aerosol concentrations within volumes are significantly ' disparate. A less pronounced effect is the additional wall and floor deposition which occurs in-the multi-cell calculations as fission products are transported through the various cells on their way to the dome region. In the present case the latter effect is dominant. The main source of additional aerosol' con-centration is sourced directly into the dome in order to approx-imate a direct heating event. 'This assumption is consistent with the argument which supports transport of' ejected debris by particle inertia rather than gas entrainment. It should be noted that the magnitude of.the effective atten-uation factor obtained in the calculations by nodalizing the containment volume is highly sequence dependent. The direct : heating event in this particular sequence forced most of the aerosol mass into the upper containment volume, so that no sig-nificant differences were seen in any of the aerosol releases. In fact, the multi-cell treatment- produced somewhat enhanced releases. from the fission product groups most heavily concen-trated in the direct heating releases (Ru, La). The fission ' product groups that were released primarily in lower containment-i volumes (CsI, CsOH, Te) did , as expected, see some attenuation in their releases (about 10%). In general, therefore, the attenuation- factors shown in Table 11 may be applied to the 50% direct heating cases, 'but should probably not be applied to either the non direct heating or the 90% direct heating scenarios.
t 4 l 5.0 TMLB' PUMP SEAL LOCA i Although .this accident sequence was intended to treat the f situation in which. the seals in the main coolant puros were
~
induced to fail due to the loss of coolant flow,.the only. available thermohydraulic analysis for this sequence was
- performed. with the pump seals failed at the outset of the
- accident. At present there is no information available from I which to estimate the mode of pump seal failure or the time l required to induce failure under accident conditions. In the i absence of the necessary information, and of a RELAPS calcula-tion that incorporated a. lag time for seal failure, the present l
- calculation assumes that the seals fail immediately. In'the j strictest sense, then, the sequence analyzed here is not actu-
{ ally an induced pump seal LOCA. The break size, assuming the failure of all four pump seals, is i less than 2 in so that this sequence is similar to an S D j with loss of all primary system makeup and cooling as well as 2a failure of containment sprays and coolers. . The pump seal LOCA sequence was analyzed assuming that a direct heating event did not occur. The primary system pressure at the time of vessel failure was about 750 psi lower than in the high pressure ejection cases, but there is no reason to believe that { an ejection of core debris could not occur with the RCS at 1700 j psia. Core debris ejection at the lower pressure would be less
- severe and probably involve smaller fractions of the core. This i being the case, and since' direct heating has already been exam-ined in some detail for the high pressure scenario, the pump seal LOCA sequence was analyzed assuming direct heating did not occur.
l The assumption has been made in this analysis that as long as j there is water in the reactor cavity the debris is coolable and l does not attack the concrete basemat. In terms of thermohy-l draulic loading on containment this is probably a conservative assumption. In terms of radiological sources the assumption i delayed the onset of core concrete reactions and the radiologi-cal sources arising from that reaction. However, the effect may . not be significant, since releases into the containment from ! core / concrete interactions are inhibited by the scrubbing effect i, of the overlying water layer. l The sequence of events for the TMLB' Pump Seal LOCA scenario is shown in Table 12 and involves steam generator dryout at 4.5 l minutes followed by core uncovery at 26.7 minutes. Incipient I core melting commences at about 43.6 minutes with core slump l taking place at 60.5 minutes and vessel head failure at 68.6 { minutes. Containment leakage was assumed to be induced at'the i time of vessel breach. The reactor cavity dried out at 455 l minutes at which time a core / concrete interaction ensued. Table i 13 shows the matrix of cases analyzed for the pump seal LOCA. ! The main parameters that were varied in this sequence were the j size of the containment leak area and the presence or absence of ______ __ _ _ _ _ _ - . _ _ __.-.._ _ _ _~. _ _
- l 4 ,
i the natural convection heat transfer mode in t'he RCS fission j product transport code calculations. There were no comparable 4 BMI-2104 analysis with which to compara results, and therefore, j no calculations were done using BMI-2104 RCS releases. l In the absence of an estimate for leakage in the Bellefonte i plant, the l ance Working leakage Group (1gyrve 1 developed by the containment Perform-for the Zion plgnt was used. This curve yielded a leak area of about 1 in at 65 psia which is ! the containment pressure at vessel failure. The paraagtric leakage area that was used was arbitrary. Areas of 1 in and 1 2.4 in2 were assumed to open at the time of vessel failure and to remain constant thereafter. 5.1 CONTAINMENT LOADING In the absence of direct heating in this sequence the peak load- ) ing was less severe than was calculated in the high pressure i ejection cases. The containment pressure and temperature
! responses for a typical case in the matrix (Case-030) is shown i in Figure 40 and 41. Figure 40 shows that the peak pressure
- occurred at about 5 hours when the water in the reactor cavity l had boiled off. The smaller leak area cases -(Case-030 and 031),
as expected, showed slightly higher pressures (34). The peak
- pressures varied between about 65 and 72 psia for the four
] cases. The cases which incorporated natural convection in the I RCS (Cases-031 and 131) showed pressures that were elevated by j about 8%. The higher pressures were due to the additional fis-j sion product heating of the atmosphere. This is reflected in the containment atmospheric temperatures (Figure 41) which were j also slightly elevated for the cases with enhanced RCS releases. a i Although the pool of water in the reactor cavity was boiled off ] by about 5 hours, overflow from the containment sump kept j refluxed coolant entering the reactor cavity until about 450
; minutes. This water did not accumulate but was revaporized upon I contact with the debris. Core / concrete interaction commenced at l about 450 minutes and' hydrogen began to accumulate in the con-
- tainment atmosphere. A hydrogen burn did not occur within the
- 20 hours to which the cases were run, but the hydrogen concen-q tration at 20 hours was about 7.9% mole fraction and a burn
, would likely have occurred before 24 hours. The magnitude of ! the burn would have been similar to the one that occurred in l Case-000 and would have yielded a pressure spike on the order of { 100 psia, assuming it- was initiated at 8% hydrogen concentra-
- tion.
I 5 5.2 PRIMARY SYSTEM RADIOLOGICAL RELEASES
- Two configurations were analyzed for the TMLB' induced pump seal
- LOCA sequence using the MCT code package. The analysis was I performed both with and without implementation of the natural
! convection models in MCT. Since this sequence has higher flowrates through the primary system than the TMLB' high pressure case, natural convection has somewhat less of a role in l
-.- - . _- , - . . _ _ . _ _ _ , _ . - . _ _ . _ . _ _ _ _ _ _ , ~ _ . _ _ _ _ . - _ , _ . _ . ~ . . _ - _ _ - - .
4 I
. heating the primary system.- Nevertheless, the combination'of a higher. primary system temperatures- associated with natural l
- convection together with the quantity of flow through the system i j due to - the break, yields a striking increase in primary system i
- releases. !
i I i Figure 42 describes the primary system flow- path for this j sequence. 'The Bellefonte primary coolant system is a B&W design
- that has a set-of 8 vent valves which allow flow from the outlet I
] plenum back into the inlet plenum.- These valves were designed l l to prevent. " steam binding" and assure that ECCS flow can be j injected into the vessel during small break LOCAs. The RELAP5 1 calculations that were performed for .this sequence indicated that the loop seals did not clear so that the primary flow path l was from Lthe core into the upper plenum, to the outlet plenum,
. through the vent valves, to the inlet plenum, the cold leg
! piping, the main coolant pumps, and out the failed pump seals. 1 j Figure 43 shows the core exit gas flow history for the Pump Seal I } LOCA case. This gas flow rate history was used for both var-i lations of the Pump Seal LOCA sequence described below. The I first analysis, discussed below, was performed without modeling , i natural convection in the RCS (Type A). The second pump seal , LOCA analysis was performed for the case which includes the j effects of natural convection (Type B). The core exit gas flow during the pump seal LOCA indicates that throughout the core ! melting process there will be substantial gas flows out of the
- i. core until core collapse at approximately 3600 seconds. At that j time a large gas flow swept the primary system vapors out into j the containment.
- 5.2.1 PUMP SEAL LOCA ANALYSIS WITHOUT NATURAL CONVECTION l (TYPE "A") l Figure 44 shows the temperature response of the primary system for the Pump Seal LOCA when natural convection was not modeled.
} The hottest structure is in the upper plenum as shown in Figure i 44, and reaches a peak temperature of just over 1600 0 F at l approximately 3600 seconds. At that' time the large gas flow out ' ! of the core, which results from core collapse and consists pri-j marily of saturated steam, cools the upper plenum region. The ! steam flow resulting from core slump accounts for the rapid ! turnaround in the upper plenum temperature. The structures i downstream from the upper plenum remain relatively cool l throughout the sequence and thus substantial retention of
- volatile species throughout the primary system would- be j expected. Figure 45 shows the fission product heat in each volume of the primary system. The sweeping effect of the large gas flow at core slump occurs at approximately 3600 seconds when-a substantial drop in the fission' product heat in the upper j plenum is observed.
Figure 46 shows the cesium iodide retention factor. The behav-ior of CsOH was similar. As expected, the fission product retention in the primary system is very high. This is due to the relatively cool temperatures prevailing throughout the i
.,-_.m. _ _ _ _ . . - . , , _ . . . . . . _ , _ , _ , _. _ ,. _ _ _ _ . - _ . _ - , . . _ _ . - , . _ _ _ , - - ,
l accident sequence. Because of the almost' continuous flow of gases from the top of the core into the remainder of the primary system, cesium. iodide and cesium hydroxide are spread out throughout. the primary system with most of the retention occurring in the upper plenum. Figure 47 shows the behavior of tellurium in the primary sys-tem. The turnaround in the upper plenum retention is due to the heating up of the upper plenum. structures. The revolatilized tellurium is carried by the gas flow into the downstream volumes where it is readily chemisorbed. Because this revolatilization takes place late in the accident, tellurium is released to the
- containment in appreciable amounts commencing at approximately
+ 3600 seconds, the core collapse time, as shown in Figure 48.
! Prior to that, the tellurium release rate was very small as it was readily removed from the gas stream by chemisoption and condensation on- structures. The releases to containment of CsI i is shown in Figure 49. Like tellurium, CsI was released to the i containment throughout the accident sequence at a slow rate ! until core collapse at 3600 seconds. At that time, the releases j to containment increase rapidly as the residence times in the l primary system volumes are insufficient to remove the fission t
products from the gas flow stream. Note that CsI releases to i containment are primarily in particulate *orm (i.e., vapors
- condensed onto aerosol surfaces which have not yet been depos-ited). At 3600 seconds, these aerosols, primarily from the j upper plenum region, are entrained in the rapid gas flow and
! swept out into the containment.
1 The results of the Pump Seal LOCA analysis without natural con-vection indicate that a large fraction, in excess of 90%, of the l fission products released from the fuel during the meltdown 4 process will remain ' in the primary system after the calculated
; vessel failure. These remaining fission products, as in the case of the TMLB' sequence, will continue to heat the primary system surfaces on which they are deposited, and eventually the temperature of these surfaces will increase sufficiently.to i cause revolatilization. Once in vapor form these fission
] products may be moved from the primary system into the con-i tainment through the large opening in the bottom of the reactor 1 vessel. l 5.2.2 PUMP SEAL LOCA ANALYSIS WITH NATURAL CONVECTION
- (TYPE "B")
i This section discusnes the results of the analysis performed for
; the pump seal LOCA case in which natural convection is modeled.
l The results of the thermal analysis are shown on Figures 50 and l ! 51. As in the case of the TMLB' sequence, the effect of model-ing natural circulation is to increase the gas temperatures throughout the primary system which in turn increases the heat structure temperatures and affects the retention of fission products. In addition, the increased gas temperatures through-out the primary system result in an increase in the aerosol temperatures which in turn rapidly decrease the condensation of l _ _ _ _ . _ . _ _ _ . _ _ _ _ _ . _ . _ . - _ . _ _ _ . _ _ . _ . _ _ _ , _ _
g
)
volatiles onto aerosols. The end effect as discussed in con- , junction with the TMLB' sequence is that the volatile species, cesium iodide and cesium hydroxide, exist in vapor form rather than condensed on the surfaces of aerosols. This, combined with the continuous gas flow throughout the primary system into the containment, results in a small retention of volatile species in the primary system. I Figure 52 shows the retention factor for cesium iodide. During the early stages of the accident, up to about 3200 seconds, the volatiles are removed from the gas flow stream primarily by l means of condensation on structural surfaces. This is because i the aerosol temperatures are too high to allow significant con-densation to occur onto their surfaces. At about 3200 seconds, l the structure temperatures reach a level at which revolatiliza- , tion of the condensed materials in the upper plenum takes place and the revolatilized vapors are quickly moved into the down - l stream volumes where they condense. Eventually, at approxi- ' nately 3500 seconds, the temperature in the cold legs and the main coolant pumps increase sufficiently to revolatilize the material condensed in those volumes and combined with the large 1 gas flow occurring at 3600 seconds, the reevolved vapors are l swept into the containment. From Figure 52, it is seen that essentially all of the retention of cesium iodide at the time of reactor vessel failure, is in the cold leg inlet plenum and it amounts to less than 10%. Cesium hydroxide behavior is similar except that chemisorption plays an additional role in retaining some of the revolatilized cesium hydroxide and, therefore, the final retention of cesium hydroxide is somewhat higher, cpproxi-mately 25%. The behavior of tellurium, shown in Figures 53 and 54, is due almost entirely to chemisorption. Because of the substantial gas flow through the primary system and the high temperatures in the upper plenum and outlet plenum, essentially no tellurium was retained there. The cold-leg inlet plenum and the cold-leg piping retained most of the tellurium with a small fraction retained in the main coolant pumps. From approximately 3600 seconds until vessel failure there is a large flow of gases throughout the primary system which carries the romainder of the released tellurium from the fuel matrix through the primary sys-tem at such rate that the residence time of tellurium in any given volume is insufficient for significant removal. The difference between the pump seal LOCA case without natural convection and this case is due to the increased gas tempera-tures in the primary system combined with the relatively large gas flow rates through the primary system into the containment. The increased gas temperature has the effect of keeping the volatile species in vapor form whereas the significant gas flow rates have the effect of carrying these vapors through the pri-mary system at a rate which results in a short residence time in i the primary system and, thus, results in low removal rates by condensation or chemisorption. In addition, the high gas tem-paratures increase the heat structure temperatures to a level at which the condensed volatiles renvolve back into vapor form and l are carried out of the primary system into the containment. The l
j i I main difference between the first and the second case (type "A" verses type "B" releases) is a drastic decrease in the retention t of both cesium iodide and cesium hydroxide by the primary sys- !
- tem. This difference was not as noticeable in the TMLB' cases
{ because for most of the sequence duration, the gas flow rates j through the primary system and into the containment were quite
- low. In both, the second TMLB' case and the second pump seal
, LOCA case the primary form of the volatiles was the vapor form.
- Because in the TMLB' case the gas flow rates were so much lower
- than in the pump seal LOCA case,_the residence times in the i TMLB' case were considerably longer and therefore condensation and chemisorption of volatiles was more significan# in removing-these species from the gas-stream. In the pump seal LOCA case, as stated before, the gas flow rates were large and, therefore, resulted in insufficient residence times in the primary system volumes to allow significant removal of the volatiles fission
{ products from the gas stream by condensation and chemisorption. lt A summary of the primary system releases of the volatile' fission-
- products for the pump seal LOCA sequences is presented in Table-
} 14. The Type A releases (without natural circulation) are on j the order of 7 to 8% of initial inventories for CsI and CsOH with CsOH releases slightly less than-that for CsI due to the , small contribution of chemisorption for the former. Tellurium ! Type "A" releases were about 9% of inventory. The Type "B"
! releases as discussed already were markedly high for Cs! and i CsOH, 93% and 77% respectively, primarily due to the much higher l } RCS temperatures and relatively constant gas flows that tended j to sweep these materials into the containment. Due to a high ! removal rate by chemisorption the primary system releases of j tellurium were only about 50% greater (13%) than for the Type l "A" releases. < i 5.3 CONTAINMENT AEROSOL TRANSPORT AND RELEASES l
- The same procedure that was used for analyzing the fission prod-i uct transport, deposition, and release from containment for the 8
TMLB' high pressure sequences was also employed for the pump
; seal LOCA sequences (see Section 4.4). The case matrix, Table 13, consists of four cases with the parameters being the primary system releases (Type "A" or "B", Table 14) and the magnitude of the assumed containment leak areas, 1 in and 2.4 in l Case-030 as indicated in Table 13 utilizes the MCT calcylated l Type 'A' primary system releases together with a 1 in leak area in the containment boundary. The mean particle size for the MCT calculated RCS releases was about 5 microns and these relatively large aerosol particles do not remain in the atmo-
! sphere for long periods. This is clearly seen in Figure 55. The peak mass of CsI in the containment was about 4.5 lbs, but was quickly removed primarily by settling within an hour, and the accumulated releases to the environmont at 20 hours amounted to only about .004 lb. (Figure 56). The release of tellurium from the primary system was quite small for this case reaching only about 5 lbs in containment. Releases from the core / con-crate interaction which commenced at about 8 hours produced
~
significantly larger Te aerosol concentrations in containment and a total release to the environment of about 0.24 lbs. The Type "B" RCS releases for CsI and CsOH were-an order'of magnitude greater than the Type "A" releases and this fact is 4 clearly distinguished in the graphs of airborne concentration i for these species, Figures 57 and 58 (Case-031). The contain-ment releases also reflect the high releases from RCSf being about .06 lbs at 20 hours for CsI. Type 'B' Teeraleases from j the primary system were not significantly different from Type "A" releases, and in any case the core / concrete interaction releases of Te tended to swamp the RCS releases with the result i
- that the containment releases for To were not significantly l l
changed from case-030. l 1 l The effects of a larger leak area, as expected, showed an ' increase in the quantity of fission products-released to the ) environment. Figure 59 shows the CsI concentration for Case-131 which incorpogates Type "B" releases along with a larger leak area (2.4 in ) and, therefore, represents the worst casa,(con- i j sidered here) for the pump seal LOCA sequence. The releases of
; volatiles for this case are a factor of about 2.4 times' greater 1
than for the 1 in2 leak area case. Thus, for these small leak l - areas the total releases appear to be nearly proportional to the leak area.
! A summary cf the containment releases.for all the cases in the j pump seal LOCA matrix is given in Table 15. In general, for the
- same RCS release type ("A" or "B") the containment leakage for
! each fission product group was nearly proportional to.the leak j area. The CsI and CsOH containment releases for the two RCS release types were, of course, in nearly the same ratio (factor of 10) as the RCS releases. The tellurium releases which were dominated by core / concrete releases were essentially the same, , while rutherium releases were higher for the Type "B" by about 70%. The Sr and La groups, na expectedi' remained unchanged , between the two RCS release types as their'RCS releases are 7 1 governed by their release from the fuel matrix and not by their transport through the primary system. , An absence of a mechanism for early containment failure or for H
, larger induced leakage areas early in the accident makes:the l l containment releases for the pump seal LOCA quite small compared to those for the direct heating scenario. The releases are, in fact, similar to those of Cases-000 and 100 in the TMLB' high pressure scenario which were assumed not to have direct heat-ing.- It should be recalled, however, that;the primary system '
pressure at vessel failure as calculated by MARCH was about:1700 psia for the Pump Seal LOCA. This pressure is probably high enough to induce direct heating with the result that this , sequence could have results very much'like some of the TMLB' l high pressure cases discussed in section 4.0. l
'l
l 5
*l l c ;
g - a I,-_.. . _ _ , . , . .___..m,. . . . _ _ , , _ _ _ . _ , _ - . - , , , - . , , , , . . , _ . _ . -_ ,m, ,. _ . ~ , , , . . _ _ _ , ,
i
6.0 CONCLUSION
S A number of conclusions can be drawn from the results of this ' analysis. A principal conclusion is that even for large, dry
- containments, direct heating probably represents one of the most severe threats to containment integrity, and, since it has the
! potential for failing the containment early and at a time when fission product concentrations are at or near maximum levels, it probably also represents one of the most severe radiological
. releases to the environment. The magnitude of the threat i presented by direct heating may require the generation of a set
' of operator action guidelines in an attempt to mitigate the problem. Possible operator action might include a timely depressurization of the primary systems before core slump. , There is probably a small window between the time at which it ; becomes apparent that core meltdown cannot be prevented and the
- time at which RCS venting must commence in order to prevent direct heating. Since a critical decision which has to be made i during a short peticd of time would require very accurate knowledge of conditions in the RCS and particularly in the core, such a decision capability may require additional in-core i instrumentation. It is not within the scope of this report to !
make recommendations regarding strategies for preventing direct containment heating, but there are some points that can be made
- without recourse to detailed studies. Clearly, the information regarding water level, core temperatures, and vessel pressure i that are available during normal operation are even more
, critical during accident conditions. An obvious strategy for obtaining critical parameters during severe accidents is to
" harden" the existing instrumentation systems by extending their range into the severe accident regime as, for example, employing
) thermocouple that function and are reliable at higher temper-atures. Locating temperature measurements in the core in such a j manner as to yield information about the extent of core damage l would also be advantageous. Finally, the addition of a manual t depressurization system to PWRs in general could be considered.. . Calculated estimates of the primary system temperature history ! using the MCT code package indicate that primary system tam-peratures are well into the regime in which failure would be expected to occur at the elevated pressures characteristic of l the TMLB' sequence (refer to Figure 22). If a failure of the
~
RCS boundary occurred of sufficient magnitude to depressurize the RCS in advance of core slump, direct heating could be l eliminated or mitigated. A preliminary evaluation of the natural circulation model has indicated that it is overpre-dicting the transfer of heat from the core to the downstream volumes in the RCS. It is uncertain whether a more realistic heat transfer rate would significantly change the primary system, heatup characteristics. It is likely, however, that the RCS heatup would be somewhat slower, delaying the time at which failure might occur and, thus, limiting the mitigating effect it would have on the direct heating pulse. The effect of natural circulation flows in cooling the core, which has been neglected l in this analysis, will tend to delay core slump and supply some
additional. time for RCS failure to. occur before core slump. Taken together, the results of this analysis seem to suggest a high probability for a breach in the RCS before core slump and, thus, a low probability for the events. Several recent analyses g{regga
- ggytingsequenceof using multi-di-mensional codes also tend to substantiate this conclusion.
- In terms of primary system releases, the incorporation of models
, for natural convection between control volume in the RCS and post-vessel-failure -fission product heatup and reevolution of volatiles has suggested that the releases of these materials from the primary system could be significantly higher than has
, been estimated in -previous analyses. There are competing fac-tors operating which affect the calculated transport of fission
! products through and out of the RCS. The previously mentioned I overprediction of natural convection heat and mass transfer in the nCS .tends to move fission products through the system more rapidly than they would actually be moved, while the underpra-a dicted heat structure surface temperatures associated with the j lumped parameter models . tend to move materials through the system more slowly than they actually would be moved. These l effects would obviously tend to counteract each other. In any event, there is considerable range for improving the models in the existing analytical tools. The magnitude of the releases to the environment clearly depends i very strongly on the size of the pathway leading to the environ- } ment. For direct heating events involving more than about 50% of .i the core debris there exists a high probability of an induced
- gross failure of the containment boundary. For direct heating
- that does not result in gross containment failure but which
- induces leakage pathways,s it is presently not possible to predict the magnitude of the leakage as a function of the magnitude of the containment loading. The work that has been conducted to date to estimate containment leakage rates have involved the situation in which there is a gradual increase in
', containment loading conditions and have not addressed the situation in which the loading is of a pulsed nature as is the case in the direct heating or hydrogen burn situation. Additional work may be required in this area in order to characterize induced leakage resulting from pulsed loading. i j _ _
i f REFERENCES
- 1. A. M. Kolaczkowski, et al., Interim ~ Report on Accident Secuence Likelihood Reassessment, SAI Report, Aug 1983.
l 2. J. E. Brockman, W. W. Tarbell, " Aerosol Source Term in High Pressure Malt Ejection," Nuclear Science'and Encineerina, V. 88, pp. 342-356,.1984. .
- 3. M. Pilch, W. W. Tarbell, J. E. Brockman, Influence of Reactor Geometry On The Behavior of Discersed Debris, 1 SAND-1725C, The 13th Water Reactor Safety Meeting,
; Gaithersburg, MD, Oct 1984.
- 4. R. J. Lipinski, et al., Uncertainty in Radionuclide Release l Under Specific LWR Accident Conditions, SAND 84-0410, V. 2, Feb 1985.
- 5. P. P. Bieniarz, Source Term Analysis of Fit patrick Power Plant, Risk Management Associates, Albuquerque, NM, Report Under Review, Jan 1986.
- 6. J. A. Giesake, et al., Endionuclide Release Under Specific
;^ LWR Accident Conditions-PWR-Larce, Dry Containment desian (Zion Plant), BMI-2104, V. 6, . Battelle Columbus Laboratories, July 1984.
l
- 7. C, A. Dobbe, R. Chambers, Analysis of a Station Blackout
; Transient for the Bellefonte Pressurized Water Reactor, EGG-NTP-6704, Idaho National Engineering Laboratories, Oct 1984. ! 8. R. J. Dallman, Transmittal of Bellefonte TMLB' Calculations j with Pumo Seal Leakaae, Letter of Transmittal-RJDS-12-85, EG&G, Idaho National Engineering Laboratories, May 20, 1985.
- 9. C. J. Shaffer, Marcon2.OP: A Version of MARCH With State-of-the-Art Core Concrete Modelina, NUREG/CR4-5779, Sandia National Laboratories, Feb 1986.
- 10. K. D. Bergeron, D. . C. Williams, "CONTAIN Calculations of Containment Loading of Dry PWR's," Sandia National Laboratories, Nuclear Science and Encineerina, V. 90, pp.
153-159, 1985. ! 11. D. K. Gabrielson, et al., ONEDIM-A Computer Code for Solvina Non-Linear Heat Transfer Problems, Report No. i SCL-DR-69-99, Sandia National Laboratories, Feb 1970.
- 12. D. A. Powers, J. E. Brockman, A. W. Shiver, Vanesa: A-Mechanistic Model- of Radionuclide Release and Aerosol Generation Durina Core Debris Interaction with Concrete, l NUREG/CR-4308, SAND 85-1370, Sandia National Laboratories, ,
l July.1985. l
I References (Cont.)
- 13. K. D. Bergeron, et al., Users Manual for CONTAIN1.0. A Comouter Code for Severe Nuclear Reactor Accident Containment Analysis, NUREG/CR-4085 SAND 84-1204, Sandia National Laboratories, May 1985.
4
- 14. J.- Jung, Ultimate Strenoth Analyses of the Watts Bar. Maine Yankee, and Bellefonte Containments, NUREG/CR-3724, SAND 84-0660, Sandia National Laboratories, July 1984.
- 15. Containment Performance 'Workina GrouD, Draft Report, NUREG-1037, May 1985.
- 16. F. Gonzalez, I Huhtiniemi, M. Corradini, Transient Model for Direct Heatina in a LWR, UWRSR-35, University of Wisconsin, Feb 1986.
- 17. Lichung Pong, M. Corradini, G. A. Moses, Eurry S 2 D Severe Accident Containment Loads Calculations UsiMa HMC, UWRSR-34, Jan 1986.
, 18. Reactor Safety Studv--An Assessment of Accident Risks in U.S. Commercial Nuclear Power Plants, WASH-1400, NUREG-75/014, Oct 1975.
- 19. W. G. Brown, K. R. Solvason, " Natural Convection TGhrough Rectangular Openings in Partitions-1, Vertical Partitions",
International Journal of Heat and Mass Transfer, V. 5, pp. 859-868, 1962.
- 20. W. G. Brown, A. G. Wilson, K. R. Solvason, " Heat and Moisture Flow Through Openings by Convection,"
ASHRAE-Transactions, V. 69, p. 351-357, 1963.
- 21. B. C-J. Chen, et al., " Degraded Core Studies _Using the l Multidimensional COMMIX Code", Transaction of the American Nuclear Society, Annual Meeting, Boston, MA, V. 49, pp.
453-454, June 1985.
- 22. R. Henninger, " Preliminary 2-D MELPROG Calculation for the
! TMLB Accident in Surry," . Presented at Direct Containment-I Heating Meeting, Held at Bethesda, MD, April 23, 1986.
- 23. V. Shah, " Structural Failure Studies of RCS," Presented at Direct Containment Heating Meeting, Held at Bethesda, MD, April 23, 1986.
4 i L - . - - -.- - - .., - -
TABLE 1 TMLB'-HPE SEQUENCE OF EVENTS (MARCH 2.0 CODE) STEAM GENERATOR DRYOUT............................... 5.83 MIN CORE UNCOVERY........................................ 30.0 MIN CORE STARTS MELTING.................................. 53.4 MIN CORE SLUMP........................................... 75.8 MIN FRACTION CORE MELTED............................ 65% FRACTION CLAD REACTED........................... 50% THERMAL ATTACK ON BOTTOM HEAD........................ 75.9 MIN BOTTOM HEAD FAILURE.................................. 82.1 MIN DIRECT HEATING EVENT............................. 82.1-82.6 MIN CONTAINMENT INDUCED LEAKAGE OR FAILURE............... 82.6 MIN 4
l TABLE 2 CASE MATRIX TMLB' HIGH PRESSURE EJECTION
% CAVITY BREAK RCS F.P. NAT. CORE / CONC )
CASE INJECT WATER AREA REL. REL. CONV. REEV. RELEASES i 000 0 -- 1 in 2
- NO NO NO YES 100 0 --
2.4 in 2
- NO NO NO YES 001 90 NO 7 ft 2
- NO NO NO NO 011 90 NO 7 ft 2 MCT YES NO NO -- NO 001A 90 YES 7 ft 2
- NO NO NO NO Olla 90 YES 7 ft 2 MCT YES NO NO NO 011B 90 YES 7 ft 2 MCT YES YES NO NO 002 50 NO 5 in 2
- NO NO NO NO 102 50 NO 12 in2
- NO NO NO NO 012 50 NO 5 in2 MCT YES NO YES NO 112 50 NO 12 in2 MCT YES NO YES NO 002A 50 YES 5 in2
- NO NO NO NO 102A 50 YES 12 in2
- NO NO NO NO 012A 50 YES 5 in2 MCT YES NO YES YES 112A 50 YES 12 in2 MCT YES NO YES YES l
*BMI-2104 ZION RELEASES 4
l
f TABLE 3 HIGH PRESSURE EJECTION DIRECT HEATING EVENT ASSUMPTIONS CORE DEBRIS EJECTED (UO2, Zr, ZrO 2 ).................. 100% AVAILABLE STEEL EJECTED..................... 53% (100,000 lbs) INITIAL'$sBRIS TEMPERATURE........... (CORIUM EUTECTIC) 25500K ZIRCONIUM REACTED IN-VESSEL............................... 50% ZIRCONIUM REACTED IN DIRECT HEATING EVENT. 100% OF INJECTED Zr DEBRIS COMES TO THERMAL EQUILIBRIUM WITH ATMOSPHERE ADIABATIC CALCULATIONS PARAMETERS VARIED FRACTION OF DEBRIS INJECTED INTO CONTAINMENT 90% 50% 0% STEEL REACTED 0% 100% (OF THAT INJECTED) HYDROGEN REACTED i 0% 100% (OF THAT IN CONTAINMENT, 3 VOL. %) CAVITY WATER INJECTED WITH DEBRIS 0%
- 100% (75,000 lbs) 1 l
l 1
l I TABLE 4 RESULTS OF DIRECT HEATING CALCULATIONS
% %* H Fe T P H CASE EJECT. WATER BUkN BURN (sF3 co.I.3 cGa3 001 90 0 NO NO 2009 159.9 281 001 90 0 YES NO 2288 178.0 330 001A 90 100 NO NC' 1462 148.8 305 001A 90 100 YES NO 1719 168.8 356 002 50 0 NO NO 1314 116.8 172 002 50 0 YES NO 1681 138.6 236 002 50 0 NO YES 1913 158.4 265 002 50 0 YES YES 2223 173.8 318 002A 50 100 NO NO 789 96.6 181 002A 50 100 YES NO 1111 121.6 239 002A 50 100 NO YES 1326 138.2 213 ,
1 002A 50 100 YES YES 1609 160.1 334 ! 1
* % OF WATER IN REACTOR CAVITY THAT IS INJECTED INTO CONTAINMENT WITH THE CORE DEBRIS I
f TABLE 5 i INITIAL FISSION PRODUCT GROUP INVENTORIES Group 1 Mass (Kc) CsI 27.90 RbI 3.97 Total Group 1 31.87 Group 2 CsOH 180.16 RbOH 21.46 Total Group 2 201.62 Group 3 Te 33.3 Group 4 Sr 63.8 Ba 81.4 Total Group 4 145.2 Group 5 Ru 138.0 Rh 27.9 Pd 70.0 Tc 49.4 Mo 206.0 Total Group 5 491.4
TABLE 5 (Cont.) INITIAL FISSION PRODUCT
-GROUP INVENTORIES ' Group 6 Mass (Ka)
La 82.9 Y 30.4 Eu 11.9 Nd 227.0 Np 34.5 Sm 45.2 Pm 9.6 Pu 624.0 Zr 238 Ce 175 Nb 3.7 Pr 67.6 Total Group 6 , 1549.8 Group 7 Kr 17.8 Xe 146.0 Total Group 7 353.8 i l l l l
i TABLE 6 DIRECT HEATING FISSION PRODUCT REgEASES TMLB'-HIGH PRESSURE EJECTION ' i FISSION 50% INJECTION (Kg) 90% INJECTION (Kg) PRODUCT GROUP ** FINE COARSE FINE COARSE . (.7 micron) (30 micron) (.7 micron) (30 micron) Te 6.75 .11 12.1 .20 Sr --
.59 --
1.06 Ru 61.6 2.34 111 4.20 Mo 83.5 *** 150 *** La -- 7.69 -- 13.8
- Calculated using fission product inventories remaining in the fuel at vessel failure in the BMI-2104 Zion analysis.
- Molybdenum has been broken out separately here. It is actually a member of the Ruthenium group
- Included in Ruthenium group total
- Molybdenum has been broken out separately here. It is actually a member of the Ruthenium group
TABLE 7 DIRECT HEATING FISSIOM PRODUCT RE, LEASE TMLB'-HIGH PRESSURE EJECTION FISSION 50% EJECTION (Kg) 90% EJECTION (Kg) PRODUCT GROUP ** FINE COURSE FINE COURSE (.7 micron) (30 micron) (.7 micron) (30 micron) Te 3.31 0.52 5.96 .10 Sr --
.611 --
1.10 Ru 61.8 2.44 111 4.40 Mo 86.4 *** 156 *** La -- 7.75 -- 13.9
- Calculated using fission product inventories remaining in the fuel at vessel failure in the present study
** Molybdenum has been broken out separately here. It is actually a member of the Ruthenium group
*** Included in Ruthenium group total I
4 1 TABLE 8
SUMMARY
OF PRIMARY SYSTEM RELEASES + FOR TMLB' HIGH PRESSURE EJECTION CASES MCT WITHOUT NATURAL CONVECTION MCT FISSION MELTDOWN REEVOLUTION W/ NATURAL PRODUCT BMI-2104 PHASE ** PHASE *** TOTAL CONVECTION ** TYPE A TYPE C TYPE B CsI .02* .028 .21 .24 .35 CsOH .02* .030 .18 .21 .28 Te .04* .24 ---
.24 .17 + FRACTIONS OF INITIAL INVENTORIES
- ZION STUDY
** UP TO AND INCLUDING VESSEL FAILURE *** RELEASES SUBSEQUENT TO VESSEL FAILURE - CALCULATED USING THE DEPRESSURIZATION HISTORY FOR CONTAINMENT LEAKAGE SCENARIOS I
Table 9 PARTICLE SIZE DISTRIBUTION FOR PRIMARY SYSTEM RELEASES TYPE 'F. P. HEATING REEVOLUTION NAT. DIAMETER STAN. DEV. , CONV. (MICRONS) (MICRONS) A Yes No No 4.58 1.76 B Yes No Yes 4.52 2.01 C Yes Yes No 5.60 1.98 l l l 4 7 i
TABLE 10 , CONTAINMENT RELEASES * (TO 20 HRS) TMLB' HIGH PRESSURE EJECTION CASE CsI CsOH Te Ru Sr La Xe Kr 000 5.7E-5 9.8E-5 .018 2.lE-5 .0054 .0032 .071 .071 100 1.3E-4 2.3E-4 .043 4.9E-5 .013 .0077 .16 .16 001 .014 .014 .29 .38 .024 .0052 .77 .77 011 .015 .016 .13 .39 .0058 .0052 .77 .78 001A .014 .014 .29 .38 .024 .0053 .78 .78 Olla .015 .016 .30 .40 .0054 .0048 .79 .79 OllB .16 .12 .21 .41 .0073 .0048 .79 .79 002 8.0E-4 8.0E-4 .0097 .012 .0012 4.0E-5 .31 .31 102 .0018 .0018 .022 .028 .0027 9.4E-5 .59 .59 012 .0061 .0056 .010 .012 8.5E-5 3.7E-5 .31 .31 112 .013 .012 .023 .028 2.0E-4 8.8E-5 .60 .60 002A 7.2E-4 7.2E-4 .0088 .011 .0011 3.8E-5 .30 .30 102A .0017 .0017 .020 .025 .0025 8.9E-5 .57 .57 012A .0101 .0092 .0092 .011 8.5E-5 3.5E-5 .30 .30 112A .022 .02 .021 .025 2.3E-4 8.5E-5 .53 .53
- FRACTION OF INITIAL INVENTORIES RELEASED TO ENVIRONMENT l
TABLE 11 COMPARISON OF MULTI-CELL-REyASES TO SINGLE-CELL RELEASES CASE-012A FISSION PRODUCT ATTENUATION FACTOR , GROUP NO BUOYANCY BUOYANCY , CsI .77 .90 CsOH .75 .88 Te .99 .92 1 SI .90 1.07 i Ru 1.17 .97 La 1.17 1.30 Xe .95 1.03 Kr .95 1.03 i " CUMULATIVE RELEASES AT 15.5 HOURS I i i {
- l. ,
TABLE 12 TMLB'-PSL SEQUENCE OF EVENTS (MARCH 2.0 CODE) STEAM GENERATOR DRYOUT.............................. 4.50 MIN CORE UNCOVERY....................................... 26.7 MIN CORE STARTS MELTING................................. 43.6 MIN CORE SLUMP.......................................... 60.5 MIN FRACTION CORE MELTED...... .................... 65% FRACTION CLAD REACTED.......................... 50% THERMAL ATTACK ON BOTTOM HEAD....................... 60.9 MIN BOTTOM HEAD FAILURE................................. 68.6 MIN CONTAINMENT LEAKAGE INDUCED......................... 82.6 MIN REACTOR CAVITY DRYOUT............................... 454.5 MIN i. I 1 I l l l l l I l
TABLE 13 CASE MATRIX , TMLB' PUMP SEAL LOCA
% BREAK RCS F.P. NAT. CORE / CONC CASE INJECT AREA REL. REL. CONV. REEV. REL 030 0 1 in 2 MCT YES NO NO YES 130 0 2.4 in2 MCT YES NO' NO YES 031 0 1 in2 MCT YES YES NO YES f 131 0 2.4 in2 MCT YES YES NO YES
\
I i 1 .{ I .. _ , _ - . .
TABLE 14
SUMMARY
OF PRIMARY SYSTEM RELEASES + FOR TMLB' PUMP SEAL LOCA CASES MCT WITHOUT NATURAL CONVECTION
, MCT FISSION MELTDOWN REEVOLUTION W/ NATURAL PRODUCT BMI-2104* PHASE ** PHASE *** TOTAL CONVECTION **
TYPE A TYPE C TYPE B CsI ----
.084 ---
.084 .93 CsOH ----
.073 ---
.073 .77 Te ----
.089 ---
.089 .13
+ FRACTIONS OF INITIAL INVENTORIES
- ZION STUDY
** UP TO AND INCLUDING VESSEL FAILURE
*** RELEASES SUBSEQUENT TO VESSEL FAILURE - CALCULATED USING THE DEPRESSURIZATION HISTORY FOR CONTAINMENT LEAKAGE SCENARIOS l
l I TABLE 15 CONTAINMENT RELEASES * (TO 20 HRS) TMLB' PUMP SEAL LOCA CASE CsI CsOH Te Ru Sr La Xe Kr 030 6.1E-5 1.2E-4 .0032 1.1E-6 .0049 8.2E-4 .070 .070 130 1.4E-4 2.9E-4 .0078 2.6E-6 .012 .0020 .16 .16 031 8.8E-4 8.1E-4 .0034 1.9E-5 .0050 8.4E-4 .072 .072 131 .0020 .0019 .0082 8.9E-6 .012 .0020 .16 .16
- FRACTION OF INITIAL INVENTORIES RELEASED TO ENVIRONMENT 2
l l
CONTAIN CODE m
' m A
/
}
I
$f I f
/
/
/
I $1 EI l l I I y
/
I
/
l dj! I 2 MCT CODE / ,I
'l l
4 I ~ I i
/ / / I
( 21 / } I
$! $lE! ' 8
si
=
p/ ei g~p;
//
fr n
- gl+ DIRECT HEATING /
y$
/
f // RELAPj / / '
/
- A MARCON (BOIL-HOTDROP) HAND CALCULATIONS n- 4 d = .
s' . DEBRIS 6N FLOOR - CONCRETE 4 DEGASSING m l l l TIME " CORE VESSEL UNCOVERY BREACH Figure 1. Diagram of Code Linkages
I 120 i i ; i i 100 - 80 - m e W 60 - cn W cn E CL, 40 - 20 - 0 ' ' ' ' ' O 4 8 12 16 20 24 TIME (hr) Figure 2. Containment Pressure TMLB' without Direct Heating (Case-000)
1200 3 I I I . l 1000 - E 800 - L. _ . un m 3 e
< 600 -
m _ m O. I w 400 - 200 0 ' ' ' ' I
; O 4 8 12 16
- 20 24 TIME (hr) i
, Figure 3. Containment Temperature TMLB' without Direct Heating (Case-000)'
i 4 4
7 6 VOL = 2262.2 ft3 MCP = 29735.6 Blu/F CONTAINMENT PRESSURIZER AR = 55.2 ft2 AH = 1079.1 ft2 DELX = 0.445 ft o 5 VOL = 45.32 ft3 MCP = 1274.5 Blu/F PRESSURIZER AR = 0.683 ft2 SURGE LINE AH = 185.3 ft2 DELX = 0.117 ft o 4 VOL = 157.1 ft3 HOT LEG MCP = 3013.8 Blu/F TO AR = 7.88 ft2 SURGE LINE AH = 198.4 ft2 DELX = 0.24 ft n 3 VOL = 179.8 ft3 { MCP = 4693.3 Blu/F i OUTLET PLENUM AR = 23.44 ft2 AH = 313.25 ft2 DELX = 0.25 ft o STRUCTURE 1 - MCP = 3525 Btu /F VOL = 969.9 ft3 2 - MCP = 5164 Btu /F 2 UPPER PLENUM MCP = 3525 Blu/F 3 - MCP = 952.6 Blu/F AR = 86.9 ft2 4 - MCP = 3639.7 Blu/F 4 g gg A = 1817112 l l I VOL = VOLUME 1 MCP = MASS x Cp OF VOLUME AR = FLOW AREA CORE AH = HEAT TRANSFER AREA DELX = STRUCTURE THICKNESS Figure 4. Primary System Flow Path TMLB' High Pressure Ejection l l l 68
--g.- --- - . , - --r c.---,r..--, w.r,. w r,%-y ye-.--y. w- .- - - , . - - - . - -
Pts I1 n V O f h, l H ko' 1
#0,To i
Figure 5. Description of Natural Convection Model l l l 69
160 I l l 140 - O N U(1) _ O NU(2) 120 - tr y 100 - 2 3 Z p 80 -
. .a o W m
m 3 60 - Z . 40 - 20 _ I I I I i 0 0 10 20 '30 40 50 60
~
TEMPEbATURE DIFFERENCE ( F) Figure 6. Nuscelt Number Correlations for Natural Convection J s or
' ~ 2560 i I i I .I I I I 2540 -
~
$e m 2520 -
a - m m
- m
., -a E E 2500 - 4 - 3 E 2480 - 2 I J i I I I f l
! 26 t I I 40 45 50 55 60 65 70 75 80 85 TIME (min) i Figure 7. Primary System Pressure TMLB'. High Pressure Ejection i
1 i
5000 I 3 I I I I i i s = 4000 - 2 L , m m e m M 3000 - m m
-a E N
>. 2000 -
m 2 g s 1000 -
- J' '
0 I I ' I I I ' ' 40 45 50 55 60 65 70 75 80 85 TIME (min) Figur.e 8. Average Core Temperature TMLB' High Pressure Ejection m.
0.8 I I I I I I I i 0*7 - CLAD REACTED .
............. CORE MELTED .i -
0.6 - u) : z 0.5 - O : - n ./ 0 : 4 0.4 - : a E . w E / - ul : E : O .# 0 0.3 -
/ -
0.2 - i , 0.1 - 0.0 I ' ' ' ' ' ' 40 ' 45 50 55 60 65 4 70 75 80 85 TIME (min)
- t Figure 9.
Core Melt and Clad Oxidation Fractions TMLB' High Pressure Ejection 4
. 1200 I I ! I l l I I i 1000 - 3 C 800 - O tal O 3 4
^ O O 600 g - -
n. J
$ 400 - -
F-200 - - I ' ' ' ' ' I 0 40 45 50 55. 60 65 70 75 80 85 TIME (min) Figure 10. Hydrogen Generation Rate TMLB' High Pressure Ejection
, ' l ]3 ! , ) ><
- l ,
una 04m E.a O3 03> om >Im OOEm c. E. 1 2 4 8 0 _ 2 0 0 0 H 0 0 _ 4 - b F i g 2 u 8 I r e 1 1 - O u T 32 I I t I l M e E _ t S _ I G N a C s E F S - l o T 36 i I w A . R . T T M L O . B F A H C g i C 4 i I 0 I h D E . P N r T e ( s s s u x - r 1 C e 02 4 i O
)
4 R i j E E e S . c L t U n i o \MP V 4 i E 8 S i S E L . F . A I L U R 5 _ - - - E 2
4200 g g i l l l l 1 CORE SLUMP n 3600 - E L m m 3 3000 - z w (L 2
$ $ 2400 -
e 0 b w 1800 - m a ir r 1200 - l VESSEL FAILURE l 600 I I I I I ' ' ' 40 45 50 55 60 65 70 75 80 85 TIME (min) Figure 12. Outlet Gas Temperature TMLB' High Pressure Ejection
2500 g , , CORE SLUMP I O UPPER PLENUM E OUTLET PLENUM 2000 - O HOT LEG TO SURGE LINE e SU'1GE LINE ~ Q PRESSURIZER C 2 G w m 1500 - )1 - 3
<i:
- n. m 2 '
w F 1000 - Qs - 0 - n r' . -n-V v v U v 500 - VESSEL FAILURE - I I I I I I 0 24 28 32 36 40 44 48 52 TIME SINCE START OF ACCIDENT (s x 102) Figure 13. Primary System Gas Temperatures TMLB' High Pressure Ejection (Case A)
2 _ 5 _ n _ E _ R U e - _ L r . I 8 u - l _ A I 4 s F s a L e P E r S P M S U E h L V g S i E ) H R I 4 2 l O 4 0 C 1 'B x L s M ( T T s N e E r D D u 0 I t l I 4 C a C r e A p F m O e T T e R r 6 A u i I 3 TS t c E u
~v C r t)
NI SA _ S me _ E es M ta E 2 I sC v I 3 T y( NI " S L n E yo G ri MR at MU U mc ie UN S R rj NEOE E Z C PE _ EL LP TNI 8 _ I R P T GL U I 2 RE E L L EE S GS 4 _ PTTRE 1 _ PUOUR e UOHSP Q r 0NOGQ u g i
- - - - 4 F 0 0 0 0 0 O2 0 0 0 0 0 0 6 2 8 4 2 1 1 C L w E D F < E w n. 2 s m E D F o D N u y < $
=
E I i i i l l 0 UPPER PLENUM E OUTLET PLENUM O HOT LEG TO SURGE LINE 30 - e SURGE LINE - Q PRESSURIZER e E M w E 24 - 2. E t' I-4 uJ I p 18 - 0 3 O y O e E EL Z 12 - 9 m 'j m
- u. VESSEL FAILURE
_ CORE SLUMP _ n o 24 28 ' 2 36 40 52 TIME SINCE START OF ACCIDENT (s x 102} Figure 15. Fission Product Heating TMLB' High Pressure Ejection (Case A) i i i
1.0 I l l l l
@ OVERALL O UPPER PLENUM E OUTLET PLENUM
' 0.8 - O HOT LEG TO SURGE LINE 9 SURGE LINE - Q PRESSURIZER . m O l-O
< 0.6 -
u. 2 9 H ca e Z w g 0.4 - O 0.2 - VESSEL FAILURO - CORE SLUMP o o 0'0 " ' 28 32 " 36 40 44 48 52 TIME SINCE START OF ACCIDENT (s x 102) Figure 16. CsI Retention Factor TMLB' High Pressure Ejection (Case A)
i 1.0 I I I I l
@ OVERALL
- O UPPER PLENUM l
0.8 - 5 OUTLET PLENUM O HOT LEG TO SURGE LINE 9 SURGE LINE Q PRESSURIZER x O i F-i 0
< 0.6 -
- u. -
Z 4 O P i co
- h VESSEL FAILURE g 0.4 -
J g e ji i i CORE SLUMP 0.2 - i e i
- 0.0 1
. I I l y y y
) 28 32 36 40 44 48 52 TIME SINCE START OF ACCIDENT (s x 102) Figure 17. Te Retention Factor TMLB' High Pressure Ejection (Case A) i i
i l l l 35 i i i i i I , V CONDENSED ON AEROSOLS V CONDENSED ON STRUCTURE WALLS l ' i 30 - t
! 25 - -
^
cn i O w Z 20 - - w
$ E m ! 15 i m i 7
~ o 10 - - I VESSEL FAILURE i
- CORE SLUMP 5 - -
o o d I I l - I l 0 - 28 32 36 40 44 48 52 56 TIME SINCE START OF ACCIDENT (s x 102) j Figure 18. CsI Mass Retained TMLB' High Pressure Ejection (Case A) i i
lll l 0
- - - - 6
- , 6 ) 5 A e s a C ( n 2 o 5 i
)
t 2 c 0 e 1 j n x E E s R ( e U r 8 u 4 T
,L '
I s A F N s E e L E D r S I P n C S E P C h V MU A g i
, L 4 F H I
S 4 O E ' R T B O R L C A M T T y S d E e _ 0 C n - 4 N i S I a E S t L e C I S E R TL M s R L I AA T s PW a
, 6 M NN I 3
OOD e DD EB " T EE R SS NN O . EE I S 9 DDM 1 NNE 2 e OOH 3 r CCC u _ VYO i g _ F .
^ _
- - - - 8 0 8 6 2 .
4 2 0 n
^c5 Oa<>um suylE$
t m<a
I8 Cal MASS RELEASED TO CONTAINMENT (kg) M O O O O O a g nO
~ ->
m u - O c, e l I I I l} >E g o <m 1 :D O -4
$ U "N
- m -
c - x 5 8 d "
" E m m o m l}
w - m ZW U g o* m
- c. m :
-4 8 >
% 2 w~
O 6 - m mO -
- e. )
# O l>
O N b
. m m Z n
- A b p&
o ~ k 3 o u m M a <
- u. g C m O w E O _
1 $. r E. L 8 g . - j-I aN E M \ in = O E a l I I l w i i
4000 g l l ' l l l O UPPER PLENUM E OUTLET PLENUM 3200 - O HOT LEG TO SURGE LINE - e SURGE LINE Q PRESSURIZER - r C 2 E 2400 - 2 - E 2
$ 1600 -
t sn 0 800 - i CORE SLUMP VESSEL FAILURE I I I I I I 24 28 32 36 40 44 48 52
~
TIME SINCE START OF ACCIDENT (s x 102) Figure 21. Primary System Gas Temperatures TMLB' IIigh Pressure Ejection (Case B)
2000 g , g g ; , O UPPER PLENUM N OUTLET PLENUM O HOT LEG TO SURGE LINE g 600 - e SURGE LINE ~ L Q PRESSURIZER w K 3 F
<t E
w 1200 - - lE su _ i = H -
- w e '
D l-o 800 - 3 E
> v_ P' n a sn k
W CORE SLUMP 7 VESSEL FAILURE 400 - - 1 I I I I l 0 24 28 32 36 40 44 48 52 TIME SINCE START OF ACCIDENT (s x 102) Figure 22. Primary System Structure Temperatures TMLB' High Pressure Ejection (Case B)
8 I I l l t O UPPER PLENUM E OUTLET PLENUM E ~ O HOT LEG TO SURGE LINE G SURGE LINE Q PRESSURIZER O M k 16 - _s - 8 4 su E
- p. 12 -
o - 3 O - O cc Z 8 - VESSEL FAILURE O - 3 CORE SLUMP us - E 1' i 4 - 0 * ' ' = 24 28 " 32 as 40 44 48 s2 TIME SINCE START OF ACCIDENT (s x 102) Figure 23. Fission Product Heating TMLB' High Pressure Ejection (Case B)
/
0.8 g i l l t VESSEL FAILURE
@ OVERALL O UPPER PLENUM E OUTLET PLENUM CORE SLUMP O HOT LEG TO SURGE LINE 9 SURGE LINE 0.6 - Q PRESSURIZER _
m O O ta. 2 - 9 0.4 - w ..- m m
'E 4
0 1 0.2 - _
r F 'N l ' "- '
0'0 T-28 32 " 36 40 44 48 52 TIME SINCE START OF ACCIDENT (s x 102) Figure 24. CsI Retention Factor TMLB' High Pressure Ejection (Case B)
t I I I I I V CONDENSED ON AEROSOLS Y CONDENSED ON STRUCTURE WALLS 16 - cn 5 O g 12 e
, w VESSEL FAILURE 8 -
l E CORE SLUMP - \ o e n 4 - l 4 O v v i i i I 28 32 36 40 44 48 52 1 TIME SINCE START OF ACCIDENT (s x 102) Figure 25. CsI Mass Retained TMLB' High Pressure Ejection (Case B) i
, - - _ . . _ _ . + .. ._m_. ..4 . . _ _ _ . _ _ - *___._m_ ._ m 06
?.
o -
@ Cal MASS RELEASED TO CONTAINMENT (kg) o g o a N W A m m i i i i
=g i O
g a
<m 3 >>
> u 22 ua H N O -4 -
m - 2 0 i E C N
- E l ~ M -<
h k l> (n O w to mm _ M
>3 H 2
b
- D N H O 6 a: mo
$:r >
O l > N O 6 m o
- D
- a z2 H
m p #- 2 7 g n = i P W $ u s y, 2 -
^
g g-im N
$ m i I I I l e N b
) 0.8 I I I I I VESSEL FAILURE y 0 UPPER PLENUM E OUTLET PLENUM CORE SLUMP O HOT LEG TO SURGE LINE 9 SURGE LINE O.6 - Q PRESSURIZER _ m , O , F-O 4 u. 2 e O 0,4 _ _ b-Z W 8 > w l E 1 e i F-0.2 - - I -_S__ I ^
"X N I
' 28 32
- 36 40 44
- 48 52 TIME SINCE START OF ACCIDENT (s x 102)
Figure 27. Te Retention Factor TMLB' High Pressure Ejection (Case B)
"e F -
i g u r
$ 2 <ee zw W I wO b O- O O Z>< E 2 mZ> 5 a^
n e 2 0 3 2 3 4 5 6 8 _ 0 - - - - T AE _ e VP _ M AA _ _ a 3 ' PR _ s T 2 OT I i - s I RC M U R E L e S A l T _ e I N = E . a _ s C 3 , e E 6 I d S - T T M A L R B T O 4 H i F 0 8 g A - 4 h C - P C I r D C e E O R s s N 44 g E ' u T S V r ( L E e s U S E x M SE j 1 o P L - 0 _ F e 2 A c ) 4 8 t I t L i U o R n l E ( / C a - - s 5 - _ _ e 2 B
)
l lll
)
8 C 2
- - - - - e s
p- y : p C ( a n o
, ~- 4 0 j '
4 2 i t c m~ ;m 0 ,' j E e
~ A- ,- ,f )
3 e r 0 0 u s
'f ' 1 2 s
" :- , " x s e
( r P
~ n- ,~
'r T N h g
E i
~ . 1, D H 6 I 1 C '
C B
~ : 1, A L F M O T m~ y T e R r u , ~ ' 2 1
A t T a S r E e E p R C m U N e E N " LI I S T I L A E s F I 8 a E L M G G E I MR S T m MU U S e UN S R E t V NEOE E s y EL LP TNIZ i ) S I R P T GL E E U a , y
, RE L S I 4
E L GS r P TTRE a PUOUR m UOHSP i r OEO9C P
- - - - - O 9 0 0 0 0 0 0 0 2 0 0 0 0 0 0 0 5 0 5 0 5 e 2 2 1 1 r u
g g L u E 3 F- < $ n. E u & as < 0 s ls i F
,w
a 8
- 2
^
p_4 .-
'v e
r u s s
" 4 e 1
Qm - 2, ' 2 r P
-r" Ev 1, h g
i H
" Ev : >v )
3 0 l 0 1 'B 2
" - ,- ,v x L M
( s T
" - _- T s N e E r i
H ' 6 1 D I C C u t a r e yf A p m F e O T
'v -
T e R r I 2, ' 2 A u 1 T t S c u E r 4 C t) E E N SC R I N I U S me es L L I E ta i E A I 8 M sC G F I y( MR L E T S MUU S n UN S R S yo NEOE E E ri EL LP TNIZ V at mc I R P T GLE U ie l RE EL S I 4 rj E L GS . PE P TTRE PUOUR . UOHSP . . 0 ONO9O 3 e r O u 0 0 0 0 0 O g 0 0 0 0 0 i 5 0 5 0 5 F 2 2 1 1 u., WE3HgEwQlwH
< .E mE3>03EFn ygI e ea
1.0 I i l l g , t c, r @ OVERALL O UPPER PLENUM N OUTLET PLENUM 0.8 - O HOT LEG TO SURGE LINE
- c. 6 9 SURGE LINE -
Q PRESSURIZER 9 E O i H I
$ 0.6 u.
e Z 9 b-2 uJ
>=
g 0.4 -
~ _
O 0.2 - g, FAILURE g A A A l W y" i M X *
- 0.0 l l
O # _ "4 8
~
~ I E 12 16 20 24
~
28 TIME SINCE START OF ACCIDENT (s x 103) Figure 31. CsI Retention Factor TMLB' High Pressure Ejection (Case C)
i 35 , , , , , , V CONDENSED ON AEROSOLS Y CONDENSED ON STRUCTURE WALLS 30 - 25 - h - E O w Z 20 - w m e us
$ 15 -
x 3 0 10 - - VESSEL FAILURE ' k 5 - - o t 7_
'rk i i i 1 0 ' "i 8 12 16 20 24 28 TIME SINCE START OF ACCIDENT (s x 103)
Figure 32. CsI Mass Retained TMLB' High Pressure Ejection (Case C)
4 _ _ _ _ _ - 2
)
B l i
' 0 l 2 O e
s a C ( n o i t 6 c I 1 e j E e r u s s
)
r e r h ( P I
' 2 E h 1
M g i
- I T H B
L M T I s I I 8 C e n r o b
- r i
A 1 4 3 3 e r u g i F O 8 5 2 9 6 3 0 1 1 1 2-}O WZEOmE< e" _
-_ ,,m_u ,__..g g _sm sa =m-.- .=-e..- ea-- h-- - -
+ * - " " " -
i i I i i I 12 - i r i i 10 -
~
i i t n ! .m j = 8 - O ' tu ' ,( E f - t#J t ca 4 m j w . >.,. _a 6 - m ' I E a , 4 -
'c
/
s l l
' .9 - .,
- mumm ..
f )
# g
, -- 0 I I I ' ' '~
~
0 8 12 16 20 24
,- - '4
/ ,
TIME (hr) - l Y e l'. Figure 34. Leaked CSI TMLB' High Pressure Ejection (Case-OllB) oe ?
< /
(. (' +
* ~'
. _ _/- - _ _ _ . _ _ _ _ _ - _ - _
sl " 4 _ _ _ - _ - - 2
)
0 A i I 2 2 1 0 e s a
._ (
C n o i i
' 6 t 1 c j
e _ E e r _ u
)
s r s h e ( r i I 2 E P
._ 1 M
_ h _ I g T i _. H _ B L M T i I
' 8 s
_ C d
~.
e k m.. a e L i 5
' 4 3 e
._ r u g _. i F 8 O 7 6 5 4 3 2 1 0
- O 0 0 0 0 0 0 0 0
$ OW)$V AWE 3O 8
i : ! 1 1f . : !l! 4l!i; l,; iii!4 - j 'l) 1! ,4 -
l s 1- i i i i i ii i i i i l ii; . . , ,
..g
- O CASE-011B ~
O CASE-012A ~ 0.1 -
.Z- -
m - - z z - O i
>=
0 0.01 -- E < : : o m - - ,
- u. - -
m - m - 2 - 0.001 --
' I '
0.0001 ' ' ' -
' ' ' ' ' ' 'I ' ' ' ' ' ' ' '
O.1 1 10 100 t PARTICLE SIZE (microns)
.s .
Figure 36. Leaked Aerosol Size Distribution TMLB' High Pressure Ejection t
1.0 g , , , , , , , , , 0.9 - O AIRBORNE - g O LEAKED 0.8 - A CEILING O WALLS - Z V FLOOR 0 0.7 -
- 7 -
O m 0.6 - 0 V)
$ 0.5 -
2 s O 0.4 - us - O E w 0.3 - 4 O O 0.2 - 0.1 - i m g
'O 4 6 10 12 14 b 18 20 22 TIME (hr)
Figure 38. Aerosol Disposition TMLB' High Pressure Ejection (Case-012A)
r VOL = 72605.8 m3 5984.5 307.04 UPPER LEVEL RGOMS y , [
)
6 FLOW PATH AREA = 215 m2 7/ 88.6 4158.1 0.36 215 MID LEVEL 2.56 ROOMS 0.84 5 2761.1 50.03 2.18 2841.7 4.91 SG 3.83 W/ PRESS SG 2 1 3 l 3801.6 LOWER ' 45 LEVEL ROOMS ( 4 l 13.34 1387.6 l CAVITY 1 Figure 39. Diagram of Multi-Cell CONTAIN Code Model l 102
80 i ; i i g 70 - 3 60 - S - W E
] 50 -
=
- =
=
h 40 - Z W 2 Z 30 F-
/ -
Z O o 20 - 10 - 0 i I i 1 0 4 8 12 16 20 24 TIME (hr) Figure 40. Containment Pressure TMLB' Pump Seal LOCA (Case-030)
l I, I i l 320 - 280 - W E
"3 b-4 g 240 - -
W CL 5* 2 W p., 200 - - Z W 2 3 -
$ 160 Z f O
O 120 - - 0 0 4 8 12 16 20 24 TIME (hr) i Figure 41. Containment Temperature TMLB' Pump Seal LOCA (Case-030)
1.0 g g i , , , , , , , , 0.9 - 0 AIRBORNE - O LEAKED 0.8 - A CEILING 0 WALLS - Z V FLOOR t 0 0.7 - F- - U 4 C C C g 0.6 - O
- u. -
u)
- $ 0.5 -
8 2 - J O 0.4 - us - O - g m 0.3 -
-y ? V 0.2 -
0.1 - 0 0 0 - i l A I ' '
'O ~2 4 6 8 10 12 14 16 18 20 22 TIME (hr)
Figure 37 Aerosol Disposition TMLB' High Pressure Ejection (Case-OllB) i i
l o 5 VOL = 293.4 ft3 6 VOL = 646.1 ft3 MCP = 489.6 Blu/F MCP = 4409 Blu/F COLD LEGS AR = 17.56 ft2 PUMPS AR = 63.55 ft2 AH = 490 ft2 AH = 361.4 ft DELX = 0.1771 ft DELX = 0.2 ft a ir 4 lVOL = 202.3 ft3 7 MCP = 26676.1 Btu /F COLD LEGS AR = 29.6 ft2 CONTAINMENT INLET PLENUM AH = 413.5 ft2 DELX = 1.17 ft o 3 VOL = 179.8 ft3 MCP = 4693.3 Blu/F OUTLET PLENUM AR = 23.44 ft2 AH = 313.25 f t2 DELX = 0.25 ft u l 2 VOL = 969.9 ft3 STRUCTURE 1 - MCP = 3525 Blu/F UPPER PLENUM MCP = 3525 Blu/F 2 - MCP = 5164 Btu /F AR = 86.9 ft2 4 3 - MCP = 952.6 Blu/F AH = 181.7 f t2 4 - MCP = 3639.7 Blu/F HEAT STRUCTURES DELX = 0.33 ft Ib i VOL = VOLUME MCP = MASS x Cp OF VOLUME CORE AR 2 FLOW AREA i AH = HEAT TRANSFER AREA ! ! DELX = STRUCTURE THICKNESS l l Figure 42. Primary System Flow Path TMLB' Pump Seal LOCA 106 w-~ =
& w -e-- ,-s, , - - - - , , - -, __ _ __
l
- i, ! ;
S~ O4s ui dO Oab OL &I I OEu eE n 1 4 6 8 0 0 0 0 0 0 2 _ 0 _
~ - - - _
F i g u 2 _ r 4 l _ e _ 4 . 3 T I M O E - u S t I 2 N I l 8 I e C - t E G S a T s A F R l T o w O 32 C l F T O M A R L C E B C I S L D U P E M u N P m T 3 l p 6 g ( S s e x a 1 l 0 2 -
)
L O C V A 4 ' E 0 S ; S E L F A I L U R E 4 - - _ 4
2000 l l l l 1 O UPPER PLENUM N OUTLET PLENUM ACOLD LEGS INLET PLENUM ACOLD LEGS - h1600 O MAIN COOLANT PUMPS W E D F 4 m g 1200 - g - 2 ' N 1 - h W K 3 0 800 - 3 C C = G h g
- n n CORE SLUMP k
uJ VESSEL FAILURE 400 -
' I ' ' '
0 40 44 20 24 28 32 36 TIME SINCE START OF ACCIDENT (s x 102) Figure 44. Primary System Structure Temperatures TMLB' Pump Seal LOCA (Case A)
30 i ; i , ,
@ UPPER PLENUM E OUTLET PLENUM ACOLD LEGS INLET PLENUM i ACOLD LEGS 25 - O MAIN CCOLANT PUMPS
[,f - r./.
^
? "E i-M
$ 20 -
2 E F nu g g 15 -
= o a .)
O O cc VESSEL FAILURE ft Z 10 - O CORE SLUMP II) us o o
- u. <.,
5 - T r T e. - ^ 0 ^ A' I 20 24 28 32 36 40 44 TIME SINCE START OF ACCIDENT (s x 102) Figure 45. Fission Product IIeating TMLB' Pump Seal LOCA (Case A)
1.0 g l l l 1 I i
@ OVERALL -
0 UPPER PLENUM 0.8 - E OUTLET PLENUM ACOLD LEGS INLET PLENUM [ ACOLD LEGS E O MAIN COOLANT PUMPS o I-O u. 0.6 -
- 2 9
Z W .:
$ 0.4 -
m 3 U CORE SLUMP VESSEL FAILURE 0.2 - - o n
-- L 0.0 -6 '5 24 26 28 30 32 34 36 38 40 42 TIME SINCE START OF ACCIDENT (s x 102)
Figure 46. CsI Retention Factor TMLB' Pump Seal LOCA (Case A)
2 _ _ _ _ 4 E R U I L I l 0 A 4 F L E ) - S A S E e V s i P l 8 3 a M ( C U L A S E R g ) 2 0 O C
- O -
1 L C 6 x l 1 o ' 3 s ( a e T S N p E m D u I P 4 C
; 3 C '
A B L F M O T T r . R o _ l 2 3 A t c T a S F E C n N o i
- S I
t l 0 n 3 E e M t M I e U T R N S E P e L M T P i T U 8 ME P 2 . MUL T N 7 4 UN IN A NE EL SSL e L LPGGO r _ L P TLLEEO u ARE C g RELDD EPTLLI N 1 6 2 i F VPUOOA OUOCCM -
@0NAAO 6- 4 0 8 6 4 2. , _
1 0 0 0
- 0. 2 -
0 0 goFO$ z9rzEyEU eb
! jl i
3.5 i , , , , , , , E PARTICULATE 3.0 - A VAPOR -
^
m / 5 1 H Z o g 2.5 - - Z q VESSEL FAILURE Z O 2.0 - -
- U g O H
Q
$ 1.5 - -
m a w II:
$ 1.0 - -
E e 0.5 - - CORE SLUMP o l ' ' '^ ' ' ' ' 0.0 4 42 24 26 28 30 32 34 36 38 40 TIME SINCE START OF ACCIDENT (s x 102) Figure 48. Te Mass Released TMLB' Pump Seal LOCA (Case A)
2 _ _ _ _ _ 4 a
- E R
U I L I I 0 A 4 F L E S ) S E A V I 8 e I 3 s a C (
)
2 A 0 C 1 O 1 n 6 x L P 3 s l M ( a U T e L S S N E E p R D m l O 4 I u C I 3 CC P A B F L O M T T l 2 R d I 3 A e T s S a e E l C e R _ A N I l 0 S s I 3 s E a M M I T I s _ C l E l 8 . T 2 9 A 4 L U e CR I r TO u RP g g AA 6 i PV ' 2 F EA
- - 4 4 0 6 2 8 4 2 2 2 1 1 0 0 s
_a6 &ZW]abZO0 o> QWU43WG UV4y ;O s ' b) ; 0
)
B 4 e
- - - - _ 4 s a
C ( A o C O i dr E R U L I 0 4 l L I A
) a F
2 0 e 1 S L E x p S S s m E ( u i V I 6 T P n f P 3 N ' M E B U D L L I M S C T E C R A e O r (n 2 F u i C I 3 O t a T r R e A p T m M S e U T N S E E P 8 C s i L M I 2 N a P G U I T P S ME T E m e MULN N t UN I M s i E A I r SSL T y EL LPGGO 4 S i P T LL EEO C I 2 y RE r E LDD N P TLLI a PUOOA m i UOCCM r ONAAO P
- - - - - 0 .
0 0 0 0 0 0 0 0 2 5 0 0 0 0 0 0 6 0 4 8 2 6 3 3 2 1 1 e r E u mE3kEElw[<0 i g F 5
2400 , , , , ,
@ UPPER PLENUM N OUTLET PLENUM 6 COLD LEGS INLET PLENUM ACOLD LEGS 2000 - O MAIN COOLANT PUMPS -
4 g / L w p 1600 -
< t.,
m w . G. e3 lE w N g ~ k 1200 - 5 $ 3 H O D m y 800 - F .
$ CORE SLUMP VESSEL FAILURE 400 -
I I I 0 I I 20 24 28 32 36 40 44 TIME SINCE START OF ACCIDENT (s x 102) Figure 51. Primary System Structure Temperatures TMLB' Pump Seal LOCA (Case B)
0.40 , , -
@ OVERALL 0 UPPER PLENUM N OUTLET PLENUM 0.32 -
ACOLD LEGS INLET PLENUM ACOLD LEGS e O MAIN COOLANT PUMPS o F- CORE SLUMP O I 0.24 - z ..
- O "
5 P Z VESSEL FAILURE w y 0.16 -
- x O
o 0.08 - ' - 0.00
,c i X\% i -,
24 26 28 30 32 34 36 38 40 42 TIME SINCE START OF ACCIDENT (s x 102} Figure 52. CsI Retention Factor TMLB' Pump Seal LOCA (Case B)
2 _ _ - - - 4 _ E R U L
, I 0
A ' 4 F L E ) S S B E V e s
' 8 ti 3 C
(
) A 2 C 0 O 4"6 L
. ^ 1 U '
x l P 3 s ( a M e U T S L S N p E E m R D u O I P C ' 4 C 3 C ' B A L F M O T T r R o
' 2 A t
3 T c S a F E C n N o
. I i
S t
' 0 n 3 E e M t
_ M I e U T R N S E P e L P M T T U 8 P ' ME 2 . MUL T N 3 UN N I A 5 NE EL SSL e LPGGO r P T EEO u RE LLC g PTLLI N ELDD I 6 i 2 F PUOOA OCCM OEAAO 64 0 8 6 4 2 0 0 0 0
- 0. 2 0
OFOt.Z9FZ$Lm$
<u - u
,=
' , 4 4
20 i i i i i i i V CONDENSED ON PARTICLES VESSEL FAILURE Y CONDENSED ON WALLS 16 - O CHEMISORBED CORE SLUMP g 3 I 5 O y 12 -
- a
~ E m
m 8 - 2
?
4 - I ' ' ' 0 -C 'T 'h ' 24 26 28 30 32 34 36 38 40 42 TIME SINCE START OF ACCIDENT (s x 102) Figure 54. Te Mass Retained TMLB' Pump Seal LOCA (Case B)
5.4 i i i i i 4.5 - i g 3.6 -
= -
~
en O C e uJ z m 2.7 - O m E 1.8 - I i 0.9 - i 0.0 I i I 0 4 8 12 16 20 24 TIME (hr) Figure 55. Airborne CsI TMLB' Pump Seal LOCA (Case-030)
l l i l l l i l l 48 - _ I l l 4 40 - - i
*?
4 o
- v-x 32 - -
B O
- uJ u) 4 24 - -
us J uJ E 0 16 - - 8 - - O I ' I ' I O 4 8 12 16 20 24 TIME (hr) Figure 56. Leaked CsI TMLB' Pump Seal LOCA (Case-030)
4 _ _ 2 I 0 2 ) 1 3 0 e s a ( C A i I 6 C 1 O L l a e S p
)
r m h u ( P I I 2 1 E ' M I B T L M T I s C e n r I
- 8 o b
r i A 7 5
~
e _ r I
' 4 u g
i F _ O 0 0 0 0 0 0 0 6 5 4 3 2 1 2 3OW[OmE< E~ 1 11 i i
4 I I I I l 4 1 250 - i g 200 -
=
1 I
- O e
. - O
- E w 150 -
Z m
- O m
m 4 100 - - i
- 50 - -
4 i \ 0
' ' ' ' I
, 0 4 8 12 16 20 24 l TIME (hr) i i Figure 58. Airborne CsOH TMLB' Pump Seal LOCA (Case-031) l
% a. . .- & ... _ _ _.d._m, M I I I i l 50 -
- 40 -
.a .. - >
~ O N =
W Wb i N O i N W
- a z 30 -
oc -
- O s
E 20 - t i l 10 - O ' ! I I I i 0 4 8 12 16 20 24 TIME (hr) ! Figure 59. Airborne CsI TMLB' Pump Seal LOCA (Case-131) 1 s t
DISTRIBUTION: U.S. Government Printing Office Receiving Branch (Attn: NRC Stock) 8610 Cherry Lane Laurel, Maryland 20707 200 copies for RG Mr. Thomas Walker (8) Office of Nuclear Regulatory Research U.S. Nuclear Regulatory Commission Washington, D.C. 20555 Mr. Rich Denning Battelle Columbus Laboratory 505 King Avenue Columbus, Ohio 43201 Dr. W. ?. Pratt Brookhaven National Laboratory Building 130, Brookhaven Lane Upton, New York 11973 Mr. R. J. Dallman EG&G Idaho Willow Creek Building, W-3 Post Office Box 1625 Idaho Falls, Idaho 83415 Oak Ridge National Laboratory (2) Bpilding 9201-3, Station 59 Post Office Box Y Oak Ridge, Tennessee 37830 Attn: Mr. S. A. Hodge Mr. T. Kress Mr. J. Travis Los Alamos National Laboratory Post Office Box 1663 Los Alamos, New Mexico 87545 Mr. Kenneth Keith Tenncssee Valley Authority 400 West Summit Hill Drive . W10dl82 C-k . Knoxville, Tennessee 37902 Mr. Wang Lau Tennessee Valley Authority 400 Commerce W9C157-CD Knoxville, Tennessee 37902 125
Mr. P. P. Bieniarz Risk Management Associates 2309 Dietz Farm Road N.W. Albuquerque, NM 87107 Mr. J. L. Tills Jack Tills and Associates 209 Eubank, N.E. Albuquerque, NM 87123 Sandia Distribution: 3141 S. A. Landenberger (5) 3151 W. L. Garner 6400 A. W. Snyder 6410 J. W. Hickman 6411 A. S. Benjamin (10) 6412 A. L. Camp 6415 F. E. Haskin 6420 J. V. Walker 6422 D. A. Powers 6425 W. J. Camp 6425 R. D. Gasser (15) l 126 L
NEC POAM 335 U $ NUCLEAR AEGULaTORY COMME $SiON ' at'C a Y NUM8Ea 'da ,a*e e, F#OL eds von Jbe , ,# #nyt I2 See E*r"3$2- BIBLIOGRAPHIC DATA SHEET NURF /CR-4563
$tt 1%$7 AUCTIONS ON Twt stytmsg SA' 86-0576 2 T4TLE AND $VST Lt 3 LE Ab S L t..
Analysis hf Station Blackout Accidents for the Bellefonte ressurized Water Reactor j ,,,,,,,,,,,,c,,,,,,,
.ONr.
j ve. i avt-Oaisi ,/ April 1986 R. D. Gasser, P. Bieniarz, J. L. Tills
- o a t'oa' '55 8 o uo~T. via j
Sept. 1986 i nia.oa ,%G oa:..vz.ucs ~ us .~o . NG .ooaiss ,,,. e. C , g s ekOsec t rasa acao.~.1 ~vveia Sandia National Labo tories 1 Post Office Box 5800 Albuquerque, New Mexico 7185 F L't, No . A-1258 10 5pON50aiNG oaGA%ilaf SON %awt A%Q UAILING ADoaE5$ u le Carea tia TYPE 05 aEPOaY Division of Accident Evaluatl 'Pechnical Office of Nuclear Regulatory R. earch US Nuclear Regulatory Commission Washington, D.C. 20555
.J SvPett ut%Yaa y NOTE 5 i) AsstaaCY #J00 weres er essJ An analysis has been performed for e llefonte PWR Unit 1 to determine the containment loading and the ra. ologi ~l releases into the environment from a station blackout accident. -
numbe of issues have been addressed in this analysis which include th effects direct heating on containment loading, and the effects of fissi n product . ating and natural convection on releases from the primary sys 'm. The res ts indicate that direct heating which involves more than about 50% of e core can fail the Belle-fonte containment, but natural nvection in th RCS may lead to overheat-ing and failure of the primary stem piping bef e core slump, thus, eliuinating or mitigating dire .. heating. Releas from the primary systen' are sianificantly increased be re vessel breach d , to natural circulation and after vessel breach due tc reevolution of retair d fission products by fission oroduct heating of RCS structures.
'I w
v 1 I 14 oCCuwt NT AN AL v5'S a E E vnO*DS DESca* TORS it avanagig t v STATEvt%T Unlimited
'6 SECumity CL A5S p acateO%
t r.4 p.,,,
. ios%7,,,eas One% e%:so fiaws UNCLASSIFIFI ITne reporrt oUNCLA
% . a c,S.S.cis I F I E T-t$ ParCE
- U S GOvtRNMENT PRINTING ot SICE 1986- 676 444 41012
~~ ~
1 1AN1R3 120555078877 US NRC ADM-DIV POLICY & 0F PUB SVCSPUS 20555 MGT BR-POR NUREG W-501 DC WASHINGTON I e
~}}