ML20081J729
| ML20081J729 | |
| Person / Time | |
|---|---|
| Site: | Browns Ferry  |
| Issue date: | 10/31/1983 |
| From: | Harrington R, Ott L OAK RIDGE NATIONAL LABORATORY |
| To: | NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| References | |
| CON-FIN-B-0452, CON-FIN-B-452 NUREG-CR-3179, ORNL-TM-8635, NUDOCS 8311090094 | |
| Download: ML20081J729 (116) | |
Text
{{#Wiki_filter:NUREG/CR-3179 ORNL/TM-8635 e e i e
- e- e-The Effect of Small-Capacity, UNION High-Pressure injection Systems CARBIDE on TOUV Sequences at Browns Ferry Unit One R. M. Harrington L. J. Ott 1
Prepared for the U.S. Nuclear Regulatory Commission Office of Nuclear Regulatory Research Under interagency Agreements DOE 40-551-75 and 40-552-75 W. 8311090094 831031 DRADDCK05000g
?
I I l 4
~>
Printed in the United States of Amenca Available from National Technical Information Service U.S Department of Commerce 5285 Port Royal Road. Springfield. Virginia 22161 Available f rom GPO Sales Program Division of Technical information and Document Contro! U.S. Nuclear Regulatory Commission Washington, D.C. 20555 O This report *as prepared as an account of work sponsored by an agancy of the United States Go.ernment Neither the u nited States'3cvernment nor any agency I therect. nor any of their employees. makes any warranty, express or +mpled, or assumes noy legal liabnity or respon:ibility for the accuracy, completeness, or usefulness of any information, apparatus. product. or process disclosed. or represents that its use would not anfringe privately owned rights Reference herein to any specific commerc:al product, process, or service by trade name. trademark, rnanuf acturer, or otherwise. does not necessanly constitute or imply its endorsesaent. recommendation or favonna by the United StatesGovernment or any agency kiiereof The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof a d'
NURE/CR-3179 ORNL/TM-8635 Dist. Ca te gory RX, 3 S e e, - Contact No. W-7405-eng-26 Engineering Technology Division Instrumentation and Controls Divisior
' DIE EFFECT OF SMALL-CAPACITY, HIGil-PRESSURE INJECTION SYSIENS ON 'MUV SBQUENCES AT BROWNS FERRY UNIT ONE R. M. Harrington L. J. Ott Manuscript Complete - August 15, 1983 T' Date Published - September 1983 e
Notice: This document contains information of a preliminary nature. It is subj ect to revision or correction and therefore does not represent a final report. Prepared for the U.S. Nuclear Regulatory Commission Of fice of Nuclear Regulatory Research Under Interagency Agreements DOE 40-551-75 and 40-552-75 NRC FIN No. B0452 Prepared by the Qj Oak Ridge National Laboratory y Oak Ridge, Tenne ssee 37830
- operated by UNION CARBIDE (X)RPORATION for the DEPARIMENT OF ENERGY l
t 111 (X)NTENTS P_agg 4 l P0 REWORD .......................................................... y
SUMMARY
........................................................... vil ABSTRACT ..........,............................................... 1
- 1. INTRODUCTION .................................................. 1 Referem&es for Chapter 1 ...................................... 3
- 2. INITIATING EVENTS ............................................. 6 2.1 Sequence Definition ...................................... 6 2.2 Implications for Recovery By Using the SLC System and the CRD Hydraulic System ............................. 6
- 3. FLOW CAPABILITY OF SMALL CAPACITY HIGH PRESSURE INJECTION SYSTEMS ............................................. 8 3.1 Introduction ............................................. 8
, 3.2 CRD Hydraulic System ..................................... 8
' 'Ek* 3.3 SLC System ............................................... 9
, 3.4 Summary .................................................. 10
- 4. PARAMEIRIC SIUDY OF THE EFFECT OF 1EE SMALL INJECTED FLOWS IN MITIGATION OF TQUV SEQUENCES ............................... 16 4.1 Introduction .........,................................... 16 4.2 Reactor Vessel Remains at Pressure ....................... 16 4.2.1 Time to core uncovery ............................. 16 4.2.2 As se s sment of core damagt ......................... 17 4.2.3 Conclusions ....................................... 18 4.3 Sequence with Rapid Reactor Vessel Depressurization ......- 19 4.3.1 Time to core uncovery ............................. 20 4.3.2 Asse ssment of core damage ......................... 20-4.3.3 Conclusions ....................................... 21 I
- 5. NO-OPEEATOR-ACTION SEQUENCES .................................. 31 5.1 Introduction ............................................. '31
. 5.2 Reactor Vessel Remains at Pressure ....................... 32 i .
() .,x 3 5.3 Stuck Open Relief Valve .................................. 32
, 5.4 Conclusions .............................................. 33
- 6. POTENTIAL FOR RECOVERY WIIM OPERAIUR ACTION ................... 40 6.1 Introduction ............................................. 40 6.2 Reactor Vessel Remains at Pressure ....................... 41
iv
.P__agg i,
! 6.3 Stuck Open Relief Valve .................................. 42 . 6.4 Rapid Depressurization ................................... 44 6.5 Concl usions .............................................. 45 .
- 7. ACCIDENT SH)UINCE Willi OPTIMJM OPERATUR ACTION ................ 62 7.1 Discussion ............................................... 62 7.2 Procedural Considerations ................................ 63 7.3 Recommended Procedures ................................... 64 7 . 3 .1 IQUV sequence not initiated by LOSP ............... 64 7.3.2 TQUV sequence initiated by LOSP ................... 64 7.3.3 IQUV sequence with SORY (not initiated by LOSP) ............................................. 65 7.3.4 IQUV sequence with SORY (initiated by LOSP) ....... 65 7.3.5 Pl a n t specific considerations ..................... 65 7.4 Recommended Equipment Modifica tions ...................... 65 7.5 References ............................................... 66
- 8. DISCUSSION OF UNCERTAINTIES ................................... 68 ,
Reference s for Chapter 8 ...................................... 72 8P Appendit A. Modifica tions to the BWR-LACP Code for This , Study ................................................ 75 Appendix B. MARCH 1.1 Code Leprovements for BWR Degraded Core Studies ......................................... 79
! B.1 Introduction ............................................. 79 B.2 BWR Core Model ........................................... 80 B.2.1 BWR reactor vessel and core description ........... 80 B.2.2 Core heat transfer models ......................... 81 B.3 BWR Reactor Pressure Relief Model ........................ 84 B. 3 .1 Description of BWR pressure relief system ......... 84 )
B.3.2 Pressure relief model ............................. 85 B.4 MARCH Steam / Hydrogen Physical Properties and the Incore Gaseous Convective Heat Transf er Correlation ...... 86 B.5 Core Quench Model s ....................................... 88 B.6 Impact of MARCH 1.1 Code Leprovements on . Accident Events and Timing ............................... 89
-b B.6.1 Introduction ...................................... 89 B.6.2 BWR Core Model .................................... 90 -
B.6.3 SRV model, physical properties / correlation, quench model ...................................... 90 Reference s for Appendix B ..................................... 91 Appendix C. Acronyms and Symbols ................................. 105
I v FOREWORD i e g_, The Severe Accident Sequence Analysis (SASA) Program was established in October 19g0 by the Division of Accident Evalua tion of the Of fice of Nuclear Regulatory Research, U.S. Nuclear Regulatory Canaission (NRC), to study the possible ef fects of potential nuclear power plant accidents. Under the auspices of the program, boiling water reactor (BWR) studies are being conducted at the Oak Ridge National Laboratory (ORNL) using Browns Ferry Unit 1 as the model plant. Assistance and complete cooperation is provided by the plant owners and operators, the Tennessee Valley Author-ity. The function of the SASA program at ORNL is to conduct detailed an-alyses of the dominant (most probable) BWR accident sequences, which have , been identified by probabilistic risk assessments (PRAs). A rece ntly com-pleted study
- concerns the Loss of Decay Heat Removal (DHR) accident se-quence s a t Browns Ferry. Among other significant findings, the results of this study demonstrate the importance of the control rod drive (CRD) hy-draulic system in accident situations.
It should be mentioned that the effectiveness of the CRD hydraulic system in accident mitigation at Browns Ferry had previously been demon-strated in a practical manner by its use during the fire at that f acility in March, 1975. Thus it seems strange that the PRAs conducted since the date have continued to neglect this important sy s t em. O- Although the reactor ves sel inj ection provided by the CRD hydraulic system is small (60 gpm) during normal reactor operation, it increases
- significantly when a scram signal is in effect. It is important to note that this occurs automatically,t without the need for any operator action.
With the reactor vessel at normal operating pressure, the inj ection pro-vided by the CRD hydraulic system is about 110 gym with a scram in ef fect; if the vessel is depressurized, the post-scram inj ection by this system increases to about 180 gpm. About 4 h af ter reactor scram, the power generation by decay heat has decreased to the point that no cooling water inj ection beyond that pro-vided by the CRD hydraulic system is required. Maj or findings of the SASA Program Browns Ferry loss of DHR accident sequences study were dependent on this fact.
; Since the ef fects of the CRD hydraulic system were shown to be sig-nificant to the progression of the Loss of DHR accident sequences, it was deemed desirable to investigate the ef fect of the operation of this system on other BWR accident sequences. The case of total loss of emergency core cooling syst ems (ECCS) combined with loss of the reactor core isolation i cooling (RCIC) sy st em immediately after scram is at the other extreme. In this "TQUV" accident sequence,t the power conversion system is unavailable 4
(1; *S. A. Hodge et al., " Loss of DHR Sequences a t Browns Ferry Unit One l - Accident Sequence Analysis," NURBG/CR-2973, ORNL/TM-8532, May 1983. tihe scram inlet valves open, bypassing the control valve that limits the inj ection flow during normal reactor operation, tFrom the Reactor Safety Study, WASH-1400, T = transient event, Q = failure of normal feedwater system to provide core make-up water, U = 4 f ailure of EFCI or RCIC to provide core make-up water, V = failure of low-pressure ECCS to provide core make-up water,
~ -- g gi- -r7e y- - =w g
- f
vi and neither the high pressure nor low pressure ECCS systems perform their intended f unction. Thus the low capacity non-ECCS inj ection systems are the only means to provide make-up water to the reactor vessel. Two of
- these sy s t em s , the CRD hydraulic system and the standby liquid control ,
system (SLCS) are of especial interest because they are capable of inj ec-tion with the reactor vessel at pressure. This report addresses the capability of the CRD hydraulic system and the SLCS to provide core cooling during the TQUV accident sequence. The matter is important because virtually all PRAs have identified the IQUV sequence to be among the dominant accident sequence s for BWRs. Th e r e-sults of this study show that core damage can be minimized or avoided al-together in this accident sequence by prudent use of the CRD hydraulic system and the SLCS in various possible combinations. Thus it is a pri-mary purpose of this report to indicate the actions that should be taken by the operators to mitigate the TQUV sequence and to recommend the.t pro-cedures be written and that operators be trained to respond correctly should this sequence be initiated. It also seems evident that the i existence of the CRD hydraulic system and the SLCS should be recognized in f uture BWR PRAs. s Stephen A. Hodge SASA Proj ect Manager . Oak Ridge National Laboratory
^
s
vii
SUMMARY
i e gag In a TQUV* sequence, there is a transient followed by reactor scram, followed by failure of all the systems that would normally be relied upon
, to inj ect cooling water into the reactor vessel. If the reactor scram occurs f rom 100% power and no vessel make-up water is provided, the un-covering of active f uel begins af ter about 0.5 h, and is followed by de-struction of the core if sufficient makeup water is not subsequently ini-tiated. The TQUV accident sequence has been shown to be among the
, dominant sequences in almost all BWR probabilistic risk assessments (PRAs). This report provides an analysis of the ability of the Control Rod Drive Hydraulic System (CRDHS) and the Standby Liquid Control ( SLC) System to adequately cool .the reactor core following initiation of a TQUV accident seque nce . These maall-Mapacity, high pressure inj ection systems have traditionally been ignored in BWR risk assessment. The small capacity systems have a significant effect on TQUV se-quences because only moderate reactor vessel inj ection flows are required for shutdown cooling. The CRDH and the SLC systems can provide enough flow to replace coolant lost due to decay heat steam production within 0.5 h af ter a reactor scram from long-term operation of 100% power. The SLC system can, upon operator initiation, inj ect 3.53 L/s (56 spm) of unbor-ated water through a single positive displacement pump.t Following a
- reactor scram, the CBDBS, without operator action, will inj ect a flow of 7.06 L/s (112 gPa) into the reactor vessel with the vessel at normal pres-a sure [7.00 MPa (1000 psig)] . Witn operator action, this inj ection can be increased to almost 18.9 L/s (300 gpm). The CRDES centrifugal pumps must pump against the pressure within the ramctor vessel; therefore, additional CRDHS injection is possible at lower vessel pressures. Both the CRDHS and the SLC System take suction on the Unit 1 condensate storage tank (CST).
There is normally about 1,371 ms (362,000 gal) of water in the CST - j enough to last for more than 35 h of steaming on decay. heat af ter scram i f rom full power. Two computer codes were used to obtain the results presented in this report: the ORNL-developed BWR-LACP code and a special ORNL-modified ver-sion of the MARCH code. It was nece ssary to use both of these code s be-l cause the MARCH code is not currently capable of predicting the progres-sion of an accident sequence before core uncovery in a BWR nor of calce- ! lating reactor vessel water level above the top of the active fuel.
. Therefore, BWR-LACP was used to calculate the time to core uncovery and MARCH was initialized at the point of initial uncovery of active fuel. It was necessary to extensively modify MARCH to permit accurate asse s maent of core ' damage for sequences involving partial or temporary uncovering of the fuel, during periods of transient steaming conditions caused by intermit- . tent safety relief valve (SRV) a ct ua tion.
CL Li
*From the Reactor Saf ety -Study, WASH-1400, Tktransient event, Q= failure of normal feedwater system to provide core make-up water, U-failure of HPCI or RCIC to provide core makeup water, V=f ailure of low pressure ECCS 'to provide core makeup water.
tThe SLC inj ection rate is independent of. reactor vessel pressure. 1 s .
-k
vili The po st- s cr am injected flow available from the CRDHS in the case of no operator action varies from 6.55 L/s (104 spm) at a reactor vessel pressure of 7.69 MPa (1100 psig) to 11.3 L/s (180 spm) a t 0.1 MPa (0 . psig). This flow is not enough to prevent severe fuel damage for a TQUV e accident sequence initiated f rom 100% pcwer; however, if not neglected, it has a significant effect on the calculated timing of critical events in , the resulting severc accident sequence. The mitiga tive ef fect of operator actions to initiate SLC system in-jection and to increase the CRDHS injection depends on the timeliness of the operator action, the total inj ected flow made available, and on the vessel pressure control. Normal vessel pressure control implies that the operators maintain the reactor vessel at, or near, full pressure by inter-mittent manual SRV actuations. Without operator control, the SRVs will maintain the vessel at full pressure by automatic actuations of the lowest-set SRV. A TQUV sequence with the reactor vessel at less than full pres-sure could be the result of either an SORV or of operator initiation of a de pr e s s ur iz a ti on.
- For a TQUV sequence at full reactor vessel pressure, core uncovery can be prevented if the operators act to increase the total inj ected flow to 14.1 L/ s (225 spm) within 10 min af ter the scram from 100% power. This flow can be obtained in either of two ways: by running both of the CRDHS pumps, or by running one CRDHS pump together with the SLC system. In either case, the CRDHS inj ection is increased by operator action to open the pump test bypass flow path as described in Browns Ferry Emergency '
Operating Instruction No. 41, " Water Makeup Methods to the Reactor Ves-sel . " For a TQUV sequence at full reactor vessel pressure, but with less than 14.1 L/s (225 gpm) of inj ected flow, there will be a period during which the water level in the reactor vessel is below the top of the active f uel . For example, if an inj ected flow of 12.6 L/s (200 spa) is initiated within 10 min af ter trip, part of the active f uel is uncovered for a per-iod of about 6 h, starting about 54 min af ter the scram. The hottest part of the uncovered fuel reaches a peak temperature in the neighborhood of 1089 K (1500*F), but the resulting fuel damage would probably be neg-lisible. For the nearly identical case in which the initiation of the 12.6 L/s (200 gpm) inj ected flow is delayed from 10 min to about 35 min, the period of core uncovery is more extensive, and the peak fuel tempera-ture is increased to about 1478 K (2200'F). This higher temperature would allow some metal-water reaction in the hotter parts of the uncovered fuel, but the damage would not be severe and no melting or loss of geanetry would be experienced. The most immediate consequence of an SORV or an emergency depressuri-zation in a TQUV sequence is the accelerated loss of water L;ventory due to the flashing of water to steam. This results in an earlaer core un - covery than would occur in a simple at pressure boil-of f. The quicker . core uncovery of the depressurized cases is partially off-set by the j increased inj ected flow that the CRDHS con pump at lower reactor
*It should be noted that there are various low pressure non-ECCS sys-tems such as the condensate pumps that, if operable, could be used if the reactor vessel is depressurized. Consideration of the effect of these systems is beyond the scope of this report.
l ; , 11 I i ve ssel pressures. However, the CRDHS flow can increase so much that the l pump suction pressure decreases to below the setpoint of the automatic ! , " low suction pressure" trip. The ability of the operators to restart the I CRDHS pumps af ter a " low suction pressure" trip is a major uncertainty of the depressurized reactor vessel IQUV sequences. The results obtained ' ( , under the assumption that the operators would be able to prevent or j recover from CRDHS pump trip show that the depressurized TQUV sequences ! with CRDHS and/or SLC system inj ection result in negligible fuel damage. Several IQUV sequences with depressurization of the reactor vessel and with delayed operator action have been investigated. For example, a MARCH run was performed for a sequence with SORY occurring immediately following the scram, but with effective operator action to establish enhanced CRDHS flow delayed until 30 min af ter the scram, and with operator actuation of injection via the SLC system delayed until 40 min I after the scram. The MARCH results predict an approximately 3 h period of extensive core uncovery and a resulting peak fuel temperature of 1144 K (1600*F). There would be no severe fuel damage for this case. The conditions for this case assume effective, although delayed, use of the ,
! CRDH S, inc1LJing the ability to prevent or recover from the low suction l pressure trip of the CRDHS pumps that can be expected as a result of ,
decreasing reactor vessel pressure. Five modifications to emergency procedures are recommended as a way of increasing the probability of effective operator mitigative action in the unlikely event that a TQUV accident should occur:
- 1. To ensure prompt operator response, the emergency procedures should specify a quantitative guideline for the initiation of the operator l actions nece ssaqr to increase the CRDHS flow and to start SLC inj ection.
1 The criterion for triggering operator action could be based on time af ter l scram and/or on detected vessel water level. l 2. If the CRDHS is operating, the emergency procedures should not require a rapid depressurization unless the low pressure injection systems are verified to be running.
- 3. The procedures should specify that a raw cooling water pump must be started in order to provide cooling water for the CRDHS pumps af ter a loss of of fsite power.
l 4. A precaution should be added that a " low suction pressure" trip l of the CEDHS pumps can be expected when inj ecting via the pump test bypass line into a depressurized reactor vessel. Instructions for preventing or coping with the resulting " low suction pressure" irt; 'hould be included.
- 5. A specific contradiction in the Browns Ferry F.ocedures regarding
- the opening of the CRDHS pump throttling valve (No, st:-527) should be resolved. This valve must be fully opened to obtain sufficient emergency inj ection flow. There are several equipment modifications that would increase the
. reliability of the emergency use of the CRDHS:
,, 1. The maximum range of the CRDHS main and pump test bypass flow indicators (which indicate in the main control room) should be increased.
, In all of the sequences discussed in this report the indicated flow would be pegged at the high end of the meters, and the operators would not know the actual flow being inj ected into the reactor vessel by the CRDHS.
X
- 2. The time delay recently installed in the " low suction pressure" trip circuit of the Susquehanna 1 CRDHS should be investigated for appli-cability to the Browns Ferry GDHS. The purpose of the time delay is to ,
prevent momenitary decreases in suction pressure from initiating a " low suction pressure" trip of the ptop. 4 G G
l THE EFFECT OF SMALL-CAPACITY, HIGH-PRESSURE INJECTION SYSTEMS ON IIlUV SHlUINCES AT BROWNS FERRY UNIT ONE R. M. Harrington L. J. Ott ABSTRACT The TQUV [ Transient-induced scram, Power Conversion System Unavailable, High Pressure (HPCI, RCIC) and Low Pressure (RHR, Core Spray) inj ection systems unavailable] accident sequence has been shown to be among the dominant accident sequences in almost all BWR Probabilistic Risk Assessments (PRAs). This report provides an analysis of the ability of the Control Rod Drive (CRD) hydraulic system and the Standby Liquid Control (SLC) sy s t em, utilized singly or in combination, to adequa tely cool the resctor core during this accident sequence. The class of shutdowns considered consists of those Initiated by a trans-lent or a loss of of f-site power in which there is a succe ssful reactor scram and there is no pipe break. Various combinations of flow delivery capability and operator action are considered
, and an optimum procedure is recommended. A special version of the MARCH computer code is used to calculate core temperatures and estimate fuel damage for cases involving partial or full
- core uncovery. The results show that f uel damage can be avoid-ed by appropriate operator action.
- 1. INTRODUCTION This is the fourth report in a series of accident studies concerning the BWR 4 - MK I containment plant design sponsored by the Containment Systems Research Branch of the Division of Accident Evaluatzon of the Nu-clear Regulatory Commission. These studies have been conducted by the Severe Accident Sequence Analysis (SASA) Program at the Oak Ridge National Laboratory (ORNL) with the full cooperation of the Tennessee Valley Au-thority (TVA), using Unit One of the Browns Ferry Nuclear Plant as the medel design. Each unit of this three-unit plant has a maximum authorized power of 3293 MW(t), or 1067 not MW(e). The primary containment s are of the Mark I pressure suppression pool type and the three unity share a sec-
- ondary containment of the controlled leakage, elevated release de sign.
d' Each unit occupies a separate reactor building located in one structure underneath the common refueling floor.
- The obj ective of this study is to analyze the effect of the Standby Liquid Control (SLC) System and the Control Rod Drive (CRD) Nrdraulic sys-tem on accident sequences in which all other sources of injection have f ailed, but in which there is no f ailure of the control rods to scram and there is no pipe break. These sm al l-ca pa ci ty, high pressure systems can
I l 2 have a significant effect because the reactor core is on decay heat after a succe ssf ul reactor scram and because there is no significant leakage of reactor coolant f rom a line break (the special case of stuck-open relief , valve is considered). Possible initiating events are discussed in more detail in Chapter 2. The motivations for initiating this study are derived from the re-
- suit s of a recently completed ORNL study, from a review of PRAs, and from actual plant experience. The ORNL study of Loss of Decay Heat Removal sequence s (Ref.1) concluded that, with the reactor vessel depressurized, the CRD hydraulic sy stem would, without any operator action, supply all needed reactor vessel injection as soon as 4 h af ter reactor scram.
i Several recent PEAS (e . g. Ref s. 2-4) have concluded that the TQUV se-quence and similar sequences in which vessel water injection capability is Even though the SLC and lost dominate the risk of core melt for BWRs. CRD hydraulic systems can meet or exceed the vessel water injection re-quired to remove decay heat after shutdown, the effect of these sy stem s has typically been neglected by PRAs. Therefore, the results of this study provide information, not previously available, which can be used to good advantage in future risk analysis. l Fi gur e 1.1 show s t he fl ow that must be injected into the reactor ves-sel to maintain constant vessel coolant inventory by replacing the coolant boiled away by decay heat. The range of flows obtainable via the CRD hy-draulic system and the rated flow of the SLC system are also indicated. As shown by this figure the required reactor vessel inj ected flow is with- . in the capability of these systems very soon af ter the reactor scram. An example of an actual plant accident in which the CRD hydraulic system played an important part is the cable fire that occurred at the . Browns Ferry Plant on March 22, 1975. Throughout the ~7 h period when multiple f ailures of plant systems were occurring, the CRD hydraulic sys-tem, without any operator action, continuously provided more than 6.3 L/s (100 gpm) of injected flow to the reactor vessel (Ref. 5). This flow was especially valuable during periods when no other sources of injection were in service. It is shown in this report that the small-capacity, high pressure injection systems can be equally valuable in other emergency situations involving loss of all normally utilized sources of vessel water inj e ct ion. Chapter 2 briefly outlines the initiating events and other circum-stances of the loss of vessel water inj ection sequences considered by this report. Chapter 3 estimates how much flow the CRD hydraulic system and the SLC system can pump into the reactor vessel under a variety of conditions. Operator actions necessary for initiation of SLC system flow and for en-hancement of the rate of CRD hydraulic system inj ected flow are also dis-cussed. The problem of critical importance for the operators to solve in a . TQUV accident sequence is to maintain sufficient vessel injection to cool . the reactor core. The thermal hydraulic response of the primary contain-ment is of secondary importance. In a TQUV sequence, all systems which . would, or could, take suction on the pressure suppression pool are inoper-ative. Thus, the pool temperature and pressure cannot affect the supply of water for injection to the reactor vessel. The results presented in Ch aps. 4-6 do not include primary containment response.
3 Chapter 4 is a parametric study of the effect of various inj ected flow rates following shutdown for two basically dif ferent categories of reactor vessel pressure control. For the first category, it is assumed l that the reactor vessel is maintained at pressure, in the neighborhood of l 6.90 MPa (1000 psig). For the second category, a rapid depressurization l is assumed to begin shortly af ter the initiating reactor trip. l Chspter 5 presents the calculated ef fect of the CRD hydraulic system for two no-operator-action ca se s: loss of vessel water injection with no stuck-open relief valve (SORV) and loss of vessel water inj ection with a , SOHV.
- Chapter 6 presents the calculational results for at pressure, SORV,
- and f ast-depressurization sequences in which there are significant delays j before operator action to enhance the inj ected flow from the CRD hydraulic j
system and to initiate the SLC system. These results lead to conclusions regarding the likelihood that these operator actions could be performed in
- time to avert serious core damage.
I Chapter 7 discusses some problems in the present operating procedures i for use of the SLC system and CRD hydraulic system for emergency vessel water inj ection. Recommendations are made for specific improvements to the operating procedures and training to improve the chance s for succe ss-fut core cooling in the event of a plant accident in which the CRD system and SLC system are the sole sources of vessel water injection. Chapter 8 specifies known uncertainties in this analysis and consid-
. ers how these might affect the results.
TWo computer code s were used to calculate reactor vessel thermal-hy-i draulic conditions for this study: the ORNL-developed BWR-LACP code (Ref. j . 6, 7) and a special DRNL modified version of the MARCH code, The BWR-LACP i code was used to calculate the time to core uncovery; the MARCH code was initialized at the point of initial uncovery of active fuel. It was nec-
- essary to use both of these codes because MARCH is not currently capable of predicting the progressien of an accident sequence before core uncovery in a BWR nor of accurately calculating reactor vessel water level above the top of active fuel, and because the BWR-LACP code is not programmed to calculate fuel damage. Appendix A describes modifications to the BWR-LACP code required for this study. Appendix B de scribe s the modifications per-formed on the MARCH code to permit accurate assessment of core damage for sequences involving partial or temporary uncovering of the fuel. This includes periods of transient steaming conditions caused by intermittent safety relief valve actuation.
References for Chanter 1
- 1. D. H. Cook e t al. , "Los s of DHR Sequences at Browns Ferry Unit One -
, Accident Sequence Analysis," NURBG/CR-2973, Vol.1, ORNL/TM-8532/V1, j , May 1983.
- 2. Philadelphia Electric Company, Probabilistic Risk Assessment -
Limerick Generating Station, Docket Numbers 50-352 and 50-353. i t
,, ._. ,, , ._ , ,. ,- _ ,m ._-,m, _ _ , _ . . . . . , y _ , , . m -. . _ , , . _ _ ,
4
- 3. S. W. Hatch et al., " Reactor Safety Study Methodology Applica tions Program : Grand Gulf No.1 BWR Power Plant," N'JRG/ Cit-1659 (Oct ober 1981). .
- 4. General Electric Company, "GEESAR II 238 Nuclear Island FSAR," GE 2
Report 22A7007, Proprietary Chapter 15 D (March 1982). .
- 5. R. L. Scott, " Browns Ferry Nuclear Power-Plant Fire on March 22, 1975 " Nuclear Safety, Vol.17, No. 5, Sept enber-October 1976.
- 6. W. A. Condon e t al., "SBLOCA Out side Containment at Browns Ferry Unit One - Accident Sequence Analysis," NURIE/CR-2672, Vol. 1, ORNL/
TM-8119/V1 (November 1982), Appendix A.
- 7. D. H. Cook et al., " Station Blackout at Browns Ferry Unit one, Acci-dent Sequence Analysis," NURFE/CR-2182, ORNL/NURIE/Tf4-455/V1 (Novem-ber 1981), Appendix A.
4 S e
1 l e ORM.-DWG 83-10485R I 1000 -
^ 900 -
-60 m
E o u1
- c. 800 - - 50 \
cn - v v
- 700 -
- 3 40 S '
O 600 - O
-.J .J b- 500- k
.......................................,y....... -30 0
W 400-O W W W o 300 y
- CRDHS
- 20 o g
n 200 - I , Z - 10 z
' 10 0 - ------------ b-------- -
'.....y-.........,~~y-..,,--...g--n-...3,,.,....,......a
,4 . y . ... .. ..
- - - - - 0 0 1 2 3 4 5 6 TIME (h af ter REACTOR SCRAM)
' CALCULATED ASSUAWNO (OVf7 ANT MACTor FEfter AFF35MRE 1005 insteel power (decey heet per 1979 ANS sed, with actinide decey), eeelmee injected et 305 K (90*F) and reemoved as dry soeureted eseo.
Fig. 1.1. Inj ected reactor vessel flow required to replace decay-heat produced steam.
4 6
- 2. INITIATING EVENTS i
2.1 Seauence Definition The accident sequences considered in this study are commonly ref erred . to in PRAs a s MUV, MUX, TeUV, and TeUX, where: Te = loss of of f-site power (LOSP) initiator, T = transient initiator not involving LOSP (e.g. loss of feedwater,
- main steam isolation valve closure, etc.),
Q = condensa te/feedwater system f ails to provide reactor vessel make-I up, U = failure of high pressure inj ection systems (HPCI and RCIC), V = failure of low pressure emergency core cooling (ECC) systems, or f ailure of automatic depressurization system ( ADS), I = failure of operators to intiate depressurization in a timely manner (i.e. Iow pressure ECC systems are available but not used soon enough). For all of these seque nce s, there is a successful reactor scram, and there is no pipe break. If the main steam isolation valves do not close t as a result of the initiating event, they are closed shortly thereaf ter by , a Group I isolation on sensed low vessel water level (i.e., level 1 11.9 m (470 in.) above vessel zerol . The decay heat of the reactor core pro-duce s steam which is relieved from the reactor vessel, either automati-cally or under operator control, by intermittent actuation of a safety relief valve (SRV) . Although there is initially more than 3 m (10 f t) of water above the top of the fuel in the reactor vessel, the loss of coolant inventory through the SRVs lowers the water level. If the reactor scram occurs f rom 100% power and no vessel make-up is provided, the active f uel begins to be uncovered af ter about 0.5 h; this is followed by destruction of the core if sufficient make-up is not subsequently provided. 2 .2 Imolications for Recovery by Usina the SLC System and the CRD Hydraulic System The most important sequence event affecting the ability of the plant operators to use the SLC system and CRD hydraulic system to prevent or to mitigate core damage is the availability of of f-site power. For those se-quences not involving LOSP, the operation of these systems is not limited and the level of stress on the operators is not increased by the addition-al complexity of -the events related to the LOSP. The probability for re-covery is correspondingly high. , For those sequences involving LOSP, the probability for recovery would be lower, but recovery is still feasible. Since the SLC system is considered vital to saf e shutdown, it is powered through a diesel-backed , safe shutdown board and can be started by the operators anytime following the automatic start of the diesels. The CRD hydraulic system operates continuously during normal operation. The operating 1A pump is. tripped upon LOSP and not automatically restarted. The 1B pump (see Fig. 3.1) le
7 j powered through a diesel-backed board and the operator can start it from the main control room af ter the automatic start of the diesels.* In order to prevent overheating of the CRD hydraulic pump the operator would al so have to start a Raw Cooling Water pump to supply cooling water to the CRD hydraulic pump motor and oil cooler.
- *CRD pump 1A is powered f rom the 4 kv Unit Board IC, and CRD pump 1B is powered from Shutdown Board A.
S r e
8
- 3. FLOW CAPABILITY OF SMALIeCAPACITY, HIGH-PRESSURE INJECTION SYSTEMS .
3 .1 Introduction The maall-capacity inj ection systems considered in this report are the CRD hydraulic system and the SLC system. There are other unall-capacity systems which might be utilized under emergency conditions but these systems are not capable of high pressure injection and are, therefore, not considered here. 3.2 CRD Hydraulic Systen Figure 3.1 shows schematically the piping arrangement of the CRD hydraul ic system. There are two high-head centrifugal pumps associated with Unit 1. The 1 A pump operates continuously during normal power operation; the IB pump is a spare that can be aligned to operate with either the Unit 2 or the Unit 1 CRD hydraulic system. In general, the flow injected into the reactor vessel by the CRD hy-draulic system depends on reactor vessel pressure, the number of pumps in se rv ice (one or two), the number of flow paths in service, and the flow resistance in each flow path. During normal power operation the primary . flow path is through the 85-11 flow control station (see Fig. 3.1) to the cooling header, and into the reactor vessel via the 185 CRD mechanism as-semblies. The normal flow is automatically regulated at 3.78 L/s (60 spm) . by the flow control valve, and there is no flow through the charging header or the pump test line.* After reactor scram, the scram inlet valves open, allowing flow into the reactor vessel through the charging header and the ORD mechanism as-sembl ie s. This results in an increased total flow which is limited only by the restricting orifices in the charging line and by the seals in each of the 185 CRD mechanism assemblies. The flow control valve automatically attempts to hold the total sensed flow at the venturi to the 3.78 L/s (60
, gpa) setpoint, but flow is measured upstream of the connection to the charging header, so the 85-11 valve automatically closes af ter reactor scram. This is a desirable control action because it maximizes the driv-ing flow into the under piston volumes of the CRD pechanism assemblies during a scram.
A normal post-scram operator action is to reset the scram. -If this occurs, the scram inist and outlet valves would close, and the inj ected flow would decrease back to 3.78 L/s (60 spa) through the 85-11 flow con-trol station. The scram cannot be rese t in a IUUV sequence because a low reactor vessel water level trip signal will be in effect continuously , throughout the period of laterest. post-scram injection flow can be increased by appropriate operator action. For example, the operator might put flow control valve 85-11 in
- manual and remote-manually open this valve from the main control room.
*It should be noted that the scram accumulators ' float' on the CRD pump discharge pressure via the charging header during normal- operation.
l
= . ~. . _ _ = _ _ _ _ - - - _ - -
9 With the reactor vessel at 6.9 NPa (1000 psia) the resulting total in-Jected flow would be about 14.2 L/s (225 spm) with two pumps running or about 11.2 L/s (178 spm) with one pump running. This tactic is not en-ployed at Browns Ferry because the pump test bypass line, which has a di-rect path into the reactor vessel via a feedwater line, is available. In order to initiate flow into the pump test bypass line, the operator must first dispatch an assistant into the reactor building to open the hand-operated 85-551 valve and to close the breakers for motor-operated 85-50 l valve. The flow begins when the motor-operated 85-50 valve is opened f rom the main control room. The Browns Ferry Emergency Operating Instruction (E01)-41, " Water Make Up Methods to the Reactor Vessel," specifies that the hand-operated 85-527 throttling valve at the discharge of the CRD pumps should be opened fully to maximize the inj ection flow. This instruction is contradicted by
; the System Operating Instruction for the CRD hydraulic system (OI- 85 ) ,
which states that the handwheel has been removed and the valve position should not be adj usted. In this study, both possibilities are considered f or the 85-52's valve: the original valve position might be unchanged and on the other hand, the valve might be opened fully. As demonstrated in . Chapter 4, it is necessary that this throttling valve be opened fully in order to allow sufficient flow through the pump test bypass to avoid fuel damage in a TQUV accident. Figures 3.2 and 3.3 show the calculated post-scram vessel inj ection
- e flow as a function of reactor vessel pressure for, respectively, one pump, j and two pump operation. Appendix A present s the de tail s of the calcula-tions. These results show that significant flow enhancement is possible
, . by operator action, but they al so show that the CRD pump low suction pres-4 sure trip [setpoint 60.7 kPa (-5.9 psig)] can interrupt flow when the re-actor vessel pressure is below about 4.38 MPa (620 psig).* The occurence of a low suction pressure trip is less likely when two CRD hydraulic sys-tem pumps are operating because there is much less flow per pump than when only one puwp is operating.
3.3 SLC Sy s t em The SLC system piping arrangement is shown schematically by Fig. 3.4. In order to initiate inj ection of non-borated water, the operator must first send an assistant into the reactor building to manipulate several hand operated valves to switch the suction from the Standby Liquid Control Tank to the de-lonized water supply. This is dona by closing valve 63-500, and opening valves 63-502, 63-509, and 63-511. Inj ection into the
. reactor vessel is initiated from the main control room by opening the two explosive-operated valves and starting one of the two positive-displace-ment pumps (an interlock prevents simultaneous operation of both pumps). l
*With the condensate storage tank (CST) full. As shown on Figs. 3.2 and 3.3, the pump low suction pressure trip would occur at higher reactor '
vessel pressures if the CST is partially depleted. However, this tank is espected to be nearly full at the start of the accident and holds enough water for some 35 h of inj ection. 1
- . _ . ~ _ , , _ , _. ,,, _,mm .. . . - _ ,4 -- - ~ . , , _ ~ , - , -
10 Since the SLC sy stem is equipped with positive-displacement pumps, the inj ection flow rate is almost totally independent of reactor vessel pr e s s ur e . Section 3.8.3 of the Browns Ferry FSAR specifies that the flow , i s "~50 gpm" ( 3.15 L/ s ) a nd i t is noted on Fig. 3.3-2 of the FSAR that the flow is "between 50 and 56 gpm" (3.15 and 3.53 L/s). 3.4 Summary The CRD hydraulic system and the SLC system can be used singly or in combination. Table 3.1 summarizes the inj ection flows available via dif-ferent combinations of these systems under various circumstances of opera-tor action. e 4 4
l l l 1 i 11 1 Table 3.1 Examples of inj ected flow available by use of CRD hydraulic system and the SLC system Reactor No. of CRDHS njected
. vessel CRDHS inj ect ed pressure punps peratog f g o, flow wf th
[NPa (psig)] running
*** * [L/s (spm)]
ggj ,)) 6.90 (1000) 1 NOA 7.06 (112) 10.6 (168) 2 OA1 7.56 (120) 11.1 (176) 1 OA2 8.63 (137) 12.2 (193) 2 OA1+0A2 9.77 (155) 13.3 (211) 1 OA2+0A5 13.1 (208) 16.6 (264) 2 OA1+0A2+0A3 18.8 (298) 23 (354) 0.69 (100) 1 NOA 11.0 (174) 14.5 (230) 2 OA1 11.9 (189) 15.4 (245) 1 OA2 13.4 (213) 17.0 (269) 2 OA1+0A2 15.2 (241) 18.7 (297) 1 OA2+0A3 +0A4 16.4 (260) 19.9 (316) 2 OA1+0A2+0A3 29.1 (462) 32.6 (Sibi "CRDHS operator action key: NOA = no operator action OA1 = start and initiation of flow through second pump OA2 = opening of the test-bypass line for inj ection directly into reactor vessel 0A3 = opening of the 85-527 throttling valve OA4 = operator actions to avoid low suction pressure trip of ORDUS pumps (bypassing of suction strainers and/or throttling of pump output to control flow to (16.4 L/s (260 gpm). b SLCS flow taken to be 3.53 L/s (56 spm).
12 _ Q 3M 5= s 5 f!
!--- SfI 6 ll: 8J sg 9 $ w5 jos e
1
- hh m v
E vigs 5. .
, il !
- h \ ~~~]s3 $!:"miil! 1
. g:E E i'liAN N! . t~ c - -Q- 2 59 2 mg%gs :
- N N_..<. i ,t '
- s - - a_ t9 .a ca !
\"I ,\
s
!! Eh I z
i \ o 'l "i o i s o
.g g8 i _ . E , .
.c ,' -
s si x x -
. s
- a
.$.,'\\s./?_
87 N e mm y 1 C h Ll C b_h___ h__________! s 1 4
= b
- W g g! 4 g :,l: .g
. , ] (.i e w
?$ Gd Q b5 (b5 k
- am m g j W ===eja v_
_ 8" i l Dl5 bla
!s =g == ==
==!s g -
ey e 5
)=e [eg
-5
13 I ORNL-DWG 83-12050 350 -
-22 A - NO OPERATOR ACTION 8 - OPERATOR INITIATES PLAf' TEST BYPASS -20 m BUT DOESpri ADstJST 85-527 VALVE E 300 -
C - SAME AS B EXCEPT 85-527 WLVE IS T v g OPEG FULLY -18 h v d 250 ' PUMP TRIP.
-16 $
$ l g o y l
-14 5 g 200 PUMP TRIPee -----------------
l
------------- h -12 e S
o i U
< B <
' -10
E " 15 0 - lA O o w ' -8 + z z
. O_ 10 0 -
-6 9 y ' Low suction pressure trip, with CST fut y t4J taJ 7 ** Low suction pressure trip, with CST empty -4
. E 2 -
E' i -2 i 0 ' ' ' ! ' I o O 200 400 600 800 1000 REACTOR VESSEL PRESSURE (pSig) I l i I I I l l 0 1 2 3 4 5 6 7 REACTOR VESSEL PRESSURE (MPo) Fig. 3.2. GD hydraulic system inj ection with 1 pump operating. l e l l I
14 ORNL-DWO 83-12049 700 - 650 -
- Low suction pressure trip, with CST ful _
T 600 -
** Low suetion prensure trip, with CST empty 7 v
8 550 - -35 O j ,.W..TF1P_e_______________________________________
$ -30 m y 450 -
y 400 fUMP IRIPee
-25 8 g 350 -
c W w m 300 - -20 E O O r 25'
-15 8
5 200 - g ! . P H o 15 0 -10 0 W - A - NO OPERATOR ACTION ._ N Z 30 0 _ B - WATOR INITIATES N TEST BYPASS ~ Z BUT DOESMT ADJUST 8S-527 WLVE 50 - C - SAME AS B Ex!TPT 85-527 WLVE IS oPENEo ruLLv 0 O O 200 400 600 800 1000 REACTOR VESSEL PRESSURE (pSig) ( I I I I I i l 0 1 2 3 4 5 6 7 REACTOR VESSEL PRESSURE (MPa) Fi g. 3.3. CRD hydraulic system inj ection with 2 pumps operating. D e i r
15
- I a
a 1 g , l l. ( :: t
!lI 'llT P
- St
- o -
a3 "l<\ O'y
!:l 6 s
a -.n , a g
- r -
\
f 1 > Ej pu% U28 ) ) % x :: ' 4 & W v -6 "'
, ,g = -
f l u %, , a 8 v .as P 1P O e
, J6 o 1r 1P 4 a o , 55
- 'E :::'E 5
~ ~
3I % ll ll I .i--1.s L I. s
= ?
e 2
' f:
. J C
< ,.e M
- :C i mm mm 44
> 1rp $ si r Jk 3 Jb u
k J
, .L
,g.
e sg 9 nr
.. . .. .- - - . . ._ . = . ._. . _ - , . - ~ ; 16
- 4. PARAMEIRIC STUDY OF 111E EFFECT OF SMALL INJECIED FLOWS IN MITIGATION OF TQUV SEQUENCES 4.1 Introduction Two series of calculations are discussed in this chapter: the first series assumes that the reactor vessel remains at pressure; the second assumes that there is a f ast depressurization caused by the opening of i
multiple SRVs. The computer runs within each series differ only in the flow injected to the reactor vessel. The result s vary from no core uncovery to extended core uncovery with severe fuel damage. The effect of delay in the initiation of the assumed injected flows is not considered in this chapter. Calculations which determine the ef-fect of delays that might accompany operator initiation of flow via the small capacity high pressure inj ection systems are reported in Chap. 6. 4.2 Reactor Vessel Remains at Pressure i This section presents the results for the at pressure series of runs. Subsection 4.2.1 discusses the BWR-LACP calculations of the time to core uncov e ry. The MARCH code calculations are initialized at the time of un- . ! covery predicted by BWR-LACP. The MARCH result s (Sect. 4.2.2) de termine the condition of the reactor core during the period of uncovery. Concin-sions regarding the minimum amount of flow required to prevent fuel damage , are discussed in Sect. 4.2.3. 4.2.1 Reactor Vessel Remains at Pressure - Time to Core Uncovery The BWR-LACP calculations begin 30 s af ter the initiating scram from full power. The water level in the downconer is initialized at 12.70 m (500 in.) above vessel zero (see Chapter 8 for a discussion of uncertainty in the initial vessel water level). Figure 4.1 shows the preesure response which is typical for each of the calculations in this series. This response reflects the operator training to utilize remote-manual actuation of the SRVs in such a manner as to avoid the automatic actuation [which would occur at a pressure of 7.72 MPa (1105 psis)]. When pressure reaches 7.59 MPa (1085 psis) the operator opens a single SRV and allows it to remain open until pressure is ! reduced to 6.21 MPa (885 psig). This cycle is repeated throughout the cal cul ation. During the first 10 min of each run, the injected flow is 7.06 L/s (112 spm). The CRD hydraulic system will deliver this much flow af ter a .-
- l. scram without any operator action (i.e., if there is no loss of offsite ;
power). The inj ected flow is assumed to be increased from the initial 7.06 L/s (112 spm)- to a higher value af ter 10 min and to remain ~ constant . at that higher value throughout the remainder of the calculation. Figure 4.2 shows the downcomer water level for four different values ,
- of the higher . inj ected flows ranging from 10.71 to 14.18 L/s (170 to 225
spm). The _ time to core uncovery beenmas longer as ~the flow is' increased; l l
17 for the 14.18 L/s (225 gpm) ca se, there is no core uncovery. Downconer water level is shown because this is the level measured by the instruments
, in the control room. During shutdown, with the core on decay heat, the collapsed water (or water / steam) level in the core will approximate the level in the downcomer. During periods of very active steaming, for , example, during SRV discharge with the reactor vessel at full pressure, the level of the two phase mixture in the core could be considerably above the water level in the downcomer. The initialization of the MARCH runs is based on the downcomer water level.
4.2.2 Reactor Vessel Remains at Pressure - Assessment of Core Damare The primary purpose of these parametric studies is to determine the extent of core damage as a function of total inj ected flow. The intent was to vary the inj ected flow so as to cover the entire range of core dam-age; from core melt and reactor vessel (RPV) failure to core recovery and no core damage. It became apparent af ter a few MARCH runs that there are two distinct subregions between the two above extremes:
- 1. a region where the core recovers but there is some core damage (i.e., measurable metal / water reaction and probable rod cladding f ailure) but the core geanetry remains intact;
, 2. a region where the core recovers but there is severe core damage (i.e., significant metal / water reaction and probable rubble bed formation).
i
- The method chosen to distinguish these regions of differing core dam-age relies on two variables computed by MARCH; (1) the total percentage of cladding reacted, and (2) the maximum core fuel temperatures observed dur-ing the MARCH simula tion. The basis for determining the extent of each subregion is outlined below:
Reaion Criteria Core recovery, no damage Negligible cladding reacted, fuel temperatures <1366 K (20000F) Core recovery, some damage Fuel temperatures (2033 K (32000F) core geometry remains intact and >1366 K (2000*F) Core recovery, severe damage Fuel temperatures >2033 K (3200*F) probable rubble bed formation Core melt, RPV failure RPV failure predicted by MARG e The method is applied by overlaying the percentage cladding reacted and maximum core temperatures vs the total vessel inj ected flow rate. That is, a MARCH run becomes a single data point in each of two overlay curves and therefore the parametric MARCH study reduce s to two curves against which the core damage assessment criteria are applied. Two "at pressure" scenarios were studied. In the first scenario the operator manually actuates an SRV to control the vessel pressure between 6.21 and 7.59 MPa (885 and 1085 psig) as shown in Fig. 4.1. In the second
- 18 scerario the SRV is alluwed to automatically actuate [i.e., the pressure would cycle be tween 7.38 and 7.72 MPa (1055 and 1105 psig)].
The second scenario was more extensively studied and the assessment , of the core damage as a function of total injection rate is shown in Fig. 4.3 . Bel ow an ~9.15 L/ s (145 gpm) inj e ct ion, the core will probably melt and slump with subsequent RPV failure; for an inj ection rate above ~12.30
- L/s (195 gpm), the core will recover with negligible damage.
For the scenario in which the operator manually actuates an SRV to control the vessel pressure (6.21 to 7.59 MPa cycle), the core damage as-se s sment ' curve' is given in Fig. 4.4. Core recovery (with no damage) is predicted for inj ection flows >~12.62 L/s (200 gpm) and the RPV is pre-dicted to fail for inj ection flows (~10.4 L/s (165 spm). Thus, in a TQUV sequence, the longer pressure cycles increase the amount of inj ection re-quired to avoid core damage. This is because of the longer time available f or the uncovered portions of the core to heat up be tween SRV actuations. The damage asse ssment " curve" for the 6.21 to 7.59 MPs pressure cycle is noticably to the right of the corresponding curve for the 7.38 to 7.72 MPa cy cl e (compare Figs. 4.3 and 4.4) and the slopes of the cladding re-acted curve and core temperature curve are increased which leads to com-paction of the intermediate core damage regions. This difference in the curves is due entirely to the increased quiescent period in the vessel between SRV " pops" for the case with operator actuation of the SRV; that is, when the SRV rescats af ter an actuation, the vessel enters a repres-surizing period in which there is boiling but the steaming rate is not
- vigorous enough to adequately cool the fuel rods and the fuel temperature increases. Therefore, for the operator-controlled 6.21 to 7.59 MPa cycle, it take s longer for the vessel to repressurize, thus allowing higher fuel
- temperatures during the repressurization period, which also results in greater metal / water reaction (if the necessary steam is available). With the fuel at an elevated temperature when the SRV is actuated, a great amount of steam is generated and the level swells dramatically. However, the heated upper portions of the fuel rods are harder to quench at the elevated temperature and there are now competing proce sres; that is, the metal / water reaction heat input may offset the cooling effect of the steam / water mixture - the result is that the rods may stay at an elevated temperature during the SRV " pop" and then begin heating up from that point af ter the SRV resents. Below the point of core recovery with no damage, the core heat-up is accelerated as the length of the repressurization cycle is increased; for instance, at 10.66 L/s (169 spm), the top of the core starts to acit at 256 min in the 7.38 to 7.72 MPa cycle, but starts to melt at 192 min in the 6.21 to 7.59 MPa cycle.
4.2.3 Reactor Vessel Remains at Pressure - Conclusions In the at pressure parametric studies, if the low capacity inj ection i sy st em fl ow can be increased to as much as or greater than 14.18 L/s (225 gpm), there will be no core uncovery (see Fig. 4.2) . However, for flows (14.18 L/s, there will be a period of core uncovery from which the core
- might or might not recover. Figures 4.3 and 4.4 present an assessment of the core damage for those cases in which the core uncovers.
19 For the scenario in which an SRV is allowed to automatically actuate (7.38 to 7.72 MPa pressure cycle), the core will probably melt and slump
- and the RPV subsequently fail for total inj e ct ed f low s (~9.15 L/ s (145 3pm). For inj ected flows >12.30 L/s (195 spm), the core will recover with negligible damage. In between these limits the core will probably recover but with varying degrees of damage (see Fig. 4.3) .
For the scenario in which the operator manually actuates an SRV to control the pressure (6.21 to 7.59 MPs cycle), core recovery with no dam-age is predicted for inj ection flows >12.62 L/s (200 gpm) and the RPV is predicted to fall for flows (10.4 L/s (165 spm). The heat-up of the core (with resultant core damage) is accelerated in this scenario primarily because of the increased quiescent period in the vessel between SRV
" pops".
Up to this point, the study has been essentially an academic exer-cise; that is, the inj ection rate to the RPV has been varied without any regard to the actual capabilities of the low capacity injection systems. j To make any realistic conclusions about the survivablility of the BFNP i Unit I reactor during a TQUV accident sequence, one must integrate the core damage asse ssment curves from Figs. 4.3 and 4.4 and the injection capabilities of the CRD and SLC systems from Table 3.1. To this end, a "real world" picture of the situation is presented in Fig. 4.5 in which the asse s sment curves (from Figs. 4.3 and 4.4) are over-laid with actual BFNP Unit 1 injection capabilities for different CRD and SLC system alignments with varying degrees of operator action. It should be noted that these inj ection rates are those corresponding to a RPV pres-sure of 7.58 MPa (1100 psia) whereas Table 3.1 was generated using an RPV pressure of 6.90 MPa (1015 psia). The flows associated with the centri-
)
fugal CRDil system pumps are slightly lower at the higher discharge pre e- I s ur e , but the difference is not significant to the damage assessment. Ref erring to Fig. 4.5, the case of no operator action results in a ) l core melt and subsequent RPV f ailure and containment f ailure (the results of this study will be presented in detail in Chap. 5); however, it should be noted (stealing some results f rom Chap. 5) that the effect of the in-jection from the low capacity CRD system in the no operator action case (without LOSP) does delay RPV f ailure by 70 min and containment f ailure by 134 min over the case with no injection at all. Therefore, even with no operator action, the effect of the CRDH system will significantly impact the progression of the accident and the timing of fission product re-leases. Also referring to Fig. 4.5, there are at least three reasonable CRD and SLC system configurations (and realistic operator interaction) which will result in core recovery and 30 core damage. Some of these scenarios will be discussed further in Chapters 6 and 7. 4.3 Seauences with Ranid Reactor Vessel Denrossurization , This section presents the results for the rapid depressurization series of runs. This rapid depressurization could result from emergency operator action to open several SRVs or to actuate the automatic depres-surization syste:s (which controls six SRVs) in order to reduce vessel pressure suf ficiently to allow inj ection by the low pressure injection l l l
20 sy s t em s. In a TQUV sequence, no low pressure inj ection systems are opera-tional, so a rapid depressurization would not be an effective emergency maneuver. , Subsection 4.3.1 discusses the BWR-LACP calculations of the time to core uncovery. The MARCH calculations are initialized at the time of un-covery predicted by BWR-LACP. The NARDI result s, Sect. 4.3.2, de termine the condition of the reactor core during the period of uncovery. Conclu-sions regarding the minimum amount of flow required to prevent fuel damage are discussed in Sect. 4.3.3. 4.3.1 Rapid Deoressurization Cases - Time to Core Uncovery The BTR-LACP calculations for these runs were initialized at the same starting point as the at-pressure cases. For the first 5 min, the reactor vessel remains at pressure and the inj ected flow is 7.06 L/s (112 spm). Af ter 5 min, a rapid depressurization begins and the inj ection flew is in-crea sed.
- The flow remains constant at the higher value for the remainder of the calculation.
Within 10 min af ter initiation of the rapid depressurization, the reactor vessel pressure (Fig. 4.6) is below the point at which the low pressure ECC systems would begin inj ection if they were functional. The rate of depressurization using 3 SEVs is slower than would result from opening of the six SRVs that are dedicated to the Automatic Depressuriza-
- tion System (ADS); the ADS does not actuate during a TQUV sequence because the low pressure ECC systems are, by definition, not running.
As a result of the rapid depressurization, water in the reactor ves-sel flashes to steam and this loss of water inventory results in a much quicker uncovery of the core than would occur if the vessel remained at pr e s s ur e. The time for downcomer water level (Fig. 4.7) to reach the level of the top of the active fuel is not very sensitive to the rate of inj ected flow because the primary mechanism for uncovery is the loss of water inventory due to flashing. 4.3.2 Ranid Doorossurization Cases - Assessment of Core Damare The purpose of these parametric studies was basically the same as those of Sect. 4.2.2; that is, to determine the extent of core damage as a function of total RPV inj ection flow. The damage assessment technique employed in this section is outlined in Sect. 4.2.2. The assessment of the core damage as a function of total inj ection rate is shown in Fig. 4.8. Bel ow ~10.85 L/s (172 spm) inj e ction, the core will probably melt and slump with subsequent RPV failure; for an inj ection rate above ~14.2 L/s (225 spm), the core will recover with negligible , dam a ge .
~
*By operator action, but same increase would also occur because the ORDHS centrifugal pumps would be pumping against a lower pressure. The exact mechanism for the flow increase is not important here because these runs are intended to be a parametric study of the effect of injected flow.
1 l 21 j 4.3.3 Rapid Depressurization - Conclusions i
- Within 10 min af ter initiation of rapid depressurization, the reactor vessel pressure (Fig. 4.6) is below the point at which the low pressure l ECC systems would begin inj ection. However, during the TQUV sequence, l
- these systems are by definition not running. As a result of the rapid depressurization, much of the reactor vessel water inventory flashes to steam, thus resulting in a much quicker core uncovery (Fig. 4.7) than would occur if the vessel remained at pressure (Fig. 4.2). For this sce-nario it would require ~31.5 L/s (500 gpm) inj ection rate to keep the core covered, and for flows (31.5 L/s there will be a period of core uncovery.
Figure 4.8 presents an assessment of the core damage as a function of the inj ection rate. For inj ection flows >14.2 L/s (225 gpm), the core will recover with negligible damage, and for flows (10.85 L/s (172 spm) the core will probably melt and slump with subsequent RPV f ailure. The assessment curve from Fig. 4.8 can be combined with the known inj ection capabilities of the BFNP Unit 1 low capacity injection systems to give Fig. 4.9. Referring to Fig. 4.9, no operator action [10.6 L/s (168 spm) inj ection flow] will result in severe core damage if not core slump and RPV failure. However, with proper operator action (there are several reasonable alternative actions in Fig. 4.9), core damage can be otally averted. I 9 0 e
0 . D 0 T 3 E 9 2 7 4 . 3 8 0 G i7 W 2 D L N R O 0 4 a2 e r u s s e 0 r 1 p i2 l e s s e a8 0 v 1 r o
) ) t N c a
I e f A 0( M r . FO a5 OT 1 E l NS M s OE I I I e 2 TL2 T i r . AP4 P U AG e T 0 s CEFI i2 1 ACA N e r RR I O T N u T W s AE O s l R R H e b E U S r P S e0 p O SS E E 9 YR S t BP AC A L S OH R I RT T U O 1 N - F 0 O - E i6 4 CVH ER T g
% RSF UEO i SL F S G H E C i RINA i3 0
tf PSE
,l J
0 . 8 o8~ 8* ge 8" e8 8* oy 8n k b O _C )yb ats$W gWiu - - ~ - - - _ - e s * - n c _ O 2g JE8W'Ey (
e e e e e o ORNL-DWG 83-4730 ETD 8 ts
- o. NOTE
~
FOR ALL CASES. FLOW INJECTED INTO REACTOR VESSEL = 7.061/s (112 gpm) FOA g- FIRST 10 mm. AFTER 10 min: A: FLOW = 10.711/s (170 gpm) B: FLOW = 11.341/s (180 gpm)
%- h- C: FLOW = 12.601/s (200 gpm)
D: FLOW = 14.181/s (225 gpm) g o .. o .
- to
^
N J- Z 8-u J J U D J yo_
>o U o- '-
g- Je b J w -._--......-- --------_-------------_-------------.------------_..--------------..F.U.E.L--. TOPOF ACTIVE U
$ n~ _
_c W _ _
-~ B co - > A g ..
8-m
- - - - . . . . _ . . . _ - _ _ _ - - _ _ _ - . . . . . _ . - - - . - _ - - - - - . . . . . . . . _ _ - - . _ . . . _ . . . - _ . - . 99119 9-Q F ,A_C_T I V E _F_U E_L_ _ ,
8-n
- 8. -
8
- i i , . . 4 i i i 0 30 60 90 120 150 180 210 240 20 300 TIME (MIN)
Fig. 4.2. At pressure series - downconer water level.
-_ ~
ORNL- DWG 83-4 731 ETD PROS. CORE RECOVERY SIGNIFICANT CORE DAMAGE PROB. RUBBLE BED CORE MELT : : : : : : CORE RECOVf HY
'iPV F AILURE NO DAMAGE (K) ( F)
~
CORE RECOV. SOME CORE DAMAGE PROB. CLAD FAILURE [- CORE GEOM. INTACT 20 -
- 250c _ ,ogo O
- w S 5 y - 2'Kn Q
g - 3000 c: a: a o 2 5
- u y 9 1500 o &
O 10 - _ 2000 s i 3 3
\ - 1000
\
g 1000 N N
-====TSA T - 500 0 l l l i N % l l 300 100 120 140 160 180 700 220 240 760 (gpm)
, I I i i i l I i l l 7 8 9 10 11 g gj,3 12 13 14 15 16
> > TOTAL INJECTION RATE Fig. 4.3. Core damage assessment for the RPV maintained at pressure (SRV automatic actuation with pressure cycles of 7.38-7.72 MPa).
l 1 . . . . . -
e . e e e . ORNL-OWG 83-4732 ETD PROS. CORE RECOVERY SIGNIFICANT CORE DAMAGE 30 - PROB. RUBBLE BED CORE RECOVERY
- NO DAMAGE CORE MELT '[ \ ' (K) (OF)
RPV F AILURE 1 \ CORE RECOVERY - 3000 - 5000 , ', SOME CORE DAMAGE PROS. CLAD F AILURE CORE GEOM. INTACT
-'2 # 4000
_ 20 - d 3 E G B 6 - 2000 $
-3 OE 5 -
3 r o w
- 1500 o u
10 - - 2000 s 3 3 X
- \ - 1000 $
\
N 1000 TSAT - 500 )
.0 i i i i i 300 l
_o l 120 140- 160 180 2 220 240 260 2ts0 (gpm) i I i i I I I I i I 8 9 10 11 12 13 14 15 16 17 (Lis) TOTAL INJECTION RATE Fig. 4.4. Core damage assessment for the RPV maintained at pressure l (operator actuation of SRV with pressure cycles of 6.21-7.59 MPa). l t. V _________ ________________ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _____________
3 'i
)
$' S 1
NOCRD P U
#/
g i pd NOSCR OP AM MP
- , k
/
+ # ~2CRD ERATRESET NO V UP OR A g '
g ALVMPS EM CT ION
*/
A a , p '""'""*i ANtP P I, ss 0 U CROP V MPT UULATION MP g ALVEEST e4 e D / g4bp6t 85-5 BYP '
# 7 #, 27 ASS e.
, 4 fy -2CROPU P OP NOT ENED 3r
- Y
%o V MPT U ALVEEST S MP MOVED A g
*p o
# O' 85-5 BYP p* '
O 99 0p 27 Oj s
/ e$ ASSOP 3% # GAD _I NOT M ENED o f7o #
NOCRDV UP OVED ' p ~1CRO V P
- C UMP ALVE MP+S g
j ALVES 8 MA LC 511 NIP s ,
/ g./ef 9 6S 9 AND 8 5527ULATION e
$, 1 P CRDP F
@j ,- [ V UMP UM ULLY gS 993 ALVE8TEST P+S BY LC OP gg EN 1 5-5 P P CRD P 27 ASS OP
- S T ,
h$4 gs _,, 2 V UMP UMP 9 4 ( 4, g 4@ gi CRO P ALV TE NOT M ENED
' 9 P E 8STBY OVED p9 t,' & [ e , V UMP UMP 5-5 P
, E, d> 6,
. ~2 ALVE V CRDP 8 TESTB S 27FASS U OPE
+SLC 4
3 a
,'*c c \ 49h pT L
A.VES UMP 5 S 27 YP 55 ASS OP LLYOPNED EN
# s 51 1 NOT M ENED 4' Y j V
p ~1 AND 8 OV ED e V CRD P
" e U 5527 ALVES F r 8 MP+S UL f 51 LC LY 0
4
$b oO 1
AND 8 OP EN
,9 6 5527 F b*
- e Y,
+
p o 4+ ggs ULLY k, 1 OP EN
/
P CRO V UM P P [ U 4 MP
- A 4,p 9 LVE
^ 2 TEST 855BYP +SLC
' "V CRD P 27 FASS
,s OP UMP ULLy y ALVES8 opENED og 51S+S1 LC gN 1 *
~P U ANDB s'
y e% 'g #
$, V MP 5527 F
r
, t, h, ALVE TE 8 ST 5-527BYP ULLY OP EN ASS i v ,, F OP st *%<
- e 2
PU CRDP ULLYENED OP
. g Y V MP us3P EN ALVE8TEST B S+S 5527YPASSLC k F OP t
ULLY OP O ENE EN s a
+,sx o 9E
'I
- e o e e ORNL-DWG83-4734 ETD
~
DEPRESSURIZATION BEGINS AT 5 min. OPERATOR OPENS FIRST ONE AND THEN AN N' ADDITION AL TWO SRVs
-A.
H-
@~ NOTE: THIS PRESSURE TR ACE APPLIES TO E ACH OF THE FOUR CASES SHOWN ON fig. 4.7 g,
E- a H-E G M H-
$ ,._ $ H-M t:
E g ., _ 'p g-e a h-PRESSURE CONTROL BY OPERATOR N- ACTUATION OF A SINGLE SRV g_ - V H- N nd 2SRV SHUT, REOPENED.THEN SHUT g_ 3rd SRV SHUT o- o . . . . . . . , , 0 30 60 90 120 150 180 210 243 270 300 TIME (MIm Fig. 4.6. Fast depressurization series reactor vessel pressure.
ORNL-DWG 83-4735 ETD es , 8 -.
- O m.
8-m FOR ALL CASES. F LOW 1NJECTED INTO RE ACTOR VESSE L = 7.06 Us (112 gpm) FOR FIRST 5 min. AFTER S men: I~ h' A: FLOW = 12 60 L/s (200 gpm) B; F LOW = 14.18 L/s (225 gpm) C; F LOW = 16.13 L/s (256 gpm) r D; F LOW = 31.51 L/s (500 gpm) h-a n- E y u as - J 3 W _2 >o y o ., Mo-(n -+ I* b. J g1 a ... . ..............................................................T.O.P.O.F..A.C.T.IV.E..F.U.E.L...... , U1 m - C u tn
- oo S ..
,_ ca
. . . _ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .B.O_T_T_Op] ,O_F A CT I V E_ F U E L , _
o W t 8-8 . . . . . . . , . 0 30 60 90 120 150 180 210 210 2~0 300 TIME (MIN) Fig. 4.7. Fast depressurization series - downcomer water Icvel.
CANL- DWG B3 4 736 E TD PROBABLE CORE RECOVERY SIGNIFICANT CORE DAMAGE 3o _ [ PROBABLE RUBBLE BED
- 7: :: : :
CORE MELT CORE RECOVERY RPV F AILURE NO DAMAGE CORE RECOVERY SOME DAMAGE gg) qur, PROB CLAD F AILURE _ _ em l CORE GEOM INTACT l l t
\ - 2s* _ 4 0m 20 _ ?
e \ U E
< 3 E
\ -
20 % N e - Joa0 m 5 g
\
- 9 E w 10 _
N} _
- 15m
- 2000g 8
E E 1
- 1000
- 1000 l
i
~
Tsar O 1 I I I I I - I L - 300 _ o 120 140 160 180 200 220 240 260 25 300 (qom) l l l l l l l l 8 l 1 8 9 10 11 12 13 14 IS 16 17 18 (L/d TOTAL INJECTION RATE Fig. 4.8. Core damage assessment for the rapid depressurization of the RIT with subsequent operator control of the vessel pressure at 1.3 8-1. 7 2 11Pa .
B- g o O 1 e Y
', &* j 4 _N CRD O NOSCR MP P
U
,. 0 3 OP AM @ D,
- e ,
,A k _2 ER AT RESET OR
/
4 c' Vg NOCR0P V UMP ALVE ACT ION p ( O, MS ANIP . g - U
#o
%fY .p LAT pg
/,
Dp j [ N[p4 e ION 7 #
, o , e Jp F gt GY d
M p _1
, ~
A
' 2 NOCRDP a op _P CRO U P V tUM
# O #, 4 V MPTVMP AVEP+S M LC
/
j
# 6 ALVE8EST S ANIP
( BYP U
# 4 5-S27 LAT Y ASS
-O ge kO p $
O M NOTOVED OP ENED fON
- o - 4 hD [ 1 o 4 P CRO
/
- _B UMP PUYP O o h #-
4 THYP ASSTEST Q#g LO OF SYP AF ROTT 85LSO W L AT IN CT SU OP ASS a 1 260 OF 8 - 11 G ION ENED F CRD gom85-5 P STR
,p @ g g G* UMP T ULLY MP U27 AtN TOER$
, 2 EST POP W
# _P CRO BY EN tTH V CONT 4
U V MP PUMP ASS RAND
- 7 , ALVE S ANDNO ALV E O.t /OR
/#
4p f TESTS 8 55 YP +SLC 27 ASS OP 4o e 1 M NOT ENED O@g CRD P OVED s, $# u p ' TEST BYP BYP UMP
, , T ASS OF ASS OP +S,C t
HROfT P
$ CRO F UMP 4
f e g ALSOLOWLtNGSuCTENED WIT OF ION
'd s,- H I iTHNO P AT PU 85 STR
,% 0 2 VALVCRD UMP VP 260 427 AIN
+gpm TO ERS V CRO E T SLC A
+ ALV PU 8 EST 1 BYP CON 0'O
+ ss8 MP S 51F TROL R
. f 2
511 UL ASS LY AN f OP O V CRD AND8 EN P p ALV U 5E27 S [f ES8MP F j o 51S+S1 LC UL LY 2 ANDP OP f P CRO EN U S$27
- V MP PUMP F ALVE8TESTS ULLif BYP OP 5527ASS F OP EN Ut e tY ENED OP EN OE
- l l
l 31
- 5. NO-OPERATOR-ACTION SF4UENCES e
5.1 Introduction
- This chapter analyzes the effect of the small-capacity high pressure injection systems on TQUV sequences in which there is no operator action.
Under this ground rule, only the CRD hydraulic system would provide inj ec-tion because the SLC system can only be initiated by operator action. In the event of a loss of of f-site power (LOSP), there would be no inj ection even from the CRD hydraulic system because the operating CRD pump for Unit 1 is automatically tripped upon LOSP and operator action would be required to restart a pump. The CRD hydraulic system continuously pumps water into the reactor vessel during normal operation and will continue to pump af ter a reactor scram (without LOSP) . As discussed in Sect. 3.2 and shown in Fig. 3.2, l the po st- s cr am injected flow for the case of no operator action varies l from 6.55 L/s (104 spm) at a reactor vessel pressure of 7.69 MPa (1100 psig) to 11.34 L/s (180 gpm) at 0.1 MPa (0 psig). This flow is not enough to prevent severe fuel damage for a TQUV accident seque nce initiated from 100% reactor power; however, as analyzed in the following subsections, it is enough to have a significant effect on the timing of critical events in the resulting severe accident se que nce . In Sect s. 5.2 and 5.3 accident scenarios are discussed which result (as predicted by MARCH) in RPV f ailure and subsequent containment f ailure.
- It should be noted that one of the greatest uncertainties in application of MARCll to BWR accident analysis is the MARCH treatment (i. e. , model s) of in-vessel structural deformations and f ailures. MARCH allows the user to select one of three nonnechanistic core melt models* and essentially allows the user to specify when the core " collapses." (The entire core is assumed to collapse when a specified input fraction of the core becomes moltea.) None of the three MARCH models treat the rod, canister, control blade s sit / flow / freeze phenomena or core collapse mechanistically. Al so, due to the design of the BWR core support system (see Appendix B), it is very unlikely that a BWR reactor vessel would f all via a corium attack on the vessel lower hea,d. The corium would more probably fall the vessel by crosion of. one '(or several) of the lower head penetrations (i.e., URD guide tube s, stub tubes, in-core instrumentation tube s, etc.). The point is that MARCH handles these phenomena simplistically and with regard to BWRs, nonrealistically. Thus the times specified in Sects. 5.2 and 5.3 for core slump and RPV failure should be used for comparative purposes and not for realistic estimates of when such an event would occur.
, *Mel t down Hodg1_j, It is assumed that the excess heat in the molten pool above that required to j ust keep the pool' molten is transferred down-ward. , Meltdown Model B2 It is assumed that the excess heat in the molten pool is transferred upward, and none is transferred down.
Feltdown Model C. It is assumed that when a fuel node melts, it in-mediately falls to the bottom of the pressure vessel. s
32 5.2 Reactor Vessel Remains at Pressure Two cases are reported in this section. In the first case, the flow e inj ected into the reactor vessel average s 6.68 L/s (106 gpm) and the re-actor vessel pressure is controlled at an average pressure of 7.55 MPa (1080 psis) by the automatic actuation of a single SRV. For the second . ca se , the reactor vessel pressure control is identical but there is no inj ected flow. The BWR-LACP calculation was initialized 30 s af ter the reactor scram f rom full power, with an initial downcomer water level of 12.7 m (500 in.) above vessel zero. The reactor vessel pressure (Fig. 5.1, curves C and D) cycles be tween 7.72 and 7.38 MPa (1105 and 1055 psig) due to the auto-matic actuation of a single SRV. The downcomer water level (Fig. 5.2, curves C and D) decreases steadily. For the case with CRD hydraulic sys-tem inj ection, the downconer water level reaches the level of the top of the active fuel about 10 min later than the case without any inj ection. The MARCH runs discussed below were initialized with downcomer water level at the level of the top of the active fuel [9.14 m (360 in.) above vessel zero] at the times predicted by BWR-LACP; 28 min af ter reactor scr am for the no-inj ection case and 37 min af ter scram for the case with inj ection. For the first case, the inj ection average s ~6.68 L/s (106 gpm). As noted in the parametric studies in Sect. 4.2, this flow is insufficient to prevent core melt / slump and subsequent RPV f ailure [9.15 L/s (145 spm) is , required to avoid RPV failure]. Melting begins in the core at 105 min (af ter reactor scram) and the core slumps at 276 min (Fig. 5.3) . The col-lapsed water level (Fig. 5.4) decreases in a sawtooth f ashion until core slump occurs; each sawtooth drop in Fig. 5.4 occurs with an SRV actuation and the loss of water inventory due to flashing. Prior to each SRV ' pop' the water level increase s because of the CRD inj ection. The first sig-nificant metal / water reaction starts at ~82 min (Fig. 5.5). A tabular comparison of the MARCH code results for Case 1 (injection flow of 6.68 L/s, described in the previous paragraph and in Figs. 5.3-5.5) and Case 2 (no inj ection flow) is given in Table 5.1. For Ca se 2, which is the " classic" TQUV accident, the total time span be tween reactor scram and containment f ailure (due to overtemperature f ailure of the dry-well electrical penetration assemblies, the uncertainty of the overtemper-ature f ailure criterion will be discussed in Chap. 8) is only 267 min. The initial core melting in Case 1 is slightly (9 min) later than Case 2, but the core slump is 131 min later (4.6 h as compared to 2.4 h) and con-tainment f ailure occurs at 7.17 h (Case 1) compared with 4.45 h for Case 2. Even though the inj ection rate in Case 1 is not suf ficient to pre-vent core welt and RPV fcilure, it does significantly delay these events. 5.3 Stuck Onen Relief Valve (SORV) The two cases reported in this section are identical to those of the . preceding section, except that an SRV is stuck open throughout the se-que nce . As a result of the SORV, the reactor vessel pressure is much low-er and this results in a greater injection flow by the CRD hydraulic sys-tem (see Fig. 3.2) . l
33 The BWR-LACP calculations show that reactor vessel pressure (Fig. 5.1, curves A and B) decreases rapidly at first and then less rapidly as the vessel continue t to depressuriz e. Even though the lower vessel pres-sure results in a greater CRD inj ection flow, the downcomer water level (Fig. 5.2, curves A and B) decreases more rapidly than the corresponding a t pressure case because of water inventory loss by flashing. The MARCH runs, described belew, were initialized with downcomer water level at the level of the top of the active fuel at the times pre-dicted by BWR-LACP,19 min af ter scr am for the case with inj ection and 17 min after scram for the case without inj ect ion. For the case with inj ection, the pressure decreases rapidly with a resultant increase in CRD inj ection (approaching 11.36 L/s (180 gpm) at 60 min). How ev er, the increased inj ection rate is able to halt the rapid loss of water inventory only af ter the level has dropped. to 6.15 m (242 in.) above vessel zero (Fig. 5.6) at ~57 min into the transient. At this point 3.09 m (10.14 f t) of the core is uncovered and, even though the steaming rate is high, there is insufficient cooling of the upper part of 4 the core; the core begins to heat up very rapidly af ter ~29 min (Fig. 5.7) with melting beginning at 79.5 min. Core melt progresses rapidly and is delayed only by good steam cooling in the bottom-middle of the core. The increased inj ection flow is more than off set by the increased core metal / water reaction heat input and the progression of the slumping heat input (MARCH mel t model A) downward; thus, the water Icvel remains essen-
. tially constant until 160 min, when it starts a steady decrease until core slump at 206 min.
A tabular comparison of Case 1 (SORV with inj ection) and Case 2 (SORV
. without inj ection) is given in Table 5.2. For Case 2 the total time span between reactor scram and containment f ailure (due to overtemperature j failure of the drfwell electrical penetration assemblies) is 286 min. The initial core melting in Case 1 is 7 min later than Case 2 but the core slumps 95 min later (at 3.4 h as compared to 1.85 h) and containment f ail-l ure occurs at 7.87 h (Case 1) compared with 4.77 h in Case 2.
The inj ection f rom the CRD system does not prevent core mel t and RPV failure in Case 1 but it does significantly delay these events. 5.4 Conclusions For the TQUV sequences (with and without SORV) in which there is no operator action, the small-capacity high pressure inj ection systems (only the CRD system is included since operator actton is required to bring the SLC system on line) cannot prevent core melt and subsequent reactor ves-sel failure. However, the CRD system can significantly delay RPV f ailure and the resultant containment f ailure. I
- In the sequence in which the reactor remains pressurized and there is no inj ection, the containment f ails in 4.45 h; but with the ef fect of the
! CRD inj ection included, containment f ailure is delayed to 7.17 h. For the TQUV sequence with SORV, the containment f ails a t 4.77 h with no inj ection, and with CRD inj ection f ails at 7.87 h. f
34 Table 5.1 4 Comparison of accident event timing for the at pressure cases (with no operator action) , Case 1 Ca se 2 Start of fuel melting 105 96 , Core slump 276 145 RPV head f ailure 280 211 Containment f ailure# 430 267 aMinutes af ter reactor scram. b Pressure control be tween 7.38 and 7.72 MPa (1055 and 1105 psis) with average C2D inj ection of 6.68 L/s (~106 spm) . Pressure control be tween 7.38 and 7.72 MPa (1055 and 1105 psis) with no CRD inj e ct ion. d Case 1 head failure caused by vessel overpressurization af ter core slump rather than corium attack on head, as in Ca se 2. Due to overtemperature f ailure of dry-well electrical penetration assemblies. d Tabl e 5.2. Ognperison of accident event timing for the SORV cases with no operator action Case 1 Case 2 Start of f uel melting 80 73 Core slump 206 111 RPV head f ailure 434 253 Containment failure" 472 286 aMinutes after reactor scram. b CRD inj ction increases from 104 spa at 1100 psia to ~180 spa after 60 min. .
#No inj ection -
d In Case 1 the corium debris is quenched when the core slasps; thus the corium must re-
- heatgeforeattackingthebottomhead.
Due to overtemperature f ailure of the drywell electrical penetration assemblies.
ORNL-OWG 83-4738 ETD o R. C D g_ PRESSURE CONTROLLED BY
, AUTOMATIC ACTUATIONS OF m- A SINGLE SRV 8-cn 1W" h b u
~$ STEADY PRESSURE DECLINE y b g_ DUE TO STUCK OPEN SRV e ~ .\
a e 0 ._ a g. \N's\ E y t: A { A g_ H "- uE m M h~ g CURVE A: SORV, NO REACTOR VESSEL INJECTION CURVE 8: SORV.1 CRD PUMP RUNNING CURVE C: VESSEL AT-PRESSURE. NO VESSEL 8-
~
" INJECTION CURVED: VESSEL AT-PRESSURE.1 CRD PUMP RUNNING 8-n E-o- o i . . . . . , .
0 10 20 30 40 50 60 70 00 90 TIME (MIN) Fig. 5.1. No operator action runs reactor vessel pressure.
II l! illi y D 0 T 9 E , , 9 , , 3 N , 7 4 O I , L, T , E, 3 C L , U, 8 E F, E G w J N S S E O Q E y .8 0 O I U, L GV j
- R F, T, L
N ENO SI C E, C A ,_ R S N N, 1 V, l O EN E E T_ F, i U R R q Q_ RR U U A, 9 0PS S S F, p 9 1 MSE E 1
.7 0
CU R R G P, 1 AP P P N g 9 l ED- e R - NI T_, !_ ve OCA RTNTO ANU _ l N.1.ELITC LR E VV S S P _ ._ r RRS E JSM _ _ ,0 e OOEVPN SSVI EU _ D _ 6 t a ABC D _ w EEE VVV E V _ _ r RRR R _ _ e UUU U e s CCC C _ ,0 ) c
._ _ 5 I
N n w M ( o . d _ _ E - _ _ M
._ .0 I n C _
4T o . i _ t _. _ c _ . a
. B .
. ._ r
_ _ o _ _ .0 t 3 a _ r e p A _ _ _ o _ . o _ 0 N
. _ .2
_ 2 _ . 5 _ ._ g _ 0
._ 1 i
\ _.
F
- ~ - - - ~ ~ - ~ - . 0 8k 8e h 8* R S' Q, o8 0 $ 8._ 8 .
)3 .
EC 3J>33LL- uL> S : c o *- e e E 2_WUJ_ 3.J ).L y l l l lll 1lll Iljll l1 l [ . l
e , 6 e *
- m o OANL-0WG S3-4740 ETO ,, e, onst_ cwa g3_4y4, gro 5 0
$~ "d
~- ) glne er k
__________.141____________ s,
/ /j CORE SLUMP AT 726 men y $ =
c o Cp1E M{LyiG
~#
rs jg jI y, 1, CAUSED BY SRV ACTUATIONS 2 g_ TEMPE R ATURE f, ej , - k; G %C 1 ~ I g g MELTING STARTS AT 105 m.n $ f :O "g
=,
M M
=- ==
, g ;l g V'www CORE SLUMP
$ -a, =~ L',m. al .?,,I i
e- e m i 2 31 T] p /g !/ e[ /j i e l _z w w
% % I
- O 0
O o a i M
-____________B_A_F___________.l m o > il Ep ,,
2 9.. E9 H -l :.4 L D% 2 Dk. A ah y
~
~ .a d
x x 'TEMPE R ATURE > ME LT TEMPE R ATURE INDICATE [
< < DtSJOINT MELTING IN THE COPE. THE MARCH MELT w t--
3 2 MODELS ONLY LOOK AT THE FIRST CONTIGUOUS d 0
- O hELT ZONE IN EACH RADIAL ZONE OF THE CORE ,3 'o '
E- g- 2d $5 S f
~
U
% 9- 5 ,,
4_ o " 2 . ve wo e4 e on te24 .m 2 sco w0 ca e 133e ,e . 23; aco TIME - MINUTES TillE - MINUTES
' I'i g . 5 . 3 . Maxista core temperature for the Fig. 5.4. Collapsed water level for the pressurized run with no operator action and an pressurized run with no operator action and a inj ection flow of 6.56 L/s (104 gpm). inj ection flow of 6.56 IJs (104 spm) .
)
CcEE CU1!ULATIVE ENERGY FRolf AlW-I!20 REACTIONS (KWf S8, !8 , ,
" co 12 1
.4 30
- 4. e no e
O ma ea a*- en CUlfULATIVE ENERGY FRO 11 llW-il20 REACTIONS (IrMQ, m #" E? 0a is .: 3 34 as >7 WEE
*w O. p .
o e O na t'* 4 oe*
%4 O M m M Q p e >l n 2.,
-*me a bd O O e5 O *t M o C e w pa. m d 'L
~
em
.o de p~3oDeee s
* ~ * '
- _~ta : "E t
. .-- \
m ., :~ i "*e 23 O ep O. 't
- M A
g. m ., au 8 lg* 9 L 5D Pt h wnoeM r GG N 7 oa m .M ~ g H m %O f
*" O O. W O N sa- 9 t3 *** W **t a o?
n k q_ o L b Q o o 9 COLLAPSED WATER LEVEL (M. AB0VE VESSEL ZERO) gg
, u 1su u.s0 4 m. e .401 80,2e u, s2 e COLLAPSED WATER LEVEL (IN. ABOVE VESSEL ZERO) lg 60 0 124 0 188 0 2s20 316 0 380 0
= =g e zo , -i i
d ** U I
)
awm I kp* 3 .. I
-O w l i I
OO e- i i 4# M
=na d j j g ,
E"E noe
- M S~
- l l
l
" O' I; I; aea w E l- l-
"O
" l 8 l M O N O @dB-de I
m I l j w l . mo ;
- I o D I h E.
-N y I I
a a i a l 7
- m l 5 o D' 81 i O n ow U -
I / 1 I I E
** $ 5 SE e
I i n 39 t 0 I I
- ORNL-DWG 83-4744 ETD a o M' l CORE SLUMP AT 206 mm e a MELT STARTS AT 79 5 mm E 6 dn do 2 - 2 .-
w w a: a: 8e 8o 2 t. M S_ bh DS 2 2 E E 8 < < 2 2 o o k' $~ 9 il : 19 0 S'7 2 9h4 133 6 N1.8 210.0 TIME - MINUTES Fig. 5.7. Maximum core temperature for the SORV run with no operator action and with one CRD pump inj ecting water into the RPV. e
l 40
- 6. POTC1TIAL FOR RECOVERY Willi OPERATOR ACTION 6.1 Introduction The results of Chap. 5 show that without operator action the CRD by- .
draulic system cannot inj ect enough flow to prevent severe fuel damage. In Chap. 4, it is demonstrated that the increased inj ected flow obtainable by operator action can prevent severe fuel damage if initiated soon (with-in 10 min for the at pressure cases and within 5 min for the f ast depres-surization cases) af ter the reactor trip and initiation of the IQUV acci- , dent se que n ce. This chapter describes analyses of sequences in which there are significant delays between initiating events and the operator actions to enhance or initiate flow from the maall-capacity, high pressure inj ection systems. The delays considered in this chapter are on the order of 20 to 30 min between the initiating reactor trip from 100% power and the completion of operator action to start, or to enhance inj ection via the maall-capac-ity, high pressure inj ection systems. The first part of the total delay would be consumed by attempts to start the swveral inj ection systems (HPCI, RCIC, RHR, Core Spray) that would normally be available af ter a scram. The last part of the delay would be consumed by the completion of hand-operated valve manipulations. As discussed in Sect s. 3.2 and 3.3, an
- assistant must be dispatched into the reactor building either-to initiate SLC inj ection of nonborated water or to enhance the CRD hydraulic system inj ection by opening the pump test bypa ss flow path.
Both at pressure and depressurized reactor vessel sequence s are con-sidered in this chapter. Normal reactor vessel pressure control af ter reactor scram would result in the vessel being controlled at pressure, be tween 6.21 and 7.59 MPa (900 and 1100 psia). The MARCH result s for the at pressure sequence s, discussed in Sect. 6.2, establish the longest per-missable delays consistent with no severe fuel damage. For depressurization to occur would require either a stuck open re-liet valve (SORV) or an operator error, i.e., rapid manual depressuriza-tion with no low pressure ECC systems running. The automatic depressur-ization system (ADS) will not actuate unless at least one low pressure ECC system is running. The MARCH results for SORV (Sect. 6.3) and for f ast depressurization (Sect. 6.4) do not establish limiting delays since the peak fuel tempera-tures are well below the severe damage range. They do, however, show that significant delays can be tolerated. As in the previous chapters, the MARCH code was initialized with col-- lapsed water level at or slightly above the top of the active fuel at the time af ter reactor scram and at the reactor vessel pressure predicted by the BWR-LACP code. It ,was not nece ssary to make new BWR-LACP runs e spe- . cially for this chapter. The n&-operator-action runs of Chap. 5 and the f ast depressurization runs of Chap. 4 provide the desired starting condi-tions for MARCH. , l
41 i 6.2 Reactor Vessel Remains at Pressure Significant events are summarized in Tables 6.1 and 6.2 for the two
- at pressure sequences presented in this section. In the first seque n ce ,
i there is no inj ection prior to operator action. This represents the loss
- of all vessel water inj ection initiated by loss of offsite power. In the j second sequence, there is an injected flow of 6.55 L/s (104 gpm) prior to
- operator action, as there would be for any react or scram not involving i loss of offsite power.
For both sequence s, there is no operator action (or, more precisely, !' operator action does not become effective) until the time of initial un-covering of fuel at the top of the core (28 min without any inj ection and 36.5 min with the 6.55 L/s (104 gpm) postscram inj ection) . At this time, the operator establishes a flow of 12.6 L/s (200 gpm) which is maintained constant at this value for the remainder of the sequence. Operator con-
- trol of the SRVs also begins at this time. The Browns Ferry emergency operating instruction
- for main steam isolation valve closure requires remote-manual operator control of the SRVs between 1100 and 900 psia (7.59 and 6.21 MPa) in order to avoid automatic SEV actuation.
The BWR-LACP results for reactor vessel downcomer water level and reactor vessel pressure for the pre-core-uncovery parts of these sequences are discussed in Sect. 5.2 and plotted on Figs. 5.1 and 5.2. The MARCH
. run s, de scribed below, were initialized with collapsed water level at the
- top of the active fuel at the times predicted by BWR-LACP
- 28 min for the
- sequence with no inj ection before operator action and 36.5 min for the
, sequence with 6.55 L/s (104 spm) inj ection.
The period of core uncovery for both these sequences is dominated by j the SRV actuations. The reactor pressure for the first seque nce (initi-l ated by LOSP) is given in Fig. 6.1 [the MARCH run for this case used a i manual actuation pressure for the SRV of 7.72 MPa (1120 psia) instead of ) 7.59 MPa. This difference insignificant 1y affects the computed results.]
- Each pressure cycle consists of a long repressurization period followed by
] a rapid 1.38 MPa (200 psi) depressurization when the operator actuates an 4 SRV. The pressure spikes in Fig. 6.1 actually occur af ter the SRV opens. 1 This is not realistic but is caused by the inadequacy of the MARCH flash- ! ing model. The uncertainties due to this model will be discussed in Chap. i 8. The sudden 1.38 MPa depressurization af ter SRV actuation results in core water flashing and a resulting drop in inventory and collapsed water level as shown in Fig. 6.2. The SRV actuations cause flashing and in-l creased steaming rates which in turn swell the incore swollen mixture water level dramatically (Fig. 6.3). As shown in Fig. 6.3, each level
- sw el l, caused by at SRV actuation, goes above the top of the active fuel, thus quickly cooling the fuel rods (Fig. 6.4) which have been heating up rapidly during the repressurization cycle. Even though nearly 75% of the core is uncovered (Fig. 6.2) five times be tween 150 and 255 min, the fuel is cooled dramatically with each SRV " pop."
The repressurization period before each of the (5) maximum uncovery
; dips yields the highest core temperatures (with a maximum fuel temperature
*The setpoint for automatic actuation is 7.72 MPa. Per instruction, i the operator should manually actuate at 7.59 MPa.
I l I
I 42 of 2766*F at 253 min) and the highest metal water reaction bursts from the f uel cladding (Fig. 6.5), the canister (Fig. 6.6) and the control blades (Fig. 6.7). There are significant burst s of metal / water reaction, but
- the final figures for total retal reacted for the cladding, canisters, and control blades are 0.8 82, 0.460, and 0.347%, re spe ctively. After the j minimum core level at ~200 min, the core inj ection of 12.6 L/s (200 gpm) turns the level decrease around (Fig. 6.2) and the core completely recov-ers a t ~541 min, when the collapsed water level is greater than the top of i the active fuel.
l Sequence No. 2 of the at pressure cases has an inj ected flow of 6.55 L/s (104 gpm) prier to operator action (the significant events for this case are given in Table 6.2). This sequence is basically less severe than the sequences initiated by LOSP. The events for this sequence are again domina ted by the SRV act ua tions, as evident from the response of the key core variables: collapsed water level (Fig. 6.8), swollen water level (Fig. 6.9), and fuel temperature (Fig. 6.10). As in sequence 1, each SRV actua tion swells the water level dramatically above the top of the active fuel thus basically quenching the fuel. The inj ection prior to operator action in this sequence allows the core to recover quicker (at 508 min) and the core is considerably cooler [by ~333 K (600*F)] than in case 1, thus there is less metal / water reaction. The final figures for the per-centage of metal reacted for this case are 0.235 for the cladding, 0.082 for the canisters, and 0.006 for the control blade s.
! Fren Chap. 4, the parametric at pressure studies show that negligible .
core damage wculd be expected for either automatic SRV actuation or opera-ter manual operation of an SRV if the total inj ected flow were greater t h an 12.6 L/ s (200 gpm). In these ' parametric studies, there is an initial , fl ew of 7.06 L/ s (112 gpm) followed by an increased inj ection flow occur-ing a t 10 min af ter reactor scram. Realistically, operator enhancement of the inj ect ion flew might cccur later than 10 min, especially if the se-quence is initiated by LOSP. What has been portrayed in this section is realistic e stimates of operator actions and timing, but, after operator enhancement of the inj ection, an inj ection flew of 12.6 L/s (200 gpm) is
, assumed rather than the flows given in Table 3.1 or Fig. 4.5. Even with the delayed timing and the assumed flew of 12.6 L/s, in both at pressure sequences the core is predicted to recover with some damage (in case 1, <
l 0.9% of the cladding react s) and with the core geometry intact. It should j be pointed out (referring to Fig. 4.5) that the capability at BFNP unit 1 l exists for combinations of system operation and valve alignment resulting in enhanced inj ection flew rates greater than 12.6 L/s (200 gpm). One of these [15.71 L/s (249 spm)] is applicable to the sequence initiated by LOSP because it requires only one CRD pump to be running (see Sect. 2.2). 6.3 Stuck Onen Relief Valve Significant events are summarized on Tables 6.3 and 6.4 for the two SORV sequence s presented in this section. These sequences are analogous
- to the two at pressure case s of Sect. 6.2 in that the first sequence has no inj ection prior to operator action at the time of core uncovery and the
! second sequence has inj ection f rom the 1A CRD pump that is normally run-i ning af t er react or scr asi. l _ _ _ .-. -- - .- . - . ,- , ~ -
43 For these sequence s, the CRD hydraulic system inj ected flow could not i be treated as a constant. The large pressure decrease that take s place during the early part of the sequences causes a si gnif ica nt increase in inj ected flow. For both the BWR-LACP calculations and the MARCH calcula-tions, programming was employed that predicts the inj ected flow as a func-tion of both reactor vessel pressure and total flow resistance (se e Appen-dix A for description). The increased CRD flows at low react or vessel pressure cause in-i creased pressure drop between the condensate storage tank and the suction of the CRD pumps. Sensed suction pressure below 60.7 kPa (-5.9 psi g) initiates an automatic " low suction pressure trip" of the operating pump or pumps. As shown on Fig. 3.2 and discussed in Sect. 3.2, low suction pressure trip is a distinct possibility for reactor vessel pressures below about 4.4 MPa (625 psig), and is much more likely if only one CRD pump is running. The low pressure sequences in this chapter are based on the assump-tion that the operators would recover from the pump trip by bypassing the pump suction strainers, thereby decreasing losses and increasing suction pr e s s ur e. Since there is a specific annunciator panel in the main control room for CRD pump low suction pressure, the operators would know of the condition, and to bypass the suction strainer would be an obvious way to increase the suction pre ssure.
- This action would take at least 5 min to complete since the requisite hand-operated valve manipulations must be com-
+ picted in the reactor building.
! The BWR-LACP result s for reactor vessel downcomer water level and reacter vessel pressure for the pre-core-uncovery periods of these se-
. quences are discussed in Sect. 5.3, and plotted on Figs. 5.1 and 5.2. The MARCH ru .s, de scribed bel ow, were initialized with collapsed water level at the top of the active fuel at the time predicted by BWR-LACP: 17 min for the sequence with no injection before operator action and 19 min for the sequence with inj ection f rom one running CRD pump.
In the 50RV sequence in which there is no injection for the first 17 min (possibly initiated by LOSP), the operator starts a CRD pump at 17 min af ter having opened the pump test bypass line, and the operator also initiates inj ection flow to the RPV via the SLC system at 30 min. The com-bined flow from the CRD and SLC systems replaces the water lost from the vessel during the rapid vessel depressurization. The collapsed water level (Fig. 6.11) in the vessel reaches a minimum of 6.73 m (265 in.) above vessel zero (~68% of core uncovered) at 44 min. Then with increased inj ection flow due to decreasing vessel pressure, the water level starts to increase with full core recovery occurring at ~226 min (i.e., collapsed water level greater than top of the active fuel). However the core is effectively recovered much earlier at 118 min by the swollen liquid level. Since an SRV is continually open during this sequence, there is a con-
, tinuous high steaming rate within the core, and with the decreasing vessel pr e s s ur e (and decreasing steam density) the water level within the canis-ters swell s dramatically. For the period from ~40 min to ~120 min, the swollen level is 1.2-1.8 m (4-6 f t ) above the collapsed water level (as seen in the core barrel and in the interstitial region). The swollen level thus also dictates the period of temperature excursion within the
- Conversations with operators indicate -that they are aware of this.
_ = _ . l 44 core as seen in Fig. 6.12. The maximum temperature attained in the excur-sion is ~950 K (1250*F) during the period be tween 67 and 96 min when the upper part of the core reaches essentially steady-state conditions as the . water level is slowly increasing. The temperature excursion period f rom 28-118 min define s the period of incore metal / water reaction (Fig. 6.13). Note, with the continuous . steaming caused by the SORV, the reaction is continuous in contrast to the high energy burst s observed in the at pressure cases (Figs. 6.5-6.7), but the peak energy shown in Fig. 6.13 is a f actor of 2200 times less than that of the bursts in Fig. 6.5. This sequence result s in a fully recov-ered core af ter 226 min with negligible core damage [<0.007% of the clad-ding react s and ( 9(10)-8% of the canister react s]. 1 In the second SORV sequence there is CRD injection during the first 19 min, but when the operator attempt s to enhance the flow by opening the j pump test bypass line and opening the throttling valvo (No. 85-527), the
- resulting high flow causes a low suction pressure trip of the pump. An 11-min lapse in injection is assumed before the operator restarts the pump at 30 min; he also starts the SLC at 40 min.* The system behaves similarly to the first SORY sequence; the water level (Fig. 6.14) reaches a minimum
- of 6.4 m (253 in.) above vessel zero at 40-45 min. The core recovers com-pletely at 203 min with the fuel rods being quenched by the swollen liquid level at 108 min. The core is uncovered slightly quicker (~2 min) and -
- more (0.335 m) than in the first case, and the core excursion temperatures ,
(Fig. 6.15) are greater at 1154 K (1617aF): as a result, the extent of l the metal / water reaction (Fig. 6.16) is greater. Even so, only 0.18% of the clad and 0.029% of the canister react s - minor damage with cos plete core recovery. 6.4 Ranid Depressurization Significant events for the f ast depressurization sequence are summar-I ized on Table 6.5. The operator decision to depressurize would, in this case, be a mistake because the low pressure ECC systems are, by defini-tion, not operational. The depressurization begins af ter 5 min with the opening of 3 to 6 SRVst and proceeds rapidly, decreasing pressure to below the shut ef f pressure for the ECC system pumpst within less than 5 min.
; When this pressure is reached, the operator halts the rapid depressuriza-i tion by closing SRVs and, thereaf ter, maintains an average pressure of ^
about 1.55 MPa (210 psig) by intermittent actuation of a single SRV. The rapid depressurization result s in an initial uncovering of active fuel within about 8.5 min of event initiation. Without any operator ac-tion, the singic running CRD pump is able to inject a flow of about 10.4 L/s (165 spm) into the partially depressurized reactor vessel (Fig. 6.17). Af t er 20 min, the operator attempt s to increase the inj ected CRD flow by . opening the pump test bypa ss flow path.
'The assumed timing of these events was ' developed af t er discussion with instructor personnel at the Browns Ferry simulator.
t0perator initiation of ADS would result in the opening of 6 SRVs. ii.e., less than about 2 MPa (275 psig).
45 l In this sequence, just as in the SORV sequence s, the low suction l pressure trip of the CRD pump is a potential problem. When the pump test bypass flow path is opened, the resulting high flow rate intmediately causes a l ow suction pressure trip. For the next 5 min, there is no in-Joction into the reactor vessel. Af ter the suction st rainer is bypa ssed, and the pump restarted, flow inj ected into the react or vessel increases to about 17.3 L/ s (275 spm). At 31 min, the operator initiates the SLC thus f urther increasing the inj ection t o 20.9 L/s (331 gpm). The BWR-LACP result s for react or vessel pressure and downcomer water level are essentially the ser.e as those discussed in Sect. 4.3.1, and plot ted on Figs. 4.6 and 4.7. The 11ARCll run, di scussed bel ow, was ini-tialized af ter an elapsed time of about 8.5 min, with collapsed core water level near the top of the active fuel, and with reactor vessel pressure bel ow 2 MPa. The rapid depressurization sequence is dominated primarily by two items: (1) the initial depressurization and resulting rapid loss of ves-sel water inventory, and (2) as in the at pressure cases in Sect. 6.2, the SRV act ua tions. There is an initial period of rapid water level (Fig. 6.18) decrease, caused first by the depressurization and then by the trip of the CRD pump, which is finally termina ted a t 25.5 min when the (RD pump (with the bypa ss line open) is brought back on line. Af t er an SRV act ua-tion which drops the level to it s lowest point [6.88 m (271 in.) above vessel zero] a t 30 min, the enhanced CRD inj ection plus the SLC inj ection e completely recovers the core at 155 min (1.c., collapsed level greater than top of active fuel) . Except for a period from 21-55 min, the swol l e n water level (Fig. 6.19 ) , wit h dr am a t i c surge s caused by SRV act ua tions,
, tends to quench the uncovered part of the core, and after 120 mir. the mat-imum core temperature (Fig. 6.20) is less than 564 K (5558F). During t he period f rom 21-55 min, the core does rapidly heat up, reaching a maximum of 1015 K (1368'F) at 41 min, llow ev e r, af t er 44 min, the SRV actuations and increased inj ection flows quickly cool the core with essentially neg-ligible damage ((0.0012% of the cladding react s).
6.5 Conclusions The result s presented in this chapt er demonst rate that there is c o n-siderable potential for recovery with operator action. For the more probable at pressure ca se s not involvir.g rapid depressurization or SCRV, severe fuel damage can be prevented if the operators can obtair a total inj ection flow of 12.6 L/s (200 gpm) before the time of initial uncovering of active fuel. The required 12.6 L/s inj ection flow can be obtained two ways: (1) the CRD hydraulic system pumping with the two CRD pumps in par-allel, or (2) the CRD hydraulic system pumping with only one pump supple-
, mented by an additional 3.5 L/s (56 gpm) of inj ection f rom the SLC sy st em.
The predicted flow rate (Table 3.1 and Fig. 4.5) for either of these con-binations exceeds 12.6 L/s (200 gpm), if the 85-527 throttling valve is
, f ully opened. Thus, even for the event of a loss of of f-si te power,
- in
*Which would trip the operating 1A CRD pump, and requi re operator action to start the ill pump.
46 which a delay of 28 min could be tolerated, there is adequate time for completion of the decision process and the requisite local hand-operat ed valve manipulations. , For the seq ue nce s involving either rapid depressurization or SORV, there is also adequate time for operator action. Even though the loss of vessel water inventory by flashing begins to uncover the core sooner, the , CRD hy draulic sy stem flow is greater wnen inj ecting to a depressurized or partially depressurized reactor vessel. There is an additional demand on the operators when the vessel is depressurized in that they must be able to recover from, or prevent, the low suction pressure trip of the CRD p ump s. This point is discussed further in the next chapter and in Chap. 8, Uncertainties. 4 1
47 Table 6.1 At pressure sequence resulting in recovery without severe fuel damage (case of LOSP) Time Event ' (min) 0 Reactor scram from 100% power 0 Loss of all vessel water inj ection, including the CRD hydraulic system 0-28 No water inj ected into reactor vessel 0-28 Reactor vessel pressure controlled to average value of 7.55 MPa (1080 psig) by automatic SRV actuations 28 Vessel downcomer water level has decreased to the level of the top of active fuel (9.14 m above vessel zero). 28 Operator act s to start the IB CRDHS pump and establish an inj ected flow of 12.6 L/s (200 spm) 28 Operator begins reactor pressure control be tween 6.21 and 7.5 9 MPa (885 and 1085 psis) by runote-manual SRV actuations 98 Significant metal-water reaction begins 201 Collapsed water level in core reaches minimum (~75% of core un cover ed) 252 Hottest region of core reaches maximum temperature of 1792 K (2766*F) 392 Significant metal-water reaction stons (total fraction of l clad reacted (1%). l 541 Core completely recovered. Collapsed water level above the l
. top of the active fuel
..a... .. .
_ _ _ . _____________________.._______________._________.._.______.s
48 Table 6.2 At pressure sequence resulting in recovery without severe fuel damage (without LOSP) I"' Event (min) 0 Reactor scram from 100% power 0 Loss of all vessel water inj ection except the CRD hydraulic system (1 pump running) 0-36.5 Without operator action to enhance flow, the CRD hydraulic system inj ects an average flow of 6.55 L/s (104 spm) into the reactor vessel 0-36.5 Reactor vessel pressure controlled to an average value of 7.55 MPa (1080 psig) by automatic SRV actuations 36.5 Reactor vessel downcomer water level has decreased to the level of the top of the active fuel 36.5 Operator act s to increase inj ected flow from 6.55 to 12.6 L/s (fram 104 spm to 200 spm) 36.5 Operator begins to control reactor vessel pressure between 6.21 and 7.59 MPa (885 and 1085 psis) by remote-manual SRV actuations 102 Significant ' metal-water reaction begins 229 Hottest region of core reaches maximum of 1462 K (2173*F) 232 Collapsed water level in core reaches minimum (~72% of core uncovered) 365 Significant metal-water reaction stops (total fraction of reacted clad < 0.25%) 508 Core completely recovered (collapsed water level above the , top of the active fuel)
I 49 i i Table 6.3 SORV sequence resulting in recovery
, without severe fuel damage (LOSP ca se) e (min) Event 0 Reactor scram from 100% power 0 Loss of all vessel water inj ection, including CRD hydraulic system 0-17 No water inj ected into reactor vessel. Reactor vessel pressure decreasing steadily due to the stuck open SRV 17 Reactor vessel downcomer water level has decreased to the level of the top of the active fuel 17 Operator starts the IB CRD pump and opens the pump test by-pass, and partially opens the 85-527 throttling valve. Re-j .
sulting inj ected flow is ~14.5 L/s (230 gpm) 30 Operator initiates 3.53 L/s (56 gpm) inj ected flow to reactor
, vessel by starting the SLC system 17-35 Vessel pressure continues to decrease, causing CRD pump flow to increase. Operator must act to bypass the suction strainer before 35 min to avoid low suction pressure trip of the pump 35-end Increased throttling using valve 85-527 may be necessary to limit CRD pump flow and prevent resulting low suction pressure trip of the pump 44 Collapsed water level in core reaches minimum (68% of core uncovered) 67-96 Hottest region of core reaches maximum temperature of 950 K (1250*F) (negligible metal-water reaction at this t empera ture) 118 Fuel rods quench due to swollen steam / water level greater than top of the active fuel 226 Core completely covered (collapsed water level above the top -
of the active fuel)
)
50 Table 6.4 SORV sequence resulting in recovery without severe fuel damage (case without LOSP) , Event I[) , O Reactor scram from full power 0 Loss of all vessel water inj ection except CRD hydraulic system (1 pump running) 0-19 Without operator action the CRD system inj ect s a flow, ini-tially 6.55 L/s (104 spm), that increases as the vessel is depressurized 19 Reactor vessel downcomer water level has decreased to the level of the top of the active fuel 19 Operator opens the pump test bypass path and fully opens i throttling valve 85-527. The resulting high flow causes a l l ow suction pressure trip of the CRD pump 19-30 No flow inj ected to reactor vessel 30 Operator bypasses the CRD pump suction strainer, restarts the . tripped CRD pump. The initial inj ected flow is about 16.1 L/s (255 gpm) 3 0-e nd Increased throttling using the 85-527 valve may be necessary to limit the increasing CRD pump flow and prevent another low suction presure trip of the pump 40 Operator initiates an addi::ional 3.53 L/s (56 spm) inj ect ed fl ow to reactor vessel by starting the SLC system 40-45 Collapsed water level in core reaches minimum, 76% of core uncovered 40 Metal-water reaction begins 92 Hottest region of core reaches maximum tanperature of 1154 I (1617'F) 108 Metal-water reaction stops (total fraction of clad reacted (0.2%) 108 Fuel rods quenched by swellen water level 203 Core completely recovered (collapsed water level above the top of the active fuel)
)
J 51 Table 6.5 Fast-depressurization sequence leading to
- recovery without severe fuel damage e
Event (min) O Reactor scram from 100% power 0 Loss of all vessel water inj ection except for the CRD hydraulic system 0-5 Reactor vessel pressure controlled to an average value of - 7.55 MPa (1080 paig) by autumatic SEV actuations 0-5 W1'hout operator action, the CRD hydraulic system inj ect s an average flow of 6.55 L/s (104 spm) into the reactor vessel 5 Operator begins an emergency depressurization by opening three or more SEVs 5-8,5 Reactor vessel pressure rapidly decreasing. CRD hydraulic system injection increasing from the initial value to 10.1 L/s (160 spm) at 2.07 MPa (300 psia) 8.5 Downcomer water level has decreased to the level of the top
- of the active fuel 8.5 Reactor vessel pressure below 2.17 MPa (200 peig). Operator closes all open SEVs except 1 and thereaf ter controls
! pressure to an average value of ~1.55 MPs (210 pois) by remote manual SRV actuatiods 8.5-20.5 Without operator action, the CRD hydraulic system continues to inject about 10.4 L/s (165 spa) into the reactor vessel 1
20.5 Operator attempt s to enhance MD hydraulic system flow by opening the pump test bypass, but this action increases flow excessively and causes a low suction pressure trip of the CED Pump 20.5-25.5 No flow injected into reactor vessel 25.5 CRD pump suction strainer bypassed, one CRD pump restarted, and valve 85-527 partially opened with resniting flow of about 17.3 L/s (275 spm) 30 Collapsed water level in fuel reaches minimum (62% of core l uncovered)
, 31 SLC system injection initiated, adding 3.53 L/s (56 spm) to 4 inj ection flow 41 Hottest region of core reaches manimum temperature of 1015 K (1368'F) (negligible metal-water reaction at this t empe rat ur e) 155 Core completely recovered (collapsed water level above the top of the active fuel)
I
~ _ _ _ - ,_ . _ _ ,_.
3 s l l l ,
\
) 'g *1Cf PRIMARY SYSTEM PRESSURE (PA)-
. 627 70.9 79 0 872 95.3 1034 g5 PRIMARY SYSTEM PRESSURE (PSIA)
, .,oe 910 0 1028.0 11460 1264 0 13820 1500 0 s
g _- __ .~
,s a -
Ew b % I a: w i
@A oo 3 A_
l N $- % l 0% 1* h._ i E g A_ S. E N-. I
"" g $~ % i g" % __ l o V
. W I !
4Y $~ h i 30 6 gQ " 8 ** R o N .I l
]b
~ {x
.$ *n* m I -
O s - O e .
~
*y
,, , COLLAPSED WATER LEVEL (M. ABOVE VESSEL ZERO)
*o 1,3500 7 0612 7.7724 84936 91948 9 9060
.n e o.
- a. d y COLLAPSED WATER LEVEL (IN. AE0VE VESSEL ZERO)
\ M *- 250 0 ,_ s 2760 3060 334 0 362 0 390 0 E :*> $ s*~ l 4 " .* m
_C-l C --
- F
- I
.-, I
.R*" - -
I
*E N E- N l Re n= -
2 8. n e 2 1 lE l
- 7 ~
l 27 s$ 5 x _ ;. g" ~
- c. N i e F.= S si l i 0
~
2 56 N_ ; E- 9 e8 - 7
-m
= g 8
g2 o l I
]e OU i
5 h 7% CORE RE COVE RY AT 641 mm '% b n a a ZS
^%
r yg SWOLLEN MIX. LEVEL (M. ABOVE VESSEL ZERO) g ,6350 7.7,22 9 093
, 10,46s 114, 36 11,208 m = "is
*" ,C SWOLLEN MIX. LEVEL (IN, ABOVE VESSEL ZERO)
. 250.0 304.0 ass 0 41z0 4se.0 s20 0
, u = -
nm a a- ~ 2 ,. I g N ! ) .- E N i l :g ' N_ ! eo g '% i
.E " ;;; N- i R. R E- N l E$ l* A l l *: = N !
gg g N- ! - x ' o d $ -- A 1 R
'a h" N !
" o N i
~
=
,. 7 E-
=
N _i ! l x [, w, g a >s
- N ,.
! E.E h i n $ o a j o 1 6 8Y MAXIMUM CORE TEMP. (K) qg 5ss.4 eri.6 1079e 1342 0 1s04 3 laces
.5 MAXIMUM CORE TEMP. (F)
"y s40.0 1012.0 1484.0 1956.0 2428.0 2900 0 s
- 3. >,m ! *- '
ce
- o.
. ==" r
~
N_ i aE e I 'S E E
- ,, m !-
ag i* N_
" g _
E
$P8". c % DE ax
~a W 75 s
'i -
EE 3
~m s b
-n a- .
O ",
- N ma ~ _ T $
-* ~ ,
M n G O
" c g
Os ~ M m & h 4 o m o es s
ll 1I'(Il l u^ 0 D 0 T 6 E 5 n 0 5 7 o i 4 t e 3 , m cc c 6 an w O d 3 ee L
- 4 ruq N
R re O es . S t E a hU 27 T w "e
/ r 3
l N l u I as J M t s ee j 6 - mrpP . /4 0 i 4E r S 2M etO I t aL T s" l i y nVb 4 aU J 4 Cl id i 3 1 T e t
.ea
- 6. tht i 0 6 i 6
2 rn
.oi y : g gf os !- e o~ i )
-s hn0 >o%M$ zo 5 <W
. oy5 xom F y1
- d. j g
r . mn n = g eo no nN
- g. M >=E?0z8U0MoU E xoA e( .
D 0 T 0 E 6 5 n 9 74 o i 4
- t 3
8 ce G 6 ac D w 3 en
,5 re L
N 4 u R rq O ne ee m t s
. S a 2
E 9 3 i 2 T w"
,74U /e 3 l r N au I
t s J M es J mer . 8 l 0 d pP J 4E a S 2M l t O l T cal di l y I eVb 4 4 uU 3 FOd ne 1 Te t m 1 8 . ea . 9 0 5.h t ti 8 6 i 2 rn L . oi e= E oe % : 8 gf . br E=x gm seMgM 8 U@m om?5 Cju i F y1
)
g
- - - - - . r .
=- b e:: ? : 3 e o
_," $5ZM $E<E 8%i5 !ao ! nN br e( i llll ft t
. _ ._ =_ . . _ . - __ __
i g o g g*W 9091V AW3A033u luo3 $ b 3 f,
~ *l X- I X.
w_ n
< g a
w I g l 8 $
- I X S g I
'w ! g l :5 % :,, .
si N Ei ;a l .I ed i
-Q A ta 3*
l ~N. pg 1 ~~ P a I N o3 1
'm ] *?*
X so : I %
) l -'
ooot och cotc coic coes cose OE (083213SS3A 3A08V 'NI) 13A31 H31YM G3SdVT103
- 5, 091'ot acce scie wit etit osed
, (083Z 13SS3A 3A00V 'W) 13A31 H31VM 03SdVTIO3 J B o I e
; a ::
I *i. l . :: g
-3 5:
a=
)I 53 oabb e 8
2 4 h eb
' n es C *
%.D
-5 ad...
q i o9 99 l *ha 9t EC 9k o'0 pI41/010) ADH3N3 NOI13V3M 0211-RN 30Y16 'I0HINO3 EEO a . DEZ 99T tti t'6 49 - 00 0j pt.(M) ADHan3 N0113Y38 0211-an 30Vla 10HINO3 *-
onht -onces-47s3 Ero ca%L-N a3-47s4 Evo E- $ E- h
- ^
max TEMPERATURE
- 217FF N e N , ,
AT 229 nwa d h- dO h-
~
I
~
@ d 2
c as w l si si Si-s Ni l
/
/
w w l TAF $ h J ' ' 1
!!- bi-- I' I /
.-a si a ;
2 2 i s . W g l 5 E s s, I 2 s, e
)
E- 3 s,E- sR~ jg / o d : i
- f. d 1
q
/ l f ((>,-
R
=_ $ , . . .
$ 1 530 0 76 0 1348 2336 3324 4312 530 0 36 0 134 8 2336 332 4 4312 TIME - MINUTES TIME - MINUTES Fig. 6.9. Swollen incore water level for Fi g. 6.10. Core fuel temperature for MUV MUV "at pres sure" sequence No. 2. "at pressure" sequence No. 2.
l l l l l
- a. COLLAPSED WATER LEVEL (M. ABOVE VESSEL ZERO)
! ;;; ,88040 22:3. va232 e432s 80m ess20
- C we am COLLAPSED WATER LEVEL (IN. ABOVE VESSEL ZERO)
$ ,",' 260 0 284 0 3080 33a0 356 0 380 0 ,
,. g .__ -3 i
l D' M I m .- , O 8- l EnO
- l
^W E I 5
'e d }
O
't3 g *" ,.
m G " w o. I : l O l b h R$ l
-- z_ s 80 i oa C"- > <F I
.m_. 8 " He .99 l
%- @ ~ s i oa n<a l ;t.
c l 2 a i e
%* 5- 5; I 7 E%
. o
- 85e I i l
" M l h I
qh CORE RECOVERY AT 226 min m h'
$ =
0 o e i
*$ 3ees MAXIMUM CORE TEMP. (K) e33: ws
,g 4,9 e e998 2033:
p . m MAXIMUM CORE TEMP. (F)
,, ,200.0 440 0 680 0 .20 0 1260 0 24000 i
i M & 2-DJ gi Oo
$%'N 4 . JI r-E~
I,. , = Eh r ,. -e~ m $m mu
^ 5 ("
d Dr 3 Q{E u-M 2-
-s
- n 5 O I U OO *
*t 5 ~
-o -
Rg 5 Z D km. ,, C m. O ,o. 2 u . on Ad 9 Y * '9
, e .
, ?
?* 1" $5 i
-p ; , t L
7$ e 9
- 226(COMPLETE CORE RECOVERVI h n a I
LS
fe uh $8bo %dEo,. dHm$ on . ya re _o o k l" g H2-r - - - - fg iy F ujO r i 7eo g<5 $ g 6e> ~mg b rd
. s fg t o.
oo E - e* t s. , Eo T r 1 0 1 6 7 I' . Q1 4 mU3 4
. i V.
n 5,
. s 9 6
)
.ql eC ua ed T I
~
nd ' ci M1 eng E 0, 2 2 - w - u im ., s t e M ht I a N1 Sl U 4, 4 O/ T 8 Rw E Ya S t ( e O nr A o 1 F, N L r 1 D ie 4 W na jc G 8 3 et - ci 4 7 t o 5 2 7 in 3 8 i E 0 o 0 D T n
. fs ie rq uoj<2O yHd.$d a <8>w >g .
o s u yl ,g -f @
t e
_= h - - - -
nF 1ci 9eg odugm@ gp@ .$d Oz w >B .
%om
. l h M Ro j
[ Ro mw 1 9 f1; i i.6 0
.n h1 )
t _
. S .4 4 _
O 5 _ k 7 VCo 2 _ ( l M . 1l I N _ a T I L E _ Cp V M9 _ p Rs De d E5 4 E L O F
// i fo uw 2 aS s
_T ma M 5 3 ' EW co F p t e I N 1 A' oL vL _A T i r U33 4 EE nN _ n T6 0 - s W jl E 4 F _ ee S 5 uA E T
. cv m e t E
R O t e n . _ A il aL _ N n oE L g f m oV _ O 8 sE o L w G
. fr 8 o 3 r I' 4 D 8= 5y =l08 I /
7 5 8 U 2 1 V 0 E T 0 O
MAX! MUM CORE TEMP. (K) 366.5 5331 699 8 8665
,h, o, 10331 1199 8 MAXIMUM CORE TEMP. (F) og e3 as se se 200 0 500 0 800 0 1100 0 1400 0 !700 0
.% C.s H. .. - 3cE (A *
- 5. 4
- oe cp #m e*,.e e f EZr-
- Wf mw e CE E
u E tn B -- E . m I O M &C he E "s N
- EiE Sgi~l g I $ U 2-
=3 is !! 9
*p 3 Em $ f Es
.N
- i 85 gr E
= $
oe " g s
**.H.
p C y s se N o m O 8 pa 00 CLAD ZR-II20 REACTION ENERGY (W) 47 94
*16 .
t e 14 1
- 18. 8 23 4 e,
} o4+
$ CLAD ZR-il20 REACTION ENERGY (BTU /11R) *1d g e.ei s 00 16 32 48 64 80 m o .e .
' i m o
,O He 1 CM EP N$
g ,, g. _
.ne-* -
r+ .o C N NO O. d f
*RS m . "*E . , 1 [-
K
.w -
# Z.
M" C os s .8 kN~ $
- g .
e.o IC
*6 W L B; 3 s
'O O o y o o 65
60
** SSL AV AW 3 AO33W 3WO3 h $W c
- 2 w .s3 m' ow l
- N y i j i 3 l l
I D N f
'8 3a e l / L e .o.e wa i
i D N a j S '3 w H
@Q **g" .
- a 3g N, E .
4
~' > ?,
- mW s 8 .w
- 5
- b l
o-I a v-
.e ,
W 0 00C 0 99C 090C O ttC 0 262 0 042 e4 e (OH3Z 13SS3A 3A08V 'NI) 13A31 H31VM G3SdVTIo] EE 029'9 6 206'O 6 tFCS'S 992b4 99th4 0999 9' O' (ouaz 13sS3A 3Aosv n) 13A31 H31va a3SavTIoa : - 3 2 8 8 8 a I - o 8 3 ( - a'
= 1 g .
-I, =
g c
! ! ~
$p s 3
- ; i a-gh h N h d b #
oo > 2 a to.e y5 o oo 5< <8 p g -
= E! 3 2 .a S
9
$2 due @
2 E 0, s'
-m W m u E*
G e o s *y aa E f $ Y 5
-r h
9 . n a gb. i a y . y .
.4 il ? m 0000C
~
000t2 . 0 0001 00d21 00b9 00 dI e (N/81) 13SS3A 3HnSS3Hd 01NI 31VH Mold 333 E* i i i i i . .a - 909 t ** 09922 tt!91 909CI 2/46 000 0 (s/ox) 13ss3A 3HnSs3Ha 01NI s1VH mold 003 ,
e e o e e e oest-Dac a3-4m Eva oes a g3 a e4 gr3 N -. k
=3 h_
3 e C W Au TEMPER ATLdE OF 1366 F AT af me 2 W l q, o c' c dEj E3~ 8 E' fj J i3 ' @ cr c
~ j4 b
s y
- l, it M c O
>o 4
> c. > 2 ev i2%
o ga Q :- x ;;" Wb' H H~ ($.'/ - j Q-1 q' l [ x 3 x ,
~:s j ,
'( l9 5 u
5 u 4 ! f dg I gg aj de _l l. i 2E , ga 2 g 9 w T*F ~ -
# g- e l l/ l li
~
. ji R E x~~ x
~
i l u 1 2 l I i < ! 1 1 ! l j Il l l I.
'l t'
$N :rw"$- Il l
!j. { h, f ! f!
^=i
('
/ < F; 3 l -i o ) 5
, I l
/
/ ',
j'l t l;
*1l l 1 1 i ;l: I l g be j
'i .
'I ! l fg. e ,
.. E aso assa este 9944 129 72 :sooo am m es:s n44 an ise m TIME - MINUTES TIME - MINCTES Fig. 6.19. Swollen water level for TQUV Fig. 6.20. Maxiet:m core temperature for sequence with rapid depressurization of the RPV. EUV sequence with rapid depressurization of the RPV.
62
- 7. ACCIDENT SEQUENCE WI'Ill OPTIICM OPERATOR ACTION 7 .1 Discussion ,
The only sure way for an operator to prevent severe f uel damage dar-ing a 10UV accident is to inject enough water into the reactor vessel sooa . enough to prevent the uncovering of the core. As shown in Chap. 4, initi-ation of SLC system and enhancement of CRD hydraulic system (CRDHS) inj ec-tion within 10 min af ter the scram is sufficient to keep the core cov-cred. Delayed initiation of the CRDHS and SLC system flows af ter a TQUV accident will cause at least some of the core to become uncovered, and increase the chance of severe fuel damage. TQUV sequence s were presented in Chap. 6 in which the operator actions were significantly delayed. As a result of the delays, there was in each sequenca a period of partial core uncovery with only steam cooling for the uncovered fuel. Elevated temper-atures were reached in the uncovered fuel, but fuel damage was generally slight. In order to understand why any core uncovery leads to a higher chance of severe fuel damage, it is necessary to consider how the oper-ators might interpret plant operating procedures in an actual emergency. Figure 7.1 is an event tree diagram of possible operator actions dur-ing a TQUV sequence. The consequences range from no core uncovery to com-plete core uncovery and resultant melting. Fast depressurization is a , very undesirable operator action because it causes rapid core uncovery and, in the TQUV sequence s, there are no low pressure systems (i.e. , low pressure ECCS or motor-driven condensate pumps) available to refill the
- vessel af ter pressure is reduced. In addition, the reliability of the CRDilS is decreased at low pressure because of the likelihood of " low suc-tion pr <ctre" CRD pump trip and the chance that the operators might not be able to prevent or recover f rom this trip.
The case of f ast depressurization has to be included in Fig. 7.1 be-cause the plant emergency procedures make it a distinct possibility. Browns Ferry E01-46, " Loss of reedwater in Conj unction with RPV Isola-tion," direct s the operator to initiate SLC and increase CRDHS injection but also to " manually" open relief valves (i.e., to depressurize) if there is no level increase. From Fig. 4.2 of this report, we see that even for the case of initiation of 14.2 L/s (225 spm) total inj ection within 10 min af t er scr am, the water level centinues to decrease for the next 90 min, even though the core does not become uncovered. Therefore EDI-46 would probably be interpreted by the operators to require a f ast depre.suriza-tion even in the most f avorable case in which suf ficient CRD and SLC flow is started soon enough to prevent core uncovery. This action would transform an acceptable sequence with no fuel damage into one with the possibility of fuel melting. In the future, the E01s are to be replaced with, or supplemented w it h, a Browns Ferry-specific version of the Emergency Procedure Guide-line s (EPG s) .1 The EPG for level control is pref erable to E01-46 because it does not require a rapid depressurization if the CRDHS is running.
63 7.2 Procedural Considerations
. This section does not presume to specify every instruction or pre-caution that the emergency procedures should have. Features are discussed that, if incorporated into the procedures, would increase the chances that . the operators would, in the unlikely event of a TQUV accident, be able to prevent core damage.
The emergency procedures applicable to TQUV sequences should have a criterion to assure timely operator action to initiate SLC injection of non-borated water and to increase CRDHS flow. For example, an instruction for the operators to begin these actions before level decreases to more than 0.4 m (18 in.) below the setpoint for automatic start of RCIC and HPCI would ensure prompt operator response. A vessel water level 0.46 m (18 in.) below the RCIC and HPCI aut o-initiation setpoints is still more than 2.44 m (8 f t) above the top of the active fuel. If the operator actions to initiate SLC injection and to increase the CRDHS flow are succe ssful, the TQUV emergency procedure should not require a rapid depressuriz ation.* This recommendation is not intended to pro-hibit the 55.6 K/h (100*F/h) controlled depressurization that the emer-gency procedures require whenever supprossion pool temperature exceeds 322 K (120*F). The instructions for emergency use of the CRDHS should warn of the
- possibility of low suction pressure trip when the pump test bypass line (see Sect. 3.2 and Fig. 3.1) is used to inj ect water into the depressur-ized reactor vessel. Instructions for recovering from the pump trip should include not only a method for bypassing the pump suction strainers
! but al so for throt tling the pump discharge - (e.g. , by using valve 85-527) to limit excessive flow. Since low suction pressure trip is less likely with two CRD pumps running, it would be desirable to start the spare pump (if available) before opening the pump test bypass. The instructions for emergency use of the CRDHS should include an alternative method for enhancing injection to the reactor vessel by plac-ing the 85-11 flow control station (Fig. 3.1) in manual and fully opening the flow control valve. This would increase the vessel injection by by-passing the restricting orifices in the charging header. The extra flow through the 85-11 flow control station would be less than that possible with the more direct pump test bypass flow path; however, the 85-11 can be i opened quickly from the main control room without the requirement for operator action in the reactor building to open hand-operated valves. The l time advantage of opening the 85-11 control valve instead of using the i pump test bypass line would become more important if the reactor building were temporarily uninhabitable because of smoke, steam, or radiation. This sugge sted option for use of the 85-11 flow control station does not preclude the use of the pump test bypass line es presently described in
. plant operating procedures. . *Although not.in the scope of this report, a situation might arise in which the TQUV accident is accompanied by a loss of the high pressure CRD hydraulic and SLC systems but a non-ECCS low pressure inj ection source such as the condensate transfer pumps or the RER drain pumps was avail-able. In this very unlikely event, rapid depressurization would be re-quired.
64 The contradiction between 01-85 and E01-41 (see also Section 3.2 of this report) regarding the opening of the CRD pump throt tling valve 85-527 should be resolved. The 85-527 throttling valve must be opened fully in a IDUV emergency if fuel damage is to be avoided. If the hand-wheel has been removed as indicated by 01-85, then instructions should be provided on how it may be found and utilized in an emergency. In order to assure adequa te cooling for the CPD pumps af ter a loss of
- of f si te power, there should be a reminder in the procedures for the oper-ator to rest art a raw cooling water pump.
7.3 Recommended Procedures This section will not attempt to present every de tail of necessa ry instruction for the operators; rather, it will be a summary of the maj or items that should (or should not) be done if a TOUV accident occurred. The instructions for each of the following cases should result in core i recovery with no core damage. l 7.3.1 TQUV secuence not initiated by LOSP The maj or instructions for this scenario should include:
- 1. do not initiate emergency depressurization of the RPV,
- 2. start the second CRD pump, ' '
I
- 3. open the pump test bypass line (valves 85-551 and 85-50), open the 85-527 throttling valve, and
- 4. control the RPV pressure by manual actua tion of an SRV [6.21-7.59 MPa
- I (900-1100 psia)]. '
As shown in Fig. 4.5, these actions would inj ect 17.41 L/s (276 gpm) into the RPV and prevent core damage. If these steps are accomplished within 10 min of scram, the core will not be uncovered. How ev e r, to provide additional margin:
- 5. realign the SLC system for inj ection of demineralized water and ini-ti n te the syst em.
7.3.2 TDUV seatence initiated by LOSP The maj or instructions for this scenario should include:
- 1. do not initiate emergency depressurization of the RPV,
- 2. provide cooling water for the IB CRD pump (this may require some raw cooling water system realignment),
- 3. start the IB CRD pump,
- 4. open the pump test bypass line (valves 85-551 and 85-50) and fully open the 85-527 throttling valve,
- 5. initiate the SLC system (demineralized water injection),
- 6. control the RPV pressure by manual actua tion of an SRV [6.21-7.59 MPa (900-1100 psia)].
If these steps are accomplished within 10 min of scram, the core will not be uncovered.
65 7.3.3 TQUV secuence with SORY (not initiated by LOSp) The maj or instructions for this scenario should include:
- 1. start the second CRD pump,
- 2. open the pump test bypass line (valve s 85-551 and 85-50), and
- 3. open the CRD throttle valve 85-527, but throt tle as nece ssary to a limit total fl ow to (25.2 L/s (400 gpm), or bypa ss the CRD pump suc-tion strainers in order to avoid ' low suction precsure' trip of the CRD pumps (see Fig. 3.3).
To provide additional margin:
- 4. initiate the SLC system (demineralized water inj ect ion) . If these steps are accomplished within 20 min the core will be partially un-covered but will subsequently recover without severe f uel damage, f
7.3.4 14UV seauence with SORY (initiated by LOSP) The maj or instructions for this scenario should include:
- 1. provide cooling water for the IB CRD pump (this may require some raw cooling water syst em realignment),
- 2. start the IB CRD pump,
- 3. Initiate the SLC system (demineralized water inj ection),
- 4. bypa ss the CRD pump suction strainers,
- 5. open the pump test bypa ss line (valve s 85-551 and 85-50),
- 6. throttle the CRD flow via valve No. 85-527 to keep the CRD flow (16.4 L/s (260 gpm) to avoid pump trip on low suction pressure.
If these steps are accomplished within 20 min the core will be partially a uncovered but will subsequently recover without severe f uel damage. 7.3.5 Plant specific considerations As indicated on Figs. 4.5 and 4.8, there are many combinations of operator actions that can ef fectively prevent core damage through use of the CRD hydraulic and SLC syst ems. The steps listed in Sect s. 7.3.1 through 7.3.4 were selected based on considerations of the available equi pment a t the Browns Ferry plant. Other sequences of operator actions to accomplish the same goal might be preferable or nece sna ry at other plants. For example, many of the recent plant designs do not include the pump test bypa ss line to the reactor vessel . The operators could, how-ever, enhance the CRDHS inj ection flow by fully opening the 85-11 flow control valve, putting an extra drive water filter into service, and ope n-ing the 85-527 throttling valve ( se e Fig. 3.1) . Conversa tion with General Electric Co. personnel indicates that the resulting flow would be about 14.2 L/s (225 spm) with two pumps inj ecting against a reactor vessel pres-sure of 6.9 MPa (1000 psia). The corresponding one pump flow would be about 11.2 L/s (178 spm).
, 7.4 Recommended Eauinment_ Modifications In using the CRDHS as suggested in this report, the operators would not have an accurate indication of the flow being pumped by the CRDHS into the reactor vessel. Several flow indicators are provided in the main con-trol room, but these .are of limited maximum indicated flow: 6.3 L/s (100
66 gpm) for the system indica ting flow controller, 5.04 L/s (80 gpm) for the cooling water flow indicator, and 3.15 L/s (50 spm) for the return water fl ow indicator. These maximum indicated flows are well below the in-creased system flow that occurs af ter every reactor scram and the ~12.6 L/s (200 gpm) flow that would have to be inj ected in a TQUV emergency. The vessel water level indicators provide some flow-related information; how ev e r , the vessel water level would be slow to respond to flow changes and also would respond to other events such as SRV actuations. An unex-pect edly low CRDUS inj ect ed flow (but still high enough to overrange the existing meters) could be very difficult to detect by means of vessel water level indication. Upgrading of the maximum indicated range of the appropriate CRDHS flow indicators is recommended to provide a way for the operators to ver-ify adequate cooling by multiple indications. Brewns Ferry E01-41, " Water Makeup Nethods to the Reactor Vessel," specifies that the suction strainer of the CRD pump be bypassed when ne ce s sa ry t o pr ev e nt a l ow suction pressure trip of the pump. A recent licensee event report 8 from the Susquehanna 1 plant reported that the low suction pressure trip was excessively sensitive to pressure fluctuations when the suction strainers were bypassed. This problem was resolved at Susquehanna by installation of a time delay in the CRD pump low suction pressure trip circuit. The applicability of this problem and solution to the Browns Ferry CRD hydraulic system should be investigated. 7.5 Reference'
- 1. " Emergency Procedure Guidelines - BWh agh 6 (Revision 3)," pa e-publication draft, Dec. 8, 1982.
- 2. Licensec Event Report No. 82-022, Docke t No. 50-387, Pennsylvania Power and Light Company, Oct. 14, 1982, e
e 6
. _. . . - - , ~_ _ . . . _ . - _ . - . _ - . _ . - _ . - - _ ---_-__ _ _ _ -__
e e o e e e i i i i ORNL-DWG 83-4765 ETO CRD PUMP LOW RECOVERY FROM CRDHSSLC INITI ATION CRDHS/SLC INITI ATION FAST SUCTION PRESSURE CRO NMP LOW W6 THIN 10 man WITHIN 25 men DEPRESSURIZATION C>225 gom*) ION PRESSURE C>200 gpm* *) TRIP ( AUTOMATICS Y NA _N NA NA NO CORE N \(Y; UNCOVERY CORE UNCOVERY
- NO SEVERE FUEL Y
\ DAMAGE
\
\ N NA CORE UNCOVERY, 2 SEVERE FUEL DAMAGE y_ y POSSIBLE $
N y COR E UNCOVERY, SEVERE FUEL DAMAGE SSIBLE Y = YES N=NO NA = NOT APPLICABLE - FUEL MELTING
*225 gpm = 14.2 L/s
**200 gom = 12.6 Lh FUEL MELTING Fig. 7.1.. Operator action event tree for 'IDW accidents initiated from 100E power.
4<
68
- 8. DISCUSSION OF UNCERTAU4 TIES 1he results presented in this report are dependent on the accuracy of the computer codes employed as well as the accuracy of the data utilized as input for these codes. This chapter discusses uncertainties associated with the model s inside the BWR-LACP and the MARCH code s, and significant uncertainties associated with input data.
The calculation of reactor vessel pressure and water level prior to core uncovery was performed using the ORNL-developed BWR-LACP code which incorporates reactor vessel and inj ection systems specific to Browns Ferry Unit 1. The BWR-LACP code has al so been used in three previcus SASA studies. Additions to the code for the present study are described ir Appendix A of this report. BWR-LACP results for a Sta tion Blackout accident sequence have been compared to results calculated for the same sequence by the Browns Ferry simulator and RELAP4 Mod 7 (Ref.1) . Code result s for a small-break LOCA with condensate booster pump inj ection have been compared with results calculated for the same sequence by RELAPS Mod 1 (Ref. 2) . Agre em e nt is good in all cases. The IQUV sequences studied in this report are suf fi-ciently similar to the compa rison cases that it can be concluded that the BWR-LACP code does not introduce significant uncertainties. Uncertainties do exist in the input parameters supplied to the BWR-l LACP code for the study of the TQUV accident sequence before core un- + cove ry. These include :
- 1. The primsry system event during the very brief period (~1 min) af ter scram and MSIV closure when multiple SRVs are open and the feedwater -
turbines are coasting down cannot be modeled by BWR-LACP. Normal reactor vessel indicated water level is 561 in, above vessel zero and the BWR-LACP calculations are begun at time 30 s with a water level equivalent to an indicated 500 in. in consideration of the ef fect of the level shrink that occurs following MSIV closure. This assumption is based upon the accident studies presented in Chap.14 of the Browns Ferry FSAR and upon the level indications at the Browns Ferry simulator when scram and MSIV closure are simuisted. For a IQUV sequence initiated by a loss of all main feedwater, the main turbine can continue to draw s t eam from the reactor vessel until the MSIVs automatically shut on indicated water level below 470 in. For an indicated starting water level of 470 in. Instead of 500 in., there would be ~20% less water above the top of the core. Therefore, for the case of no inj ection af ter reactor scram, the core would begin to be un-covered ~20% sooner, i.e., af ter about 20 min instead of 28 min. Th e co r-responding dif ferences for ca ses with inj ection would be proportionally sm al l e r. The calculations of this report were performed with the 12.7 m (500 in.) initial level because this is representative of the maj ority of , possible TQUV initiators.
- 2. The time at which all vessel water inj ection is lost can have a significant ef fect on the timing of events in a 10UV sequence. In all ,
cal cul a ti ons, it was assumed that all of the high pressure inj ection sys-tens f ailed at the time of reactor scram. This is conservative because if the main feedwater system, HPCI, or. RCIC syst ems conti nued t o operate f or even a few minutes af ter the reactor scram, a sizeable volume of water
69 I would be inj ected into the vessel, and the time available for operator action to prevent core uncovery would be greatly expanded. ! 3. The inj ection capability of the SLC system is assuned to be 3.53 i L/s (56 spm), the capacity specified in Browns Ferry E01-41, " Water Makeup Methods to the Reactor Vessel." On the other hand, th e Brow ns Fe r ry FSAR I specifies that the SLC pumping rate is between 3.15 and 3.53 L/s (50 and 56 spm), and the Browns Ferry Technical Specifications require periodic i survelliance t o sh ow th a t t he SLC sy st en ca n pump a t least 2.46 L/s (39
- spm) of borated water from the SLC solution tark (and recirculate it back
- to the tank). Since the SLC system has positive displacement pumps, this j surveillance should also apply to the inj ection of water into the reactor v e s sel . The uncertainty in SLC capacity is not great enough to adversely af fect the conclusions of this report.
- 4. The pumping capability of the CRDHS is very important t o t h e r e-suit s and conclusions of this report. The flow delivered into the reactor vessel depends on vessel pressure, flow resistance of the multiple flow pa ths into the vessel, the number (1 or 2) of opera ti ng CRD pumps, and the CRD pump characteristics. A model that take s all these f act ors into ac-count (see Appendix A for description) was utilized to generate the i n-jected flow results presented in Table 3.1 and Figs. 3.2 and 3.3 and with-
, in the BWR-LACP and MARCH runs of Chaps. 4, 5 and 6.
l The CRDHS flow model depends on pressure drop and fic= information taken f rom the Browns Ferry FSAR, the CRD pump head vs capacity curve sup-
, plied by TVA, and ORNL estimates of some of the flow resistances. Th e se input data are felt to be suitably accurate for this study. Pow ev e r, it should be noted that the actual inj ected flow af ter scram and the actual , inj ected flow that can be delivered through the punp test bypass path cannot be measured at Browns Ferry because the existing installed flew-meters do not have high enough ranges [e.g., the indicating flow control a
station has the highest indication, and this is only 6.3 L/s (100 gpm)]. The ORNL version of the MARCII code was used to determine the core and i RPV response during the period of core uncovery. The general limitations 1 of this code with respect to LWR accident analysis with emphasis on PWR applications have been discussed extensively by Rivard,8 and the specific limitations with regard to BWR accident analysis have been identified by Greene.* Two of the major MARCH code deficiencies noted by Greene (the core model and the safety relief valve model) were corrected prior to this study, and these MARCH code improvements are discussed fully in Appendix B. One of the greatest uncertainties in applica tion of MARCH to BWR ac-cident analysis identified by Greene was the MARCH treatment (l.c., model s) of in-vessel structural def ormations end . f ailures. MARCH al l ow s the user to select one of three nonnechanistic core mel t model s and essen-tially allows the user to specify when the core ' collapse s. ' (The enti re core is assumed to collapse when a specified input fraction of the core becomes mol ten. ) None of the three MARC 11 model s treat the rod, canister, control blade melt / flow / freeze phenomena or core collapse mechanistic-al ly.
- Also, due to the design of the BWR core support sy s t em (sec
- Efforts are currently rnder way at Rer..selaer Polytechnic Institute under Drs. R. T. Lahey, Jr., and Michael Pedowski to mechani stically model these phenomena (and al so the corium progression in the lower head and RPV failure). This work is being sponsored by the ORNL SASA Progrse.
l
i 70 Appe ndix d) , it is very unlikely that a BWR reactor vessel would f ail vis a corium attack on the vessel lower head as modeled by MARCH. The corium would more probably f ail the vessel by erosion of one (or several) of the
- lower head penetrations (i.e., CRD guide tube s, stub tubes, in-core in-strumentation tube s, etc.). The point is that MARCH handles these phe-nomena simplistically and with regard to BWRs, nonr e alistically. Thus the times specified in Chap. 5 for core slump and RPV f ailure should be used .
f or comparativo purpose s and 391 for realistic estimates of when such event s would occur. The containment failure mechanisms used in the MARCH analysis of the TUUV sequence s include (1) failure due to overpressurization and (2) fail-ure due to excessive drywell atmosphere temperature. A value of 0.91 MPa (117 psis) was assumed for the pressurization f ailure limit (based on the information presented in Ref. 5) with probable f ailure being at the 1 cylinder-sphere intersection in the drywell.s For those TQUV sequence s studied (Chap. 5) in which the RPV failed, the drywell abnosphere tempera-ture exceeded the 411 K (2818F) design temperature of the drywell before overpressuriz ation f ail ure occurred. Therefore, overtemperature f ailure l was considered. Yue postulated
- that probable drywell failure would occur !
via electrical penetration assembly (EPA) elastomer seal degradation at drywell atmospheric temperatures in excess of 477 K (400'F). The combined j cross-sectional flow area of all the drywell EPAs in Unit 1 is ~1.95 m* (21 ft2). Thus, the assaantion was made that the EPA seals would begin , leaking at 477 K (400*F) and f ail completely at a temperature of $33 K , ) (500*F). For the RPV and containment f ailure cases presented in Chap. 5, a comparison of the containment f ailure times by overpressurization and overtemperature is presented in Table 8.1. The containment f ailure times
- by overtemperature (in' Table 8.1) that are less than tho' overpressuriza-tion failure times were used as the containment failure times in Chap. 5.
The MARCH in-vossel flashing models flash steam in bursts a ,7 that pressurize the vessel for several subsequent time steps during which all of the decay heat f rom the submerged fuel is transferred to the subcooled water and no steam is produced. This obviously af fects the Zr oxidation rate which is limited by the availability of steam. This lack of steam due to pressurization causes oscillation in the Zr oxidation rate which can decrease the total Zr oxidation. These oscillations' have no apparent physical ba sis and, in fact, are numerical in nature and caused by the manner in which flashing is modeled in MARCH 1.1.
"In-vessel flashing occurs when the predicted temperature of the re-1 sidual water exceeds the predicted saturation temperature at the system pr e s s ur e. Such flashing can be the result of system depressurization, heatup of the residual water, or both. In any case, flashing alone should not result in systen repressurization and resubcooling as predicted by the MARCH 1.1 model s. Instead flashing should produce just enough steam to keep the resulting water and saturation temperatures equal."? '
Sandia' is modifying the MARCH flashing models to eliminate nonne-chanistic oscillations in the steam generation and Zr oxidation rates in MARGI 1.1. I In light of the realistic in-vessel pressure cycles predicted with a the new ORNL SRV model, the in-vessel flashing models in ORNL's version of NARCH could have an impact on the predicted BWR core response. However, the Sandia model changes (which are in a development stage) have not yet i t
l 71 l been incorporated in the ORNL version of MARCH, and, at this point in time, it is not possible to quantify the ef fect these flashing model changes would have on the TQUV results presented here. Not only would the
. steaming rate affect the Zr oxida tion rate, but an increased (and more consistent) steaming rate would also provide better steam cooling of the core. These are essentially competing phenomena; with increased oxidation . the reaction energy input is greater and the core and steam temperature will increase, but with better steam cooling, the core temperatures will be lower and thus the oxidation reaction rate will be lower. Thi s i s a
! dif ficult question (quantifying the ef fect of the flashing model) and the effect is not intuitively obvious. Until both the SRV model and the more mechanistic Sandia flashing model are incorporated in the same code, the ef fect of the combined models is purely speculative. 'l In Chap. 6, the results of the MARCH computer runs for the two a t-pressure cases predicted a period of high fuel rod temperatures [>1255 K (1800'F) but well below the melt temperature of ~2128 K (3370'F)] with some cladding oxidation (<0.9%) followed by complete core recovery with little damage. These temperatures might lead to thermal distortion of the f uel assemblies.
- Also, the rod thermal responses in these sequences are l cyclic in nature (i.e., the rod heats up during the repressurization of 1
the vessel and then is rapidly cooled when the SRV actuates). The high tenperatures and the thermal cycling of the cladding could result in clad-ding failure. It should al so be noted that the maximum inj ection in the
, two cases discussed in Chap. 6 was 12.6 L/s (200 gpm)t whereas (from Fig.
4.5), at BFNP the alternative inj ection capabilities (for core recovery with no core damage) range f rom 13.0 L/s (206 gpm) to 21 L/s (332 spm). With these higher inj ection rates, the expected fuel temperatures (and
! thermal cycles) would be much less than those reported in Chap. 6.
4
- Fuel assemblies in the prototypical thermal hydraulics (PTB) test series of the NRU Test Program (conducted by Battelle Northwest Laboratory) a t Chalk River, Canada, which were subj ected to ~1367 K (2000'F) temperatures were thermally distorted.'
fMinimum inj ection required for complete core recovery with no core damage from the parametric studies in Chap. 4
72 Jeferences for _ Chapter 8
- 1. R. M. Harrington, " Comparison of Station Blackout Calculations on BWit- ,
IACP, IVA Browns Ferry Simulator, and RELAP IV Mod 7," letter report to NRC SASA Program Technical Monitor, Au8ust 1981.
- 2. W. C. Jouse and R. R. Schultz, A RELAPS Analysis of a Break in the Scram Discharge Volwte at the Browne Ferry Unit One Plant, IEG--tlTAl" 5993, August 1982 (also published as Appendix G to NURED/CR-2672, ORNL/TM-8119/V1).
- 3. 3. B. Rivard et al., Interim Technical Assessment of the MARCH Code, NURIE/CR-2285, SAND 81-1672.P3, Novembe r 1981.
- 4. S. R. Greene et al., SBLOCA Outside Containment at Broene Ferry Unit One - Accident Sequence Anatyeis, NURIE/CR-2672, Vol .1, ORNL/TM-8119/V1, November 1982, Appendix B.
- 5. l.. G. Greimann et al., Reliability Analyses of Steel Containment Strength, NURIE/CR-2442, June 1982.
- 6. D. D. Yue et al., Station Blackout at Broene Ferry Unit One - Accidant Sequence Analysia, NURIE/CR-2182, Vol.1, ORNL/TM-455/V1, Novembe r
- 1981.
- 7. F. Eric Haskin, Nuclear Systems Technology Division, Sandia National Laboratory, personal communication, March 1983. .
- 8. R. H. Chapman, Manager of ORNL MRET Program, personal communica tion.
l t e 4
73
- e t
Table 8.1 Comparison of containment f ailure tinies" for the no-operator-action cases presented in Chap. 5 Failure via Failure via , Case 3 overpres s uriz a tion over t empera tur e.
"At pressure"
- 1. With no CRD inj ec- 311 267 tica
- 2. With average CRD 444 430 inj ection of 6.68 L/s "SORV"
. 1. With no inj ection 340 286
- 2. With CRD inj ection# 511 472
" minutes after reactor scram.
b Pressure f ailure limit of 0.91 KPa (117 psig). Temperature f ailure limit of EPAs at 533 I (500*F). d RPV pressure control be tween 7.38 and 7.72 MPs. CRD inj ection increases from 6.56 L/s at 7.58 MPa to 11.36 L/s af ter 60 min. 9
75 Appendix A. MODIFICATION TO IllE BWR-LACP CODE FOR 11 TIS S111DY The only modification required was the programming to solve for the CRD hydraulic system (CRDIIS) flow inj ected into the reactor vessel as a f unction of reactor vessel pressure, the number (1 or 2) of CRDilS pumps l . runni ng, and the flow resistance. Straightforward application of Bernoulli's law between the condensate storage tank (CST) and the reactor vessel (RV) yields the following equa-tion: h = 2.32 P + 35 + h g p where h = head developed by the CRDIIS pump (s) (ft), P P = reactor vessel pressure (psi g), 2.32 = 144 divided by density of water at 90*F, 35 = elevation dif ference between the CST and RV (f t), h3 = total unrecoverable losses between the CST and RV (f t). This equation is a simple statement that the CRDHS pump must develop suf-
; . ficient head to pump against reactor pressure while of fsetting the eleva-tion dif ference and unrecoverable (i.e., frictional and shock) l o s se s.
After expressing the h and h g terms as quadratics in flow, the above equation can be rearranged as a single quadratic equation which can then be solved for the unknown, flow, using the f amiliar relationship for the l a r ge s t root of a quadratic polynomial:' B , = (-b + Yb 8 - 4ac)/2a where B = fl ow inj ected into reactor vessel (gpm), 3 a= (Eg - h/Ng), 2 b = -(h 3 + 2h 2B,) /N p c = 35 + 2.32 P - ( h, + h B, 3 + h B2 2 ) , B, = flow recirculated to pump suction (gpm per pump), N = number of CRDIIS pumps running, . p K g = loss coefficient, h,h,h e i 3 = coef ficients developed f rom least squares pump curve
, fit.
The loss coef ficient, E, relates the unrecoverable losses to the inj ected g flow as follows
- Actual values of system parameters lead to only one positive root.
1 76 4 hg = K B{ , and the pump developed head is expressed as a quadratic fuiction of the bulk flow per pump: . h = h, + h (Bg 3 + B,) + h (Bg 3 + B,)2 The values for the coefficients were determined by a least squares fit to the TVA-supplied pump head vs capacity curve, resulting in the f ollowing expression: h = 3816 + 1.552(B3 + B,) - 0.02517(Bg + B,) * . The calculation of flow coef ficient, g K , is more invoM Mcaun two pa rallel flow pa ths were considered: (1) the normal post-scram flow pa th through the charging header, and (2) the pump test bypa ss flow path (which can only be opened by operator action). Thus, the loss term Kg is a com-posite of the two parallel flow pa ths and al so of a third, series resist-ance - the 85-527 throttling valve
- through which all inj ection must flow before entering the charging header or the pump test bypass header:
Kg = 1/(1/YK + 1/V K )*+K tv
- where .
K = overall loss coef ficient, g K = loss e ef ficient for flow through the charging header, d K =1 ss c ef ficient for fim through the pump test bypass, tb
= loss coef ficient for the 85-527 throttling valve.
Note that prior to operator action to enhance CRDHS flow, there is no flow through the pump test bypa ss, and K is effectively infinite, so the The calculations presented in above expression reduces to K
=Kd+E .
this report utilized a value of the (K +K ty that was krived hem Fig. 3.4-10 of the Browns Ferry FSAR: (K K ) = 0.0954 ft/(spm8). There was no FSAR data on flow {Ir+ougd*the pump test bypa ss, so a value was estimated by taking into account the known fittings and valves in each segment of piping of different diameter: K = 0.01 ft/(spm 8). The flow resistance of the 85-527 throttling valve was inferred from Browns Ferry OI-85, " Control Rod Drive Hydraulic System," which states that the valve has been set to supply a 1500 psis control rod drive pump discharge pressure (i.e. , as measured directly downstream from the valve) under normal conditions (before scram, with 60 spa inj ection flow and 80 gpm total pump flow, considering the 20 spa pump recirculation flow). This information is sufficient to allow the pressure drop across the throttling valve to be estimated because the pump developed head is avail-abic frce the head capacity curve, and because the relative elevations of the CST and the pump discharge header are known. The calculated estimate of throttling valve flow coef ficient is:
*See Fig. 3.1.
._a - y w w_ a - - - - - . r , _ _ * - _ _ _ . - -
77 K 2 0.0507 ft/(gpm2). For ca se s in which the 85-527 valve is fully opened a s specified by Browns Fe r ry E01-41, it is assumed that K is reduced by f actor of ten to 0.00507 ft/(gpm 2). As part of the CRDIIS modeling, an expression was developed to calcu-
- 1 ate the pump suction pressure. This i s nece ssa ry in order to determine at what flow the pump (s) would trip on low suction pressure (se t a t 18 in.
mercury absolvt e, or 8.8 psi a) . The cal cula tion proceeds f rom Bernoulli's Law : P =P + AP el - AP ps a tm ey 1,a , where P, = pump suction pres sure, AP,3,v = 21.2 + 0.431*(CST level) (psi), AP = unrecoverable pressure losses plus acceleration pressure drop (psi). The value of the loss term, AP3 ,, was e stimated f rom the cal cula ted ope r-sting conditions on Fig. 3.4-10 of the FSAR: AP 1,a
= 0.000516 B2p, where B = flow per pump (gpm) .
p The me; hod described above for calculating pump suction pressure was developed under the assumption that 100% of the pump suction pressure loss occurs in the pump suction piping that is specific to each pump. This assenption is reasonable since the piping between the CST and CRD pumps that is common to both pumps is much larger, and ..nce the pump-specific piping includes the suction strainer which is the largest single loss. 9
- - -- .. - - , = _ _ . _ _ _ - _ _ _ -
79 l l l l Annendix_B l l . MARCH 1.1 CODE IMPROVEMENTS FOR BWR DEGRADED CORE S'IUDIES I i
, B.1 Introduction Since its release by Battelle Colunbus Laboratories and the Nuclear Regulatory Commission in 1980, the MARCH 1.1 code has been a widely used tool for LWR accident analysis applications. A primary guideline in the development of the code was that the code must be able to analyze the en-tire, prolonged course of a postulated core meltdown accident. This was accomplished in many instances by simplistic modeling of very complex phenomena; but it did result in a fast running, integral meltdown code.
The general limitations of MARCH 1.1 with respect to LWR accident analysis with emphasis on PWR applications have been extensively discussed el sew her e.1 However, the specific limitations with regard to BWR analysis are even more confining because MARCH f ails to consider many of the basic design and operating differences which distinguish BWRs from PWRs. As an outgrowth of the Severe Accident Sequence Analysis (SASA) Program MARCH l BWR application experience at ORNL, a broad range of problems associated
! with MARCH modeling of BWR accidents has been identified and is discussed . in Appendix B of Ref. 2.
l Two of the maj or MARCH code deficiencies identified by Greene* were corrected prior to this study and they are the subj ects of this appendix.
, The limitations in the original MARCH 1.1 code that will be addressed here include :
- 1. (Section B.2) the core model does not explicitly represent the Zir-caloy channel boxes and the control blades in a BWR core, i 2. (Section B.3) reactor pressure control by the safety relief valves (SRVs) is not correctly represented in that a continuous release of steam from the reactor vessel is modeled instead of the actual peri-odic blowdowns followed by relatively quiescent periods within the reactor vessel be tween relief valve " pops."
The implementation of the new SRV model lead to the discovery of ad-ditional problem areas in MARCH:
- 1. The steam / hydrogen physical properties and the in-core gaseous con-vective heat transf er correlation which will be addressed in section B.4, and
- 2. the core quench models (to be considered in Sect. 5.5).
Finally, a brief discussion of the impact of these various model changes on accident events and timing will be made in Sect. B 6. j The primary purpose of the ORNL modification of the MARCH code has j been (1) to permit determination of the extent of BWR core damage for the l accident sequences presented in this study and (2) to improve the code l capability of analyzing the events af ter core uncovery in all BWR accident 4 scenarios. e , , - , , - --r~ ,- - -- , s -s -
80 B.2 BWR Core Model } B.2.1 BWR reactor vessel and core descriptions , The BWR reactor assembly (Fig. B.1) consists of the reactor vessel, internal components of the core, shroud, top guide assembly, core plate , assembly, shroud head, steam separator, dryer assemblies, and j et pumps. Also included in the reactor assembly are the in-core neutron flux moni-tors, control rods, ORD housings and the utDs.
! Most of the fuel assemblies (Fig. B.2) that make up the core rest on j orificed fuel supports mounted on top of the control rod quide tubes (Fig.
B.3). Each guide tube, with its fuel support piece, bears the weight of four fuel assemblies and is supported by a CRD penetration nozzle (stub tube ) in the bottom head of the reactor vessel. The core plate provide s lateral guidance at the top of each control rod guide tube. The top guide provide s lateral support for the top of each fuel assembly. BWR control rods (Fig. B.4) occupy alternate spaces be tween fuel as-semblies and can te withdrawn into the guide tube s below the core (Fig. B.3) during plant operation. The reds are coupled to CRDs mounted within housings (Fig. B.1 and B.3) that are welded to the stub tubes in the bottom head of the reactor vessel. The cruciform control rods (Fig. B.4) contain stainless-steel tubes filled with boron carbide powder (neutron a bs orbe r) . These tubes are held in a cruciform array by a stainless-steel sheath extending the full length of the tube s. A top casting and handle aligns the tube s, provide s structural rigidity, and contains positioning I rollers; and there is a parachute-shaped velocity limiter at the hottom of the control rod to mitigate rod drop or ej ection-type ac cide nt s. There
- are 185 control rods in Unit I and the other Browns Ferry reactors.
A typical BWR fuel assembly is shown in Fig. B.2. The combination of a fuel bundle and a fuel channel is called a fuel assembly. The Zircaloy-4 fuel channels direct the core coolant flow through each fuel bundle and also serve to guide the control rods. The channels give increased operat-ing flexibility since each assembly can be orificed separately thus allow-ing easy change of reload fuel design. The channels do separate the core coolant flow area (inside the core shroud) into two regions: (1) the in-terstitial region (i.e. outside the channels and not in direct contact with the fuel rods), and (2) the region inside the channels in which flow directly cools the fuel rods. Using the cycle 4 core (which is a combina-tion of 7 x 7, 8 x 8, 8 x 8 R fuel assemblies) in unit 1 at Browns Ferry i as a ba sis, the interstitial region is ~42.2% of the total incore coolant flow area but only ~10% of the coolant passes through the interstitial region. The boiling water reactor core (i.e. In side the core shroud, above the core plate and below the top 'fulde) is essentially comprised of only three components: (1) fuel assemblies (764 in unit 1 at BFNP), (2) con- . trol rods (185 in unit 1 at BFNP) and (3) in-core neutron' flux monitors. The BWR core model (to be discussed in B.2.2) will be limited to the ac-tive core region with the control rods inserted in the core. ,
81 B.2.2 Core best transfer models
- The MARG 1.1 core is divided into two heat transfer regions, a gas (steam acd hydrogen) covered region and a region covered with a water-steam mixture. For a fuel-rod node in the gas-covered region, the MARG 1.1 generalized heat balanced is l
8T RHOCU x V + R melt R " b, R + NW + mel t (B.1)
-Q ~
rad R~ G' ' where VR "
- I""* I " d*'
RHOCU = heat capacity of core mater?. d per unit volume, T = temperature of node, R Fg = fraction of node melted per unit time, A = heat of fusion of core material in the node, pVR " ""** # * "" *# * " ** ** Q = decay power in the node, D, R Oy , = heat from metal-water reaction, Qg = heat added to node from slumping during meltdown, Q rad " #* * " ** ~ #*"*'*# * * * *
- h = combined radiative and convective heat transfer coefficient 3
(to the gas), Tg= ga s temperature in the node, F. The increase in the gas temperature along the length of the channel is given by BT inC = hp(TR -T)g , (B.2) p 8Z where m = total gas flow rate, Z = distance measured up the channel, p = fuel rod circumference. In e ssence, MARG 1.1 model s (Eqs. (B.1) and (B.2), via lumped parameter techniques) the simplified nodal schematic shown in Fig. B.5. This is representative modeling of a PWR core. Significant structures of a BWR core (i.e. channels and control blade s) are not explicitly represented. The MAR G user must either ignore the channels altogether, or artificially lump the mass of the channels into the cladding of the fuel pin, or (simulating the correct Zr surf ace area) create additional (artificial) fuel pins. Any of these alternatives
82 can lead to additional uncertainties in the MARQI computed results; but to ignore these structures is probably the worst alternative since ~34% (by mass) of the in-core Zircaloy (Zr) and ~30% of the in-core Zr surf ace area , is contained in the channel walls. These alternatives can obviously have a great impact on the calculated energy release and hydrogen production a seciated with in-core zirconium oxidation reactions. , The MARQ1 1.1 modeling problems (as documented by Rivard and Greene) can significantly compromise the results of both PRA and SASA analyses due to resulting uncertainties in the computed reactor and containment re-spon se s. Given this motivation, a j oint interlaboratory upgrading of MARQ1 1.1 was undertaken in 1982 (with the release of MARGI 2.0 expected in January 1983) . As part of this maj or code upgrade, ORNL was given the
^
task of developing simplified channel / control blade models to analyze the heatup, oxidation, and melting of BWR channels and contrci blade s. These models were to be used in conjunction with the existing fuel rod heatup model in the BOIL subroutine of MARQl; that is, the models were to be de- . veloped and incorporated in the ORNL version of MARGI and then (after com-plete debugging and testing) transferred to Bette11e Columbus Laboratories who had the primary responsibility for the MARGI 2.0 and 2.1 development. The schematic shown in Fig. B.6 depicts the heat inputs and the hest transf er mechanisms included in the ORNL models for the maj or structurst - components in a BWR core. Each radial region (up to 10) and axial node (up to 50) in the core is represented via the following generalized heat- - bal ance s: ~~ ' fuel rod, ROD ROD ROD at MELT ' ROD ROD
+O -O ~0 (B.3)
MFLT RAD RAD ROD ROD-GAS CORE
~ kADROD-STRUCTURE OD ROD ~ kAS WRE OR WATER l channel box, BOX BOX BOX at MELT BOX ROD-BOX BOX-01 ,
O ~O ~ R RAD BOX BOX-SIRUCTURE
- BOX-GASQ)RE OR (B.4)
WATER l 1 - --
83
~
I BOX ~ GAS CORE BOX-GAS INTER gg BOX ~ GAS WB03 NTER CORE INTER control blade, BT RII0g gV g +F ET g (A"} G VG"QRADg CD-STRUCIURE/ WATER G-GAS INTER NTE (B.5) x (T cg -T GAS
'+ w +
MELTg mTER G where V g = vol ume of node,
, RHO, = heat capacity of material per unit volume, T, = structural node temperature, , F = fraction of node melted per unit time,
( Ap) *. = product of the density and heat of fusion of the node ma t er.i al, Q D, R " #"I E "'# " *" Og = heat from metal-water reaction, O ET = heat added to node from slumping during meltdown, g Qg = radiation heat-transfer from x to y, x-(hA)3(Tg -T GAS = c wective heat transfer fm i to gas stream, T = ga s n d al t empe r a tu r e, GAS g subscripts: ROD = fuel rod, BOX = channel box, CB = control blade, GAS = steam / hydrogen mixture, CORE = refers to the inside of the. canisters,
. INTER = ref ers to the interstitial ' side of the canisters.
The increase in the gas temperature along the length of a channel inside
84 the canister box is given by GAS CORE
" GAS ^ CORE CORE CORE
* $0X CORE CORE CORE and out side the canister box (i.e. the interstitial region) is given by GAS P 8Z OI BOX gg
'G AS g GAS NT (TBox -T GAS ) + hgpCB (TCB -T GAS g mm where m = total gas flow (mass) rate, ggg ,
Cp g = specific heat of the gas, z = distance measured up the channel. n* = combined radiative and convective heat transfer coef ficient (to the gas), p g = structural perimeter. e These models were incorporated in the BOIL subroutine of MARCH 1.1 and use supporting calculations f rom BOIL (i.e. h's, Cp' s , etc.). The basic structure of the BOIL subroutine was not changed and existing MARCH melt model s were augmented to include the channels and control blades. The ORNL IIYR core model allows explicit determination of the tempera-ture of the active core structures (f uel rods, canisters, and control bl ade s) and gas temperatures (inside the canisters and in the intersti-ti al region) as functions of time .and position. The model detdraines the oxidation on all core surf aces. The model neglect s level swell in the interstitial region, thus there are two incore levels: (1) the collapsed level for the interstitial region and between the core shroud and the RPV wall) and (2) the swollen level (Wilson model) inside the canisters. The model does not include axial or inter-radial heat transfer within the core (nor does the original MARCH core model). B.3 BWR Reactor Pressure Relief Model . B . 3 .1 Description of BWR nressure relief system s ,s The safety relief valves are dual-function valves discharging di-rectly to the pressure suppression pool. The saf ety function include s protectior. against overpressure of the reactor primary system. The relief function provide s. power-actuated valves opening to relieve steam during 4-
85 transients resulting in high system pressure or, during postulated acci-dent conditions, to depressurize the reactor primary system. To a c com p-i list these obj ectives, thirteen safety / relief valves are located on the l steam lines within the drywell. Set pressures for the valves at BFNP are listed in' Table B-1. The 13 safety-relief valves at BFNP unit 1 are the " dual purpose" )* Target Rock type - self actuating at the set relieving pressure. valves are pilot operated which permits remote-manual opening at pressures The below the setpoint. Each valve's discharge terminates below the water level in the suppression pool to permit the steam to condense in the pool. When these valves actuate automatically on high pressure, they will closc when the pressure f alls to a preset pressure of 345 kPa (50 psi) (at BFNP) below the setpoi nt pressure. That i s, once the valve (s) opens, the RPV will essentially blowdown 345 kPa before the valve (s) resents. Also, these valves are designed to be critical flow devices: that is, the dis-charge rate through the valve is independent of the downstream pressure - it is dependent only on the pressure within the RPV. B.3.2 Pressure relief model The MARCH 1.1 computation of steam flow through a safety / relief valve is given by* W=C75 (B.8) l where W = mass flow rate, i
" C = essentially a valve flow coef ficient and conversion f actor, p = s t e am de n si ty, AP = pressure difference between the RPV and containment. .
This value is calculated in subroutine PRIMP which also contains models which "will reduce the vent valve leakage so that the system pressure does not f all bel ow the relief valve setpoint, PS ET. Thus, if the leak rate does not exceed the valve capacity, a constant leak rate is calculated, which is the leakage required to maintain the system pressure at PSET. Opening and closing cycles for the vent valves are not explicitly mod-eled."* In the LDHR accident sequence analysis 8 (recently completed by the - ORNL SASA program) the MARCH predicted reactor pressure, prior to core s1 map, is shown in Fig. B.7. Note that the pressure is ~8.55 mPs (~1240 psia) prior to containment f ailure. This is not physically possible (see a Tabl e B.1) , but, since ~ the flow model is related to the cor.tainment pres-sure (Eq. B.8), the predicted pressure will approximately be equal to (PSET + containment pressure). After containment ! allure, the containment pres sure --a tmospheric pressure and the reactor pres sure ---PSET. During this period (from containment f ailure to core slaap), the valve . steam flow
.is as shown in Fig. B.8 and the vessel collapsed water level steadily decreases (Fig. B.9).. As noted above, the pressure trace (Fig. B.7) is incorrect and the flow through the valve should resemble a square wave U
86 (or, considering the time scale, a series of delta functions) and the collapsed level should resemble stair-steps. As discussed in Sect. B.3.1, BWR safety / relief valves are critical , flow devices which actually blowdown af ter actuation by high pressure (pres sure > PSET) . Thus, the reactor pressure response will resemble a sawtooth pattern, cycling between PSET and (PSET - 345 kPa) with the , valve showing approximately constant flow during actuation periods and zero flow during periods when the pressure < PSET. Instead of a steady decline in the vessel water level, the level trace will be j agged because of loss of inventory during the blowdowns and the swollen level will rise dramatically (due to flashing) during the blowdowns. Prior to this (TQUV) study, ORNL replaced the MARCH relief valve model described in the first paragraph of this section with a model which more aptly describes the actual behavior as described in the previous para-graph. The ORNL model is essentially given by (B.9) I"I RATEL(PRPV tr JRATED p pGAS)0.5 where W = gas mass flow rate through the SRV, W ~ #" * ""** ' " #' * * "I *
- RATED P = reactor pressure, RPV pGAS " 8"
- d'"'I *7' *
=
(Pp)RA' FED the density
- Pressure product at which WRATED ***
determined. Equation B.9, toge ther with the appropriate programming logic to keep the SRV open until the reactor pressure is less than (PSET - 345 kPa), consti-tutes the ORNL BWR SRV model. Af ter implementation of the new SRV model, a sample run of the TQUV sequence was made. Plots of the sample run results for the reactor pres-sure, relief valve gas flow, and collapsed and swollen reactor water lev-els are presented respectively in Figs. B.10 through B.13. These vari-ables behave as expected and much more realistically. One surprising re- I sult was the dramatic rise of the swollen water level during relief valve actuation; even through the collapsed level falls to ~8.56 m (337 in.) above vessel zero, thus uncovering ~0.68 m of the core, the swollen level dramatically rises ~0.4 m above the-top of the active fuel during each
" pop" of the relief valves. Thus during each valve actuation the active fuel would 1e quenched.
B.4 MARCH Steam /Hydromen Physical Pronerties and the Incore Gaseous Convective Heat Transfer Correlation 1 The implementation of the new SRV model in MARCH lead to the discov- I ery of additional problem areas in MARCH Version 1.1. 1 m - -
87 ] i
- 1. the steam / hydrogen physical properties and the incore gaseous convective heat transfer correlation, and
- 2. the core quench models (to be considered in Sect. B.5) .
In the gas covered region of the core, MARCH determines
- the i convective heat transfer coefficient by taking the maximum of a natural l convective coefficient and h calculated by e C j
h = 0.0144 Cp08.8/D*.* c (B.10) j where l G = gas mass flux, j D = bydraulic diameter, i Cp = mass average of the specific heats of steam and hydrogen. ! The correlation given by Eq. (B.10) is a simplification of the Dittus-j Boelter correlation 7 and is very limited in application. Equation (B.10) j is intended 8 only for flow inside tubes with gases at ordinary pressures j and temperatures (based on Cpp/k = 0.78 and p = 1760 n s/m2 (0.0426 lb/ i h* f t * ) . This simplification is essentially for air at temperatures (310 K { (100*F); it should not be used for stesm/ hydrogen mixtures at elevated ! temperatures and pressures. At low pressures, Eq. (B.10) predict s an h l for steam which is +15% of the Dittus-Boelter prediction.
- , An attractive feature of Eq. (B.10) is that only one physical prop-
- erty of the gas (the specific heat) needs to be calculated. For the
, Dittus-Boelter correlation, the specific heat (Cp), thermal conductivity ! (k) and viscosity (p) all need to be determined, and for a ga seous mix- )* ture, these properties would need to be calculated for each species and then the mean property values used in the correlation. It is apparent ] that the computational time would increase dramatically with the Dittus-j Boelter correlation, especially if the functional form for the properties is more complex than simple polynomials. A comparison of the steam specific heat calculated 4 in BOIL and the steam tables value s* is shown in Fig. B.3 4. It is obvious f rom Fig. B.14 that the MARCH specific heat function is a temperature fit to atmospheric data. Consider the case of saturated steam at 6.9 mPa (1000 psia): the j steam table value for Cp is 1.19 and the MARCH value is 0.47. This value j of Cp is not only used in the calculation of the heat transfer coefficient 4 but also in the calculation of the temperature increase of the steam (Eqs. , B.2 or B 6 and B.7) . j It was concluded that a comprehensive accurate physical properties pa cka ge should be implemented in MARCH 1.1 regardless of the increase in j computational time and cost. A package that was readily available and that has been extensively tested in the ORNL Blowdown Heat Transf er Pro-i gram was chosen for the steam properties-(<1144 K) and then supplemented
- with high temperature data from Vargaf tik.1* Also, hydrogen physical 4
properties were acquired from Vargaf tik. The complete steam / hydrogen physical properties package is applicable over a tempcrature range of 273-
. 2810 K (32-4600*F) and a pressure range of ~0-22.1 mPa (~0-3208 psia).
- Implementation of the package -resulted in a-20-25% increase in MARCH com-putational time.
4 --n-. m - g -m 7 , y
l 1 l 88 Af ter implementa tion of the physical properties package, Eq. B.10 wa s replaced by the Dittus-Boelter correlation. 7 For comparison purpose s, a j set of calculations were made comparing the heat transfer coefficients , predicted via the new ply ical properties /Dittus-Boelter correlation and the old MARCH approach, For a constant mass flux, hydraulic diameter, and syst em pre ssure, a comparison of the predictions f rom saturation tempera- , t ur e to 2755 K is presented in Table B-2. For this example, within 80 K (144*F) of saturation the old March approach would underpredict the con-vective heat transfer by 20-80%, then for the next 555 K (1000*F) the ORNL prediction is ~10-20% greater than the MARCH prediction, after 1200K (1700'F) the predictions start to diverge due to dissociation of the steam (which is not considered by MARCH) . Probably the most critical predic-tions in the core are within 80 K of the saturation temperature. In MARCH 2.0 the user is given the option of using the Dittus-Boelter correlation and a new 1 TOPS routine; however, the specific heat correla-tions (fits) in MARCH 2.0 are still not adequate as is shown by the two ov erl ay s on Fig. B.14. Again, the region of the Cp curves that is badly missed is within 80 K of the saturation temperature. This region is crit-ical because (1) the heat transf er is better than that predicted by MARCH, thus resulting in lower structural temperatures; (2) steam in this region has more capacity to absorb the heat (i.e. the ga s doe s not hea t-up a s fast); and (3) in a BWR. several meters of the core will see a frothy, near saturation steam / water mixture.
- As a point of interest, the new ORNL physical properties package pre-dicts the 1/Cp curves in Fig. B.14 exactly.
i
- j B.5 Core Quench Models
.YL the MARCH 1.3 code for the water or steam / water covered region of the core, the generalized fuel-rod-node heat balance equation
- is O D, R + MELT ~ ( B^ (0*II R~W where h
B"*"* * # "I ** #*"*'** * '**"' l T, = water temperature (remaining variables are as defined in Sect. B.2.2). If an uncovered node is recovered due to water inj ection or level swell-ing, the fuel node temperature is the maximum of that calculated f rom Eq. B.11 or that obtained from quenching. The quenching rate in BOIL is cal-culated from the nimimun obtained using either a boiling heet tressfer coefficient or rates characterized by a time constant t. Three terms are evaluated: QBU=hA(TROD-TSAT)At g (B.12) ' QB2 = MC(TROD - RT )At/t , and (B.13) QB3 = p V hg g f (1 - a)At/t . (B.14)
89 where hB = boiling heat transf er coefficient, A = node heat transf er area, NC = node heat capacity, pg= water density, h = heat of vaporization, fg V g= node water volume, TROD = fuel node temperature, (from previous timestep) TR= equilibrium temperature, [obtained from Q = h(T R~ POOL with Q, = decay power in the node), At = timestep, r = time constant, and a = local void fraction. QB2 is the change in stored heat obtained in quenching to decay heat or equilibrium levels. QB3 is the heat required to vaporize the water in the coolant channel next to the f uel node. For a 1.0 minute time constant t, the QB2 and QB3 terms are generally found to control the quenching rate in BOIL. Af ter implementation of the new ORNL BWR core models (Sect. B.2) and
- the new SRV models (Sect. B.3), smaller MARCH time steps were required for numerical stability (0.075-0.15 min) primarily due to the solution of Eq.
B.4 ; al so, the process of quenching becomes phenomenologically very impor-tant since the level swells dramatically with each SRV " pop." Thus, at- ! tention was drawn to Eqs. (B.11-B.14), and the physical significance of a l 1.0 min time constant (especially with regard to a 0.2 cm thick channel) I became very que stionable. The approach outlined above for calculating the structural tempera-tures in the water or steam / water covered region of the core was aban-doned, and was replaced in the ORNL version of MARCH with a more mecha-nistic approach. These new models are basically simplified versions of Eq s. B.3-B.5, that is, the generalized structural heat balances (minus those terms which are not appropriate for a steam / water covered region) plus a simplified heat transfer switching logic to account for different heat transfer regimes. Therefore, both regions of the core are now handled by the same basic mechanistic set of equations. This change to MARCH resulted in ~5-10% increase in computational time. B.6 Insact of MARCH 1.1 Code Imorovements on Accident Events and Tiq1gg B.6.1 Introduction The approach taken in the assessment of the impact of the modeling-o change s discussed in Sects. B.2 through B.5 is a purely qualitative one. That is, comparisons of MARCH runs with and without (i.e. before and af-ter) the changes are made to determine differences in event timing and/or events. An attempt is made to determine the singular impact of each change.
90 1 B.6.2 BWR core model To determine the effect of the new core model, two recently completed , SASA studies were repeated (i.e. just the MAP.CH runs):
- 1. the SRLOCA outside containment at BFNP unit one,8 and
- 2. the LDHR at BFNP unit one.'
- For the SBLOCA sequence the time from core uncovery to vessel f ailure in the previous study w a s ~4 h. The maj or event s (in-ve s sel ) in this sequence occur much earlier with the new core model. There is a much f aster fuel temperaturc rise with fuel melting starting ~1 h_ earlier and core slumping ~2 h earlier. The bottom head of the RPV fails ~1 1/2 h earlier. The sequence acceleration was primarily caused by increased core metal / water reaction (a f actor of ~6 times greater with the new core mod-el).
In the LDER sequence study * , the time from core uncovery to ve ssel f ailure was ~5 1/4 h. The maj or events (in-vessel) occur slightly later with the new core model. The fuel started melting ~15 min later and the core slumped ~10 min later with subsequent head f ailure ~20 min later. The slight delay was primarily caused by better core steam cooling. What seems to be coming into focus is that one cannot make ge ne r al-l ized comments on the ef fect of the new core model on a particular accident
- scenario. There are j ust too many f actors (phenomena) that bear on the problem to risk intuitive comments.
B.6.3 SRV model, ohvsical oronerties/ correlation. auench model The ef fect of these model (code) changes will be illustrated by studying the effect of each change on one sequence. The sequence is a BWR TQUV scenario with CRD inj ection flow of 10.78 L/s (170 gpm) and with the reactor maintained at pressure. All camparisona (only the maximum core tenperature in each case) will be made at 160 min. All cases will use the new BWR core model. l The base ca se will be all the old MARCH 1.1 model s, At 160 min the i maximum core temperature is 977.6 K (1300'F). The core eventually recovers at ~450 min with negligible core damage. Introducing just the new SRV model yields a dramatic j ump in the f uel temperature [1560 K (2350*F)] and a change in the accident events (i.e. core melting starts at ~170 min). During the relatively quiescent periods between SRV " pops" (when the reactor is repressurizing), there is insuf fi-cient steam cooling of the core: thus the rods heatup f aster and then when an SRV actuates, the rod cladding, which is at an elevated temperature, re-act s vigorously with the steam and the steam cooling ef fect is more than of fset by the additional metal / water reaction heat input. Adding the new physical properties pactase and correlation te the new , SRV model results in a 277 K drop in the core temperature [1283 K (1850'F) at 160 min]. This result is d2e completely to the better convective heat transfer from the rods. Finally, including the new quench model results in another dramatic drop in the core temperature [977.6 K (1300*F) a t 160 min] . It would ap-pear that we returned to ground zero; however, in this case, the upper 0.3-f _ __ __ ., w
91 0.6 m (1-2 f t) of the core mel ted (and probably rubblized) bef ore the core recovered. Each of these model changes has had a dramatic ef fect on the core response. References for Annendix B
- 1. J. B. Rivard et al., " Interim Technical Assessnent of the MARCH Code," NUREG/CR-2285, SAND 81-1672.R3, November 1981.
- 2. S. R. Greene et al., "SBLOCA Outside Containment at Browns Ferry Unit One - Accident Sequence Analysis. " NURHi/CR-2672, Vol.1, ORNL/TM-8119/V1, November 1982.
- 3. R. T. Lahey, Jr. and F. J. Moody, "The Thermal-Hydraulics of a Boil-ing Water Nuclear Reactor," American Nuclear Society,1977.
- 4. R. O. Wooton and H. I. Avci, " MARCH Code De scription and User's Man-ual," Battelle Columbus Laboratories /USNRC Report /CR-1711, October 1980.
- 5. Browns Ferry FSAR.
- 6. D. H. Cook e t al . , "Lo s s o f DER Seque nce s a t B r ow ns Fe r ry Uni t One -
Accident Sequence Analysis," NUREG/CR-2973, ORNL/TM-8532, May 1983.
- 7. F. W. Oittus and L.M.K. Boelter, " Heat Transfer in Autanobile Radia-tors of the Turbular Type," University of California Publications, 1(193 0), p. 443-461.
- 8. R. H. Perry et al. , " Chemical Engineers' Handbook," McGraw-Hill Book Co., Inc. , Fourth Edition (1963); pp.10-14.
- 9. "ASME Steam Tables," Third Edition (1977), The American Society of Mechanical Engineers.
- 10. N. B. Vargaftik, " Tables on the Thermophysical Properties of Liquids and Cases: In Normal and Dissociated States," John Wiley and Sors, Inc., Second Edition.
I I l t U
)
1 l l
92 Table B-1. gFNP safety / relief valve settings No. of Set pressure valves [mPa (psia)) 4 7.721 1 0.076 (1120 1 11) 4 7.790 1 0.076 (1130 1 11) 5 7.859 1 0.076 (1140 1 11) Rated capacity at se t pres-sure = 385560 kg/h (850,000 lb/h) per valve. Table B-2. Comparison of heat transfer coefficient predictions Temperature Ratio" [K (*F)] 564.4 (556.3) 1.785 , 588.7 (600) 1.478 644.3 (700) 1.208 699.8 (800) 1.137 , 755.4 (900) 1.116 810.9 (1000) 1.115 866.5 (1100) 1.122 922.0 (1200) 1.132 977.6 (1300) 1.142 1033.2 (1400) 1.152 1088.7 (1500) 1.160 1144.3 (1600) 1.186
, 1199.8 (1700) 1.209 1255.4 (1800) 1.230 1310.9 (1900) 1.250 1366.5 (2000) 1.261 1644.3 (2500) 1.331 1922.0 (3000) 1.425 2199.8 (3500) 1.553 2477.6 (4000) 1.776 2755.4 (4500) 2.137 "Ra tio = he a t t r an s-fer coefficient predicted by the Dittus-Boelter .
correlation using the new ORNL physical properties divided by the MARG 1.1
- approach (Eq. B.9 and at-mospherio Cp).
Saturation tempera- ' tare for steam at 7.58 mPa (1100 psia). 1
- .,, .)
93 ORNL-DWG a s-4766 ETD
/
N
/
RF ACTOR CUTAWAY KEY
\ '
r [l
~ , '
7/ A. VENT AND HEAD SPRAY l f;Ill g: B. STEAM DRYER l C. STEAM OUTLET f ) ! Il 9 -l D. CORE SPR AY INLET E. STEAM SEPARATORS ' s F. FEEDWATE R INLET / G. FEEDWATER SPARGER H. LOW PRESSURE COOLANT INJECTION INLET ' M
'w i q '
J. CORE SPRAY PIPE w 3 5 i l Ij' i K. # E CORE SPRAY SPARGER D j L. TOP GUIDE g d M. JET PUMP l N. CORESHROUD l 0. P. TUEL ASSEM8 LIES
,ONTROL BLADE Hh k 9 O. CORE PLATE h
R. JET PUMF/ RECIRCULATION WATER INLET ,
'l- W gl S. RECIRCULATION WATER OUTLET N T. VESSEL SUPPORT SKIRT ' < a
" 1' '
l U. CONTROL R D DRIVES 8 ' ' l V. IN-CORE FLUX MONITOR PL g, ini' n!}h 5 k/
- -::!:!.:=:) . ...
kw & w , Q$W 4l _p - ~: u , l Fig. B.1. Reactor assembly.
r 94 ORNL-DWG 83-4767 ETD CO x , HANDLE , % - ll UPPER TIE PLATE ( m CHANNEL ]"""~" FUEL l' T' CLADDING # -v% f f, I . b. f."I ..b.q v 0 - FUEL ROD- ' E W , INTERIM - e SPACER ; . ! l l SPACER l l TYPICAL j l . a,3, i
" OF SEVEN j ,# {
FUEL # ' p CHANNEL +-i
SPACER CAPTURE /
WATER ROD f'd LOWER TIE
- j g- 3 PLATE i
,w -
...6 9s r ::
NOSE PIECE TIE PLATE CASTING II Fig. B.2. BWR fuel assembly. W
, 95 e ORNL-DWG 82-5179 ETD
< ~
em
's TOP GUIDE '
1'd 1 y f l FUEL ASSEMBLY [' l i l CONTROL ROD FUEL SUPPORT PIECE 3 CORE PLATE J_,=k . FLOW INLET INTO FUEL ... 4 e BUNDLE %
\ '
NCORE PLATE SUPPORT STRUCTURE I, N, . CONTROL ROD - GUIDE TUBE #E g REACTOR VESSEL CONTROL ROD r BLADE / f l E "v'e"Su"S%N#
/
/ CONTROL ROD VELOCITY LIMITER STUB TUBE '
-mr-Q Fig. B.3, BWR fuel assembly support.
I e
a D a > , T E 060OO000 eeegeege eeegeege 000OO000 8 . 6 } 7 ) 4 3 000OO000 I I eeegeege 8 G W 000,OO000 I I eeegeege D 0,0 0,e O 0 0 0 ( eeegee@6 050~OO000 L N eeegeeg@ R O 050OO eee@ee@e e O O 0~O Q.0 0 0000 eee@eege c
' g' l;
-e*fs
- w a
O o_00O000 v*@ 0_ ~0 00O000
= y 00 00O000 . O. . . d o
00.0eO000 O . r 0000e000 00 00O000 O=.~ C A". g O = l t o r n 00 00O000 C
~ o c
R E < O0 00O000 ' <
- y' R E
R W . L L B L " L O ' E O R
- L R R .
E " D E E D N T D 4 I
- A I I
U
- M U B G ^ G EH t L G R " N S R E E I
LA Y E g L P PE T i D P " UL I C W F N U ?s OE O O A Y CR L L H E
- s. , ' * ' , ,Ef'!
, . , , i 'L lg ri V
i xg ;: '.}
; j,li. i
)y!,s,,,, * * , gI. j. -
4, ' * ? ' N E T E K C
^ D O
' A S
" L G
" B .
NI L P U O C
)
97 ORNL-DWG 83-4769 ETD ORAD OMELT hA(Ta YbU5 e OD,n i / .:::;::: STEAM / HYDROGEN 1 I $ Ouw FUEL ROD Fig. B.S. Simplified MARG 1.1 nodal heat transfer schematic. 1 .k
~
i I D l
l ' ORNL-DWG 83-4770 ETD I C8 OMELTsox s O RAD ORAD k QMELTROD N N OCONV CONV OCONV 5. COR E.li + ORAD AD \ g lNTEUSTlY'l L l
~. STEAM /:i:
HYDROGEN' :. SfS'dM'/iii. . . . . . OD.R oo
... .H. .Y. .D. R.O. G. .E N q
ORAD O RAD STRUC/ STRUC/ WATER WATER s .
~
ce s h L
- Ow O MWsox Quw80*
M noo ORAD STRUC/ WATER FUEL ROD CHANNEL BOX CONTROL BLADE Fig. B.6. Simplified nodal heat transfer schematic for BWR core structures. a e e e W 4 4
3- + PRIMARY SYSTEM PRESSURE (PA) *1(f 68.9 731 772 81.4 85.5 89.6 i i i e i i PRIMARY SYSTEM PRESSURE (PSIA)
- I 1G00.0 1060.0 - 112 0.0 1180.0 1240.0 1300 0 3 $-
v , . U_ F $ = d a . mm m g.. 8 kI* " d EE i 3 E E k, S E- am
- g= ss ;
R 58 5
~
a 25 o,
$. z d
g o g O L o O I 4 STEAM LEAKAGE RATE FROM PRIMARY SYS. (KG/S)
,00 196 6 393.1 589.7 7862 9828 STEAM LEAKAGE RATE FROM PRIMARY SYS. (LB/M$l0' 00 2 52 78 10.4 13.0 3 $
? i . Ni 8~' .ao o - E 5 l- ~ l o $ e Z x = 5,
>* E d- l-c B n g 9 E 2 . = ; o .
n E ? g- 5 A o 1 n E o 66
COLLAPSED WATER LEVEL (M. ABOVE VESSEL ZERO)
,1851 4 s72 7483 10,414 issas is,2se COLLAPSED WATER LEVEL (IN. ABOVE VESSEL ZERO)
. eso 180 0 asso uoo saso e4oo y E $ !T l#= w o im
;t !B \
.'e u g-is j: 1%
1-o j l l m [ d l a n m grj _ I n o I I m so
" 1. I i g2 gj
- m I "; g g5
% E I I o# ;:;A e cH y. I i == e5m M l l <2 o A5 't
- Q l I 25 9E" I i
. o y l I B2 s< e !
r
.- g- 1 I ze i o l I l
8
* {.
I : I: l
$ 5o o
e C'P PRIMARY SYSTEM PRESSURE (PA) *1(f go ,eae 7tv as r72 ao.o a2.7 m jE5 _,y toooo PRIMARY SYSTEM PRESSURE (PSIA) io.co io.oo moo unoo 12ooo
~ 3_ em nw D*
- s mm g EE.
no oe h-nw u na D D h d 94 E H-
*e
~n i"
o 8.D.*u E o na Sk T h-
- 2: _
E* - 4 U$ i o k 5 r
-2
. : 8 m =
E t
&E sa j -
5 u l O 001
i l Fg STEAM LEAKAGE Pao =48 RATE FROM PRIMARY SYS. (KG/S) sam 28t44 so m m% a.tae
- e. m m STEAM LEAKAGE RATE FROM PRIM /.RY SYS. (LB/$(f a$h '
,00 a0 1&O 24n s2.0 40.0
- w
~
00 i 1v b= 4
. * > = :
#% d !
r E-a i! 55
*r !5 i z :
*** G O
d' 5 a< l-5 6 E" "i 1 oo =
- L m i 5 Es k' 3 om .
*$ ,aseeo COLLAPSED asees WATER asets LEVEL (M. ABOVE VESSEL ZERO) - s2es4 e4022 esoeo MNm *Q ason COLLAPSED u.mn WATER LEVEL (IN. ABOVE VESSEL ZERO) su. o wn0 soon q,. , se.a0 .
G? e O o, E-en a h >=
-=- d
. > m g D*
. .~
O Qs
$N m o K
M N Z '
$E k' -
"a m 9=
ea 9 z gE ! i
. - e EU e
$ g -
p o N TOI
102 9 ORNL-DWG 83-4177 ETD 3- a g s 6 E E N N N el- e,r B B h- < k-
< g fl 5 $ [
ds da lil
$ 3- R-
' 0jf l
d d g db > 2 x gh g!- $ k- .. il a O d o 5 5 880 1$92 1$E4 2$94 2$94 340.0 TIME - MINUTES Fig. B.13. Swollen reactor water level for TQW sample run with constant CRD inj ection of 12.6 L/s (200 gpm). 4
1 , v v .. E-103 e If E
- c:g .W7rs;4rpar14 4 up 40g ; cMARCH 2.0 (100 psia) ..gJ 7p/l M. ARCH r
4, 2.2
- -xmmcwgft a pr j; yx .1.1 7:
, 2.i L-_ito.- -t g ~2r ',2Nb h
%, q x:r ~
~
"x .
~
D . ... W: ~~"
,r,,,-11 E V,;'.-391-0,3M:M2, J. .d ittTM96 m
- . _._ _7y7__ _ _ -
2.0 z *n : xx 3
---ge ;;
X / spM 7.* d p. -. K
-rygg- ..,,.o" t;%sMf% ase
..ew::M%D"4W p
' l'9
- d. .. .._ _ __r M :. X ...X 14 .14!, .
', L gtMt WA 7-- *-. -
- r~t
/ ;L :: C r y --s.- yM s/ tr,!. "r.t%'-w+4 4
~ '
18 --: :- y T * "ir ~
~ ~
p:g
;g
;-) n ;p :- q" ;; y ~
,, .g 3
~ 3 g,7 - q- g/ t_;:Q
~ -
vtT-},z 1;
-4 M;< ;--
4 'l r::I F- ..V
- :W O
d
/ :
. y,:!"[2 : i T7 :J:
M .i' 5 MARCH 2.0~~?~ M. lbf h; h/ p h:M[ .- - s b l.6
.o T
yj 44' H1 g.500 psia l4 Ma P g gl =
/._, y j g.Mt @l ,,
_ 4 3 y lil. jp" i - 7
; W -
r: # f;T~~' n&d 1.5 m 4 / /r:.cTc
..M'le t; 44. II,&
~
y1 c4.$p,, J t 4 . JP _ _ p t ~ dEWZ, dig / 4r 4' 'b
,, 3 L
~y: 1 _ -"1MI: ,2 1'1 $'~yb' -~@=g d !*
~
3 ~ 1 .E I i}. Nf U M1 f[zn'_j b2 : 3.
- d[2 l [ {.I'.M=cy
- i,3
.j E
m:
.4_
t t! '
. 4.*,
".[:
4 A7 N 1.*"yn t +
- 2. g
= :-. 2t ,
-r :. }.r__
b
;b'2
-: rt_.. c dd.gr.ntE yf; :
F t
, y{ .
$g 1.2
.y- -
.L .T g; pg,; 4 7 =gg 77 :
. j*gj7;)
7
- t rt - FOR WATER ~
j " * *
._[
!.
- t. lff "
O i.I a
+
m Bess
. flytt gp4 r --
~r
- i. N 'N"..
.p Q.'.
D
- u. %m. J u m .a4 n l,a. }< M!' u. f ., t{h-- e:m. ;
c+M'.
- 4. .. 4' t .dif e. c ,i m.4 %g a..J, tr.c g 1.0
-9 rs r - %-
4+ r1 -
- M e w l .:. .. , ' ; 7 ae' y c.y..a 'y;.-' ' '
V- , - .
' - -.j yu o a
{@Rarebg id 1 :, a'2
}= M
- n. og r s j
"r 1- . _ .
{ 3. ?'n- 3:~1 h 4 08
.I g Eu ,j 5:- l! l '.
;*rv M U;E:;d .
2.. . w h $58 43 h . .,d [h .
":q ;4 l *: l.22 L ... #
o o i l._t h b:+pi {t d .:p,4 .6.' R,$2 M@ ,%, . . Wm 7.. w g- ^ c.. ; 7..e
~
E NkNM 2 ; o,7 ', j._ " p [ ; r. -
'~ --
t g I, ;tu .. 1] *::: .a ;; ;:1 ' M:N,0 %u.v yt ;1'&m *;M,r c4 :: g
;:. T :.
_p 9 :h, 0.6 .n ,*aq w i , _ . 4 .. .._
. Iq 0c p. ; i, ,s. . m; .,5 g
, . Q,__d .i. :._
a- ,f 1 > - . _. g 'pr M. 2 ,E t.n ' y :'t 1 O._P 1
"*l'" .# , t".
M,.,
.}.. " . &
1 . . 1 : 0.5 m . = , : .h- _;r _n 3
.. .g mIh.
7-d.Mn-{t
+---
r ::m;' +,_, . ln jw g 1 +-y ;}g .d
, ,1,1, . v.g, .. i:. o.
r _ ,- w g
._n k
m
. ',- YZ T
J.} U
... l Hn"; ;W gM .I ,
.: 3j e
.p{: h- t_ n -.q!.4 p4 n..,:E :m ;., g M-/jpg g w,/;T f. : PAGE 293 'AS Q3 J: : +--
3 pc- 7 .a
+
-:9 gg _.
g p p --RECIPROCAL CO 0.2
. t -uh 1 ;(( ;;r .
, ~; r 4 p.h yp er .Qpfi ]y[j4 P el fr 7 g- .qn .. ..w 1y 7 L ;.:1. .n .pg gJ4 1 g :r;:
- 4 q _* x _;_ -.- no _..h.. . , 4..p, a t ;_ k- o.f W-: 22 ' ' I C _4n:2o; 31 " ? h, d. . i . d
' F gghqc,h:po,HW.p, 4 ="Eg
~
- T c .J._
y } qpp-1r-- 7 nn-qrx [: ].rn {d._ c :.g g .g;g M xg y,j[ qn 7 _g
, _ :_d j < -
O.0 _ 32 100 200 300 400 500 600 700 800 900 ' 1000 t ,, TEMPER ATURE. F f g Fig. B.14. a Comparison of EARG computed Cp and the steam E h M u
.-w
O R N L- DWG 83-4778 ETD d IPERATURE FIT FOR CP ;;;;:n- -
,q..
'f. Q ~*1 } ..
.roR STd b bJ. ~""
- ). _ -
-__:{
t
'- "~ ~
$kg. *Ig--- 6{., - -
" ' ~
---4::::!:m.
-T; g:
flg$
$i n. :@agg..
+-.- :w- ..
--% ]
m 4 72: % d.-- _ . L. ...
- :TTmm.
J. .
, 2.,.:. 9e:. f. .T'";.4 + in @:t M-
~
,r',- **
T {f H{tt tii-
- t
;_ -? E'j: ?:iRhTilj 7
.1.
; af- ! .
g)1
..x ;--r -:::*-. '31t j:
$y w-7 ] t=$-f
~
MC4W 1 ;-M. :ttEWM Aft . ';$-~fi-#M
~T;i4-f. ;- f "'-
.p:y ..y . .]It ,-y
~
g :. . 2 ; p y; _g. 3 3;db q-p if[;d:hN ..: r-7'ry'. . Ef
/
k.;
-p a
.I f )f}/IQ ,d
,. Try
??p3iT ib{47n r g f %p .} .
a + :. . t
- 1 yyx . .._):h,1' ~
c F d ..: .t' /y n n - i ti. . .; 2
- r;:
in t, p' u.. f n' l.. .a{b.- { .l . 44 {tf. ~Jj ; j-.t jf. g' : [,-
' '4 +
.. ;; $ .2 :t -.. - .
4t.,. y...-..-.
, - . ~ .-
7 ,7
;+
1.; !. . g, d,-
- ; j.-
{!!! t..p . ; .; y y.g. _ _p- --- t *t 11 "+;::t t. .:: t .T: ^4-- [ ..Ik.: + y -- ' L .f ' t r}* +
;1'
._. 14 +j l-.
-+
7 . . h ._.. .
-g y,r
~
! }:p-- :p :63h. ._ --E$
STEAM TABLES,' THIRD EDITION
; TANT-PRESSURE SPECIFIC HEAT h
~
- M
/cp, LBM F/ BTU
- .E STEAM AND WATER -:
ttt .ty?t -
- t:T. + &~~
- t *h,Ly..- !.-. };i
. - 1 T: --
.:44tC--
- l100 1200 1300 1400 1500
+
able data.
t 105 Appendix C: ACRONYMS AND SYMBOLS ADS Autamatic Depressurization System ANS American Nuclear Society ANSI American National Standards Institute BAF Bottom of Active Fuel BCL Battelle Columbus Laborator?: 2 BFNP Browns Ferry Nuclear Plant BFNP#1 Browns Ferry Nuclear Plant Unit One BTR Boiling Water Reactor CBP Condensate Booster Pump CILRT Cont ainment Integrated Leak Rate Test ., CP Condensate Pump CRD Control kod Drive CRDHS Control Rod Drive Hydraulic System
; CS Core Spray System CST Condensate Storage Tank DF Decontamination Factor
. DHR Decay Heat Removal DW Drywell ECCS Emergency Core Cooling System EECW Emergency Equipment Cooling Water
. EPA Electrical Penetration Assembly EPG Emergency Procedure Guideline E0I Emergency Operating Instruction , EPRI Electric Power Research Institute FSAR Final Safety Analysis Report GPM Gallons Per Minute HCU Hydraulic Control Unit HPCI High Pressure Coolant Inj ection ID Internal Diameter EMEL Idaho National Engineering Laboratory D4TER Core-concrete interaction subroutine of the MARCH code
.! LREP Interim Reliability Evaluation Program l kPA Kilopa scal l LACP Loss of AC Power i LDER Loss of Decay Heat Removal LPCI Low Pressure Coolant Inj ection Mode of the RHR System :
; LPECCS Low Pressure Emergency Core Cooling Systems I LOCA Loss of Coolant Accident '
l LOCA/OC Loss of Coolant Accident Outside Containment LOSP Loss of Offsite Power ! MARCH Meltdown Accident Response Characteristics MPa Megapascal
; MSIV Main Steam Isolation Valve 2
Mrd/te kbgawatt Day per Tonne
- MW(e) Megawatt electrical
~
i MW(t) Megawatt thermal
. NPSH Net Positive Suction Head l NRC Nuclear Regulatory Commission OI Operating Instruction
1 06 ORNL Oak Ridge National Laboratory Pa Pascal PCV Pressure Control Valve PCIS Primary Containment and Reactor Vessel Isolation Control
- Sy st em FCS Power Conversion Syst em PRA Probabilistic Risk Assessment PSP Pressure Suppression Pool PV Pressure Vessel PWR Pressurized Water Reactor RBCCW Reactor Building Closed Cooling Water RCIC Reactor Core Isolation Cooling System RES Office of Nuclear Regulatory Research RHR Residual Heat Removal System RHRSW Residual Heat Removal Service Water RPS Reactor Protection System RPV Reactor Pressure Vessel RWCU Reactor Water Cleanup System SASA Severe Accident Sequence Analysis SBGIS Standby Gas Treatment System SGT Standby Gas Treatment System SBLOCA Small Break Loss of Coolant Accident SDV Scram Discharge Volume Int erna tional System of Unit s (Systeme Interna tional)
~
SI , SLC Standby Liquid Control SNL Sandia National Laboratories SORV Stuck Open Safety Relief Valve , SRV Safety Relief Valve TAF Top of Active Fuel TIP Traveling Incore Probe TOUV Transient event initiated by reactor scram and f ailure of normal feedwater system to provide core make-up water, accom-panied by failure of HPCI and RCIC, and by .f ailure of low pressure ECCS TVA Tennessee Valley Authority VWI Vessel Water Inj ection WW Wetwell Zr Zirconium i a
_ _~_ _ _ 107 NURIU/ CR-3179 ORNL/TM-8635 Dist. Ca tegory RX, IS - Internal Distribution
?
- 1. S. J. Ball 17. F. R. Mynatt
- 2. T. E. Cole 18. S. J. Nienczyk
- 3. D. H. Cook 19-20. L. J. Ott
- 4. W. B. Cottrell 21. R. S. Stone
- 5. G. F. Flanagan 22. B. E. Tr ammel l 6-7. S. R. Greene 23. R. P. Wichner
- 8. D. Griffith 24. A. L. Wri gh t 9-10. R. M. Harrington 25. Patent Office 11-12. S. A. Hodge 26. Central Research Library
- 13. J. E. Jone s Jr. 27. Document Reference Section
- 14. T. S. Kress 28-29. Laboratory Records Department
- 15. R. A. Lor e nz 30. Laboratory Records (RC) i 16. A. L. Lotts l
External Distribution
! . 31-32. Director, Division of Accident Evaluation, Nuclear Regulatory Commission, Wa sh ingt on, DC 20555 33-34. Chief, Containment Systems Research Branch, Nuclear Regulatory -e Commis sion, Wa sh ingt on, DC 20555
- 35. Office of Assistant Manager for Energy Research and Development, DOE, ORO, Oak Ridge, TN 37830 36-40. Director, Reactor Safety Research Coordination Office, DOE, Wa shingt on, DC 20555 41-42. L. D. Proctor, Tennessee Valley Authority, W10D199 C-K, 400 West Summit Hill, Knoxville, TN 37902
- 43. R. A. Bollinger, Tennessee Valley Authority,1530 Chestnut Street, Tower II, Cha ttanooga, TN 37401
- 44. Wang Lau, Tennessee Valley Authority, W10C126 0-K, 400 West Summit Hill, Knoxville,1N 37902 l
- 45. R. F. Christie, Tennessee Valley Authority, W10C125 0-K, 400 West Summit Hill, Knoxville, TN 37902
- 46. J. A. Raul ston, Tennessec Valley Authority, W10C126 0-K, 400 West Summit Hill, Knoxville, TN 37902
- 47. B. L. Jone s, Tennessee Valley Authority, W10A17 C-K, 400 West Summit Hill, Knoxville, TN 37902
- 48. J. T. Teng, General Electric Company,175 Curtner Avenue, San Jose, CA 95215
- 49. Z. R. Rosatoczy, Research and Standards Coordination Branch, Of-fice of Nuclear Reactor Regulation, U.S. Nuclear Regulatory Can- i mission, Washington, DC 20555 l
- 50. K. W. Hol tz claw, General Electric Campany, -175 Curtner Avenue, ;
San Jose, CA 95125 ) 51-52. Technical Information Center, DOE Oak Ridge, TN 37830 53-577. Given distriution as shown under categories RX,1S (NTIS-10)
NRC e onu 335
'" " U.S. NUCLEAi! REEULATORY COMMIS$10N NUREG/CR-3179 BIBLIOGRAPHIC DATA SHEET ORNL/TM-8635 4 TlTLE AND SUBTITLE (Adcf Volume No.,if apprQDreateJ 2. (Leave dim 41 The Effect of Small-Capacity, High-Pressure Injection Systems on TQUV Sequences at Browns Ferry Unit One 3. RECIPIENT'S ACCESSloN NO. , 7. AUTHORtSi 5. DATE REPORT COMPLETED R. M. Harrington and L. J. Ott uou rs lvE4a August 1983 , 9 PERFORMING ORGANIZATION N AME AND MAILING ADDRESS (includ'= I,o Codel DATE REPORT ISSUED Oak Ridge National Laboratory P. O. Box X e it,,v,uanas Oak Ridge, Tennessee 37830 ,,
8 (Leave Nank)
- 12. SPONSORING ORGANIZATION NAME AND MAILING ADDRESS (/nclude 2,p Codel Division of Accident Evaluation Office of Nuclear Regulatory Research 11. FIN NO.
U.S. Nuclear Regulatory Commission B0452 Washington, D. C. 20555 13 TYPE OF REPORT PE RioD COVE RE o (Inclus,ve deres/ l Topical l 15 SUPPLEMENT AHy NOTES 14 (t ear, orm A s 16 ABSTR ACT (200 words or lessi The TQUV [ Transient-induced scram, Power Conversion System Unavailable, High Pressure (HPCI, RCIC) and Low Prescure (RHR, Core Spray) injection systems unavailable] accident sequence has been shown to be among the dominant accident sequences in almost all BWR Probabilistic Risk Assessments (PRAs). This report provides an analysis of the ability of the Control Rod Drive (CRD) hydraulic system and the Standby Liquid Control (SLC) system, utilized singly or in combination, to adequately cool the reactor core during this accident sequence. The class of shutdowns considered consists of those initiated by a transient or a loss of off-site power in which there is a successful reactor scram and there is no pipe break. Various combinations of flow delivery capability and operator actions are considered and an optimum procedure is recommended. A special version of the MARCH c.omputer code is used to calculate core temperatures and estimate fuel damage for cases involving partial or full core uncovery. The results show that fuel damage can be avoided by appropriate operator action. i 17 AE Y WoROS AND DOCUMENT AN ALYSIS 17a DE SC RIP T ORS BWR Severe Accident Analyses ( i a 17t> IDENTIFIE RS ' OPE N ENDE D TE RMS 18 AV AILABILITY ST ATEMENT 19 SECURITY CLASS (Th,s reporr/ 21 NO OF PAGES Unlimited Unclassified 20 SECURITY CL ASS (Th,s paget 22 PRICE Unclassified S Nac eonu sas m m
I 1 4 1 1AN1HX11S 120555078877 US NRC ADM-DIV POLICY OF TIDC& PUB MGT BR-PDR NUREG W-501 DC 20S55 W ASlilNGTO N s I S}}