ML20003A924
| ML20003A924 | |
| Person / Time | |
|---|---|
| Site: | Davis Besse  |
| Issue date: | 12/30/1980 |
| From: | BABCOCK & WILCOX CO. |
| To: | |
| Shared Package | |
| ML20003A917 | List: |
| References | |
| RTR-NUREG-0737, RTR-NUREG-737, TASK-2.E.4.2, TASK-2.K.3, TASK-2.K.3.07, TASK-2.K.3.30, TASK-TM 685, BAW-1628, TAC-44841, TAC-45198, TAC-45205, NUDOCS 8102100074 | |
| Download: ML20003A924 (115) | |
Text
. - _ _
ggg' Docket No. 50-346 BAW-1628 Final License No. NFF-3 December 1980 Serial No. 685 January 30, 1980 Attachment A A.
I REACTOR VESSEL BRITTLE FRACTURE ANALYSES DURING
- l SMALL BREAK LOCA EVENTS WITH EXTENDED LOSS OF FEEDWATER k Applicable to Babcock & Wilcox 177-Fuel Assembly Nuclear Steam Systems In T
L i
l*
i a
1 l
I
- I l
l l
BABCOCK & WILCOX Nuclear Power Group Nuclear Power Generation Division
,f P. O. Box 1260 l '
Lynchburg, Virginia 24505 Babcock & Wilcox
t't g'
Ba'bcock & Wilcox Nuclear Power Group Nuclear Power Generation Division
'ynchburg, Virginia Report BAW-1628 Final December 1980 Reactor Vessel Brittle Fracture Analyses During Small Break LOCA Events With Extended Loss of Feedwater Key Words: Brittle Fracture. Small Break, Reactor Vessel Downcomer, High Pressure Injection ABSTRACT This report has been prepared to address issues raised in a letter from D. F.
Ross of the U. S. Nuclear Regulatory Commission to J. H. Taylor of Babcock &
Wilcox. The letter, dated July 12, 1979, is entitled "Information Request on Reactor Vessel Brittle Fracture." The investigation reported herein addresses the possibility of exceeding the fracture mechanics acceptance criteria of the reactor vessel in a nuclear steam system caused by excessive cooling by high-pressure injection flow (without reactor coolant loop flow) during small breaks (or tote.1 loss of feedwater events where the operator opens the power-operated relief valve) where the reactor coolant pressure is kept relatively high owing to choked flow out the small break (or open PORV).
i I
l t
Babcock & Wilcox
i CONTENTS .
Page
- 1. INTRODUCTION AND DISCUSSION . .. . ......'. ......... 1-1 1.1. The Brittle Fracture Concern ... ... .. . . . .. ... 1-1 1.2. Investigations ....... . .. ... . .. .. ..... 1-2 1.2.1. LOCA Analyses . . . . . . . . . . . . . . . . . . . . 1-3 1.2.2. KV Downcomer Temperature Evaluation . . . ...... 1-3 1.2.3. Linear Elastic Fracture Mechanics Analyses ..... 1-3 1.2.4 Assumptions and Conservatisms . . .... .... .. 1-4
- 2. SMALL BREAK ANALYSIS .... . . . . .. ... .... ... .. 2-1 2.1. Evaluation of Worst-Case Parameters . . . .. ... .. .. 2-1 2.1.1. HPI Flow Effect . ... .. ... .. .. . .. .. 2-1 2.1.2. Break Location Effect ... .. . ... . . . ... 2-2 2.1.3. Break Size Effect . . . .. ... . ... .. ... 2-2 2.2. LOCA Analyses Without Operator Action to Throttle HPI Flow 2-3 2.2.1. 0.007-ft* Pressurizer Break . . . . . . . . . . . . 2-3 2.2.2. 0.015- and 0.023-ft2 Pressurizer Breaks . . .... 2-8 2.3. LOCA Analyses With Operator Action to Throttle HPI Flow . . 2-11 2.3.1. Assusptions Used . . . ..... . .. . ... .. 2-11 l
2.3.2. Analytical Method . . . . . . . . . . . . . . . . . 2-12 2.3.3. Results . . . . . . . . . .. ...... .. ... 2-13
- 3. REACTOR VESSEL DOWNCCEER MIXING . . . . . ... .... ...... 3-1 3.1. MIX 2 .. ........ . . ...... . ......... 3-2 3.2. Analysis Assumptions .. .. .. . .. .. .. . . . .... 3-2 3.3. Analysis Performed ... ..... ........ ..... 3-3 3.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-4 4-1 ~
- 4. BOUNDING REACTOR VESSEL C00LDOWN ANALYSES . . . . . . . . . . . .
4.1. Assumptions . . . ... . . . .... .. ...... ... 4-1 4.2. Conditions . . ..... . . ......... . ... .. 4-2 1
4.3. Analyses .. . .... . ..... ... . .. ... . . . . 4-2 l
4.4. Results . . . . . . . . . . . . . . . . . . . . ... . . . . 4-4
- 5. FRACTURE MECHANICS ANALYSES . . . . . . . . . . . . . . . . . . . 5-1 5.1. Methodology . . . . . . . . . . . . . . . . . . . . . . . . 5-2 5.1.1. Thermal Stress Intensity Factors .... .. ... 5-2 5.1.2. Pressure Stress Intensity Factors . ... ... .. 5-3
! 5.1.3. Welding Residual Stress Intensity Factors . . ... 5-3 5.1.4. Material Fracture Toughness Data . ..... . . . . 5-4 5.2. Flaw Parameter Assumptions . . .... .......... 5-5 -
l - 111 - BabcCck & Wilcox 1
l
CONTENTS (Cont'd)
Page 5.3. Results . . . . . . . ... . . . . . . . . . . . . . ... 5-5 5.3.1. Fracture Mechanics Evaluation Criteria . . . ... 5-5 5.3.2. LEFM Results . . . . . . . . . . . . . . . . ... 5-6 5.4 Applicability of Base Case .. . . . . . . . . .. . ... 5-8 5.5. Conservatisms . .. . .. . . . . . . . . . . . . . . ... 5-8
- 6.
SUMMARY
AND CONCLUSIONS . . . . . . . . . . . . . . . . . . ... 6-1 6.1. Analyses Performed . .. . . . . . .. . . . . . . . ... 6-1 6.2. Ceneral Conclusions . ........... . . . . . ... 6-1 6.3. Specific Conclusions .. . . . . . . . . . . . . .. ... 6-2
- 7. REFERENCES . .... . ..... . . . . . . .. . . .. . ... 7-1 List of Tables
. Table 2-1. Vent Valve Opening Vs Resistance . . . . . . . . . . . . ... 2-13 3-1. Results of Generic Bounding Analysis With Instantaneous Cooldown in Downcomer Coolant Temperature . . . . . . . . ... 5-9 5-2. Comparison of Reactor vessel Materials . . . . . . . . . ... 5-10 List of Figures Figure 1-1. RCS Flow During Normal System Operation . . . . . . . . ... 1-6 1-2. RCS Flow During Total Loss of Feedwater Event With a Small Break . . . . . . . . . . . . . . . . . . . ... 1-7 1-3. Reactor Vessel Flow During Small Break . . . . . . . . . . . . 1-8 2-1. HPI Flow Rate Vs RCS Pressure . . . . . . . . . . . . . ... 2-14 2-2. CRAFT Noding Scheme, Eight-Node Model of RCS . . . . . . ... 2-15 2-3. Downcomer Temperature Vs Time Comparison of Multinode and Eight-Node CRAFT Models, Stuck-Open PORV, Two HPI Pumps, AFW @ 40 s . . . . . . . . . . . . . . . . . ... 2-16 2-4. Downcomer Pressure Vs Time -- Comparison of Multinode and Eight-Node CRAFT Models, Stuck-Open PORV, Two HPI Pumps. AFW @ 40 s . . . . . . . . . . . . . . . . . ... 2-17 2-5. Cold Leg Level 0.007-fg2 Pressurizar Break Without HPI Throttling, Node 3 . . . . . . . . . . . . . . . . . . . . 2-18 2-6. Hot Leg Level 0.007-ft 2 Break Without HPI Throttling, Node 4 . . . . . . . . . . . . . . . . . . ... . . 2-19 2-7. Pressurizer Level 0.007-ft 2 Pressurizer Break Without HPI Throttling Node 7 . . . . . . . . . . . . . . . . 2-20
- iv - Babcock & Wilcox
l Figures (Cont'd)
Figure Page 2-8. RV Liquid Level 0.007-ft2 Pressurizer Break Without HPI Thrcttling, Node 7 . . . . . . . . . . . . . . . . 2-21 2-9. Core Outlet Temperature Vs Time, 0.007-ft 2Pressurizer Break Without HPI Throttling, Node 2 . . . . . . . . . . . . . 2-22 2
2-10. Downcomer Temperature Vs Time, 0.007-fc Pressurizer '
Break Without HPI Throttling, Node 1 . . . . . . . . . . . . . 2-23 2-11. Hot Leg Temperature Vs Time. 0.007-ft: Pressurizer ,
Break Without HPI Throttling, Path 2 . . . . . . . . . . . *. . 2-24 2-12. Core Exit Flow Vs Time, 0.007-ft8 Pressurizer Break Without HPI Throttling, Path 2 . . . . . . . . . . . . . 2-25 2-13. Leak Path Flow Vs Time, 0.007-fc2 Pressurizer Break Without HPI Throttling, Path 7 . . . . . . . . . . . . . 2-26 2
2-14. Vent Valve Flow Vs Time 0.007-fc Pressurizer Break Without HPI Throttling, Path 8 . . . . . . . . . . . . . 2-27 2-15. HPI Flow Vs Time, 0.007-ft2 Pressurizer Break Without HPI Throttling, Path 9 . . . . . . . . . . . . . . . . 2-28 2
2-16. Core Outlet- Quality, 0.007-f t Pressurizar i Break Without HPI Throttling, Path 2 . . . . . . . . . . . . . 2-29 2-17. Leak Path Quality (PORV), 0.007-f t 2 pg,,,,,13,,
Break Without HPI Throttlingg Path 7 . . . . . . . . . . . . . 2-30 2-18. Vent Valve Quality, 0.007-ft Pressurizer i Break Without HPI Throttling, Path 8 . . . . . . . . . . . . . 2-31 2
2-19. Primary System Inventory. 0.007-ft Pressurizer Break Without HPI Throttling . . . . . . . . . . . . . . . . . 2-32 2-20. Core Pressure Vs Time, 0.007-fc2 Pressurizer Break Without HPI Throttling, Node 2 . . . . . . . . . . . . . 2-33 2-21. Downcomer Temperature Vs Time at 1500 psia, Comparison of Eight-Node CRAFT to Semi-Steady-State Analysis Method .... 2-34 2
2-22._ Downcomer Temperature Vs Time, 0.015- and 0.023-ft Pressurizer Breaks Without HPI Throttling - CRAFT Semi-Steady-State Analysis . . . . . . . . . . . . . . . . . . 2-35 2
2-23. RC Loop Flow Vs Time, 0.015- and 0.023-ft Pressurizer Breaks Without HPI Throttling, Flow Path 3 . . . . . . . . . . 2-36 2-24. Downcomer Temperature Vs Time, 0.007 , 0.015 , and 0.023-ft 2 Pressurizer Breaks With HPI Throttling at 100F Subcooled Core Outlet Assuming Moody Discharge Flow for 100F Subcooled Water ........... 2-37 l
2-25. RV Pressure Vs Time, 0.007 , 0.015 , and 0.023-ft*
Pressurizer Breaks With HPI Throttling at 100F Subcooled Core Outlet Assuming Moody
! Discharge Flow for 100F Subcooled Water . . . . . . . . . . . . 2-38' 3-1. Numerical Model of MIX 2 Analysis . .............. 3-5 3-2. Downconer Fluid Temperature, Profiles - HPI Temperature 40F, VV Temperature 540F, (MHPI/MVV) % 1.5.......... 3-6 3-3. Downcomer/ Cold Leg Velocities Without Density Effects .... 3-7 3-4. Downcomer/ Cold Leg Velocities With Density Effects . . . . . . 3-8 3-5. Locations of HPI Nozzles (One on Each Cold Leg Pipe) -
l Lowered-Loop 177-FA Plants . . . . . . . . . . . . . . . . . .- 3-9 3-6. Locations of HPI Nozzles (One on Each Cold Leg Pipe) -
Raised-Loop.177-FA Plants . . .... 3-10 Babcock & Wilcox
-v-
~ - . - . _
i Figeres (Cont'd)
Figure Page 4-1. 2 Downcomer Fluid Temperature, 0.023-ft Break . . . . . . . . . 4-5 4-2. Heat Transfer Coef ficient Vs Time . . ............ 4-6 2
4-3. Transient Wall Temperature Profiles, 0.023-ft Pressurizer Break Without HPI Ihrottling . .......... 4-7 4-4 Tangential Conduction . . .................. 4-8 5-1. Types of Weld Orientations . . . . . . ............ 5-11 5-2. Allowable Pressure Vs Downcomer Temperature . . . . . . ... 5-12 5-3. Oconee 1 Inside Surface Reactor Vessel -- Weld Locations of Interest . . . . . . . .... . . ............ 5-13 5-4 TMI-1 Inside Surface Reactor Vessel -- Weld Locations of Interest . . . . . . . ........ ..... . ... 5-14 5-5. TMI-2 Inside Surface Reactor Vessel -- Weld Locations of Interest . . . . . .. ....... ........... 5-15
! 5-6. Crystal River 3 Inside Surface Reactor Vessel -- Wald Locations of Interest . . ................. . 5-16 5-7. ANO-1 Inside Surface Reactor Vessel -- Weld Locations i of Interest . . . . . . . .... . . ............ 5-17 5-8. Rancho Seco Inside Surface Reactor Vessel -- Weld Locations of Interest . . ................ .. 5-18 5-9. Reactor Versel Nozzle Locations -- Inside Surface, Typical 177-FA Lowered-Loop Plant . . ............ 5-19 5-10. Reactor Vessel Nozzle Locations -- Inside Surf ace, Typical 177-FA Raised-Loop Plant (Davis-Besse 1) . . . . . . . . . . . 5-20
- vi - Babcock & Wilcox
l l
l i
- 1. INTRODUCTION AND DISCUSSION
?
t The investigations described in this report were performed to evaluate the con-cern of brittle fractura during recovery from a small LOCA with extended loss of foedwater in response to the NRC's information request dated July 12, 1979.2 This report describes the analyses performed and the results obtained.
1.1. The Brittle Fracture Concern The concern associated with brittle fracture during a small LOCA with the re-quired assumption of extended loss of feedwater can best be understood through the use of simple schematics of the system. Figure 1-1 shows the reactor cool-anc flows within the reactor coolant system during normal operation.
Figure 1-2 shows the system flows a few minutes af ter a small break typical of -
a stuck-open power-operated relief valve (PORV). During chis phase, the reac-
- tor coolant system (RCS) pressure and temperature are dropping and RCS flow is decreasing as a result of an operator requirement to trip the RC pumps on ESFAS actuation. The temperature in the RCS is higher than in the secondary side of the steam generator; thus, the generators assist in removing heat and circulat-ing the RCS coolant. Assuming no feedwater is available, the RCS must be cooled by injecting coolant from the high-pressure injection (HPI) system. The warm RCS loop water is well mixed with the cold high-pressure injection flow as it passes the HPI nozzle; thus, relatively warm water enters the vessel and downcomer.
If the transient proceeds unhindered, RCS temperature and pressure continue to fall and eteam voids will form in the system. The rate of temperature and pressure decrease and the volume of voids is primarily a function of the break size. At this point, natural circulation loop flow can cease. For the larger
~
small breaks assuming an extended loss of feedvater and no forced RC flow, loss of natural circulation could occur in 8 to 15 minutes and is partly the result of voids at the top of the hot leg and partly the result of loss of heat 1-1 Babcock &Wilcox
. . _ . ~ _ ._ - - _ _ . . . ~_ -_. _ __ _ . - -
removal capability by the steam generators. Loss of steam generator heat re-moval occurs when the RCS temperature falls below the secondary system tempera-ture.
When loop flow ceases, the cold RPI water will begin to cool the cold leg and subsequently flow into the reactor vessel (RV) downcomer. There it will mix with the warm vent valve return flow as shown in Figure 1-3. The system will stay in this condition, with the downcomer fluid temperature gradually de-creasing as decay heat decreases, until reactor coolant loop flow is initi-ated. If RC pressures remain high enough and there is insufficient vent valve flow or mixing, the RV wall temperatures may decrease to the point where cracks in the RV could initiate if flaws exist in the RV metal.
Tne potential for brittle fracture of the reactor vessel is dependent upon RV material properties, flaw size from which the brittle fracture initiates, tem-perature, and stress. The main components of stress are usually the RC pres-sure and vessel thermal gradients due to cooling. Transients that exhibit high vessel stress at a low RV temperature must be evaluated to ensure that the fractura mechanics acceptance criteria will not be violated during the transient considering the vessel irradiated material properties and postulated flaw sizes.
1.2. Investigations The analyses presented in chis report can be divided into three areas:
- 1. LOCA analysis.
- 2. Reactor vessel downcomer mixing.
- 3. Linear elastic fracture mechanics.
The LOCA analyses provided all the information necessary for performing the linear elastic fracture mechanics (LEFM) analyses except the RV downcomer tem-perature next to the RV wall. The LOCA analyses determined the RV downcomer temperature assuming complete mixing of the HPI and vent valve flows entering the downcomer. The extent to which mixing occurs is uncertain; therefore, various RV downcomer mixing calculations were made using different assump-tions.
Finally, LEFM analyses were conducted for different LOCA events assuming dif-ferent amounts of downcomer mixing. The analyses performed are summarized below. No credit has been taken in these analyses for activities to upgrade 1-2 Babcock & Wilcox
r the auxiliary feedwater systems, which will make a total loss of feedwater for an extended period of time very improbable.
1.2.1. LOCA Analyses Six LOCA analyses were performed. Three breaks (0.007, 0.015, and 0.023 fez )
were analyzed assuming no feedwater to the steam generators. They were lo-cated at the top of the pressurizer and analyzed both with and without oper-ator action to throttle back the HPI flow. For the three cases with operator action, the assumed action was to reduce EPI flow when the core outlet tem-perature reached 100F subcooled. He then maintains approximately 100F sub-cooling at the core outlet. The primary purpose of the LOCA analyses was to determine the HPI flow rate, vent valve flow rate and temperature, RCS pres-2 sure, and RV downcomer temperature. The 0.007-ft pressurizer break with no l
operator action was analyzed in detail using the CRAFT computar code for 10 hours1.157407e-4 days <br />0.00278 hours <br />1.653439e-5 weeks <br />3.805e-6 months <br /> real time in response to the NRC-requested analysis. The other fi;;
LOCA analyses used the CRAFT code only during the blowdown stage of the tran-sient. After the RCS refilled with water, a steady-state analysis was per-formed to determine the reactor vessel conditions. The steady-state analyti-cal method was benchmarked against the CRAFT analysis for the 0.007-ft 2 i pressurizer break with no operator action. The LOCA analyses are described
! in detail in section 2 of this report.
1.2.2. RV Downcomer Temperature Evaluation Fracture mechanics analyses require calculation of the RV wall temperature, which in turn depends on the downcomer fluid temperature. It is expected that there will be significant flow mixing in the cold leg piping in the area of HP injection and in the downcomer, but because of the complex geometry of the downcomer region, quantifying this effect represents the principal uncer-tainty in the investigation. For this reason, very conservative bounding cal-culations were also performed. The analyses performed to evaluate potential mixing and RV cooling are discussed in sections 3 and 4..
1.2.3. Linear Elastic Fracture Mechanics Analyses Linear elastic fracture mechanics analyses were performed for each of the break cases and for a range of mixing assumptions. These analyses are described in section 5.
1-3 Babcock & Wilcox l
1.2.4. Assumptions and Conservatisms Assumptions are used throughout this report as necessary to address the items contained in the Staff's July 12, 1979, information request. Two fundamental assumptions that must be made to address the requested information are (1) an extended loss of feedwater and (2) subsequent loss of natural circulation.
Since the issuance of the Staff request, programs have been undertaken or com-pleted that signif,1cantly reduce the potential of these situations occurring.
Extensive upgrades underway to increase the reliability of the emergency feed-water systems decrease the probability of ever experiencing an extended loss of feedwater and reduce the need to cool the RCS via the high-pressure injec-tion system. The conservatisms and assumptions used throughout this analysis
're a summarized below.
- 1. All feedwater is lost for an extended period of time. No credit is taken for upgrades to the emergency feedwater system which reduce the probability of this occurrance.
- 2. For the bounding case, a hypothetical maximum HPI flow capacity is assumed over the entire RCS pressure range analyzed. No single plant can achieve this hypothetical capacity over the entire pressure range.
- 3. Core flow into the downcomer is assumed to pass through four vent valves rather than the eight valves existing on all but one plant. This reduces this amount of warm water entering the downcomer.
- 4. HPI flow is assumed to be injected directly into the downcomer, making introduction of cold water to the downcomer more sudden. In the actual t
plant configuration, the HPI flow is injected into the cold leg pipe, and a fraction of the HPI flow is directed backward toward the steam generator.
No credit is taken for mixing and heatup of the HPI flow in the existing 11 feet of cold leg piping between the point of HPI injection and the down-comer inlet.
- 5. Little or no HPI-vent valve flow mixing in the downcomer was assumed. The 7
HPI flow was assumed to enter the downcomer and essentially streamline down the RV wall with little or no mixing with the vent valve flow.
l
- 6. Linear elastic fracture mechanics (LEFM) methods were used in the brittle fracture analysis. No credit vas taken for warm prestressing. .
1-4 Babcock & Wilcox
1 i
i i
- 7. The minimum possible BWST temperature of 40F was used.
- 8. The worst combination of RV material properties and crack orientation of I all reactor vessels was assumed.
E 4
i l
)
i i
l l
4
-E i
t t
r r
. -- 1
.1-5 Babcock &Wilcox !
Figure 1-1. RCS Flow During Normal System Operation HOT LEG PIPE m
A.. ..
COLO LEG PIPE i *.
-iF t-o r PRESSURIZER i .
Ch "
l
.fJ F CORE m SURGE LINE
, __!L__
.LL j U N REACTOR VESSEL DOWNCOMER
._ HPl INJECTION N0ZILE STEAN GENERATOR i
l
? .
l l Babcock & Wilcox 1-6
Figure 1-2. RCS Flow During Total Loss of Feedwater Event With a Small Break VALVE HOT LEG PIP TO QUENCH TANK OfE
. COLD LEG PIPE *'*e c - . ..
' d n
~
' PRESSURIZER VENT VALVE g i
h O i n l.I
( 1 LYg.,,* LL g '
h J ___ 9 o]#
a l
' CORE SURGE LINE p g
'~~" ~~
~ REACTOR i
VESSEL
.__! n L 00WNCOMER l
HPI INJECTION l
i FROM BWti
' STEAM l GENERATOR 1-7 Babcock a.Wilcox
Figure 1-3. Reactor Vessel Flow During Small Break r 1 var LTT. 4 7 W GC O.
1 tb -- b" b5 k
/3
-y a s f - 2
(' i bt) (/ !
- - ~-
, _ , - r . ,
m i e
.yf, ga//, ri ;l m ***
VJ
'x w'.
' vitet v& Lyt
- 7' ;
. L'~**
, ll ]--*-+M 4' F Flow I' ' -.
): ;; '" - mar,sozar wrter voz:tc --
,a . ,o 3 s essu.E.. / _
iz t.
v.-
- N( L I y'
.g 0'
i I: T
' ' , #~
}d
,,,, , to ,
y ~ _ - 4,;
1, 4 q
r- -- c - r, e,K.f, q:;x
,- . .cro. . ...a L
q
's n
- e
,jti :) i]' / -
p -
= ce=<a
. , 3) ;. }:
- x. W. :_ g .. - .. .o ro. v. ,,s
) ;. ,.
. < , c /, . ,' ..
, ; s.
. st ,. a.
.m 4 - ., t <v
. , /
i .
l - ,, N, l
! ' 'Q: / \. ,
D l
- ke; ' /
l3j Q I K - r, i,I'fv , ; '-
l.
, ! ! .A p s
~
k' s
% ;l N N
k' 1< ,
i II'\ k '-d q N
aff h - - - s
.h kt urrr: ,. 'N; r / s)
- -,1 > 1
(' (.w;.-
l ;4 ,-
.. 33.d.,pa
't~' . ?
0 '%x ! j a
' i./ % ,
j e.7 1 3 sw _ - -s 1 z _j l l } i 1N 1
l 1-8 Babcock & Wilcox
- 2. SMALL BREAK ANALYSIS
! In order to bound the brittle fracture concern, a number of evaluations and small-break LOCA analyses were performed; these included (1) evaluating worst-case inptics for the snalyses, (2) running one break size out 10 hours1.157407e-4 days <br />0.00278 hours <br />1.653439e-5 weeks <br />3.805e-6 months <br />, (3) analyzing a spectrum of breaks using worst-case HPI flow to help define the worst-case break size, and (4) analyzing the spectrum of breaks assuming operator action to throttle HPI based on subcooling at the core outl'et.
2.1. Evaluation of Worst-Case Parameters The most limiting transient is considered to be the one that produces a sys-tem pressure coming closest to the maximum allowable pressure as determined by a linear elastic fracture mechanics analysis (LEFM). The maximum allow-able pressure is a function of three parameters - temperature, rate of tem-perature change, and material properties. Of these three parameters, only temperature and rate of temperature change are affected by the transient vari-ables of HPI Ilow, break location, and break size. Therefore, an investiga-tion was undertaken to define the worst-case parameters (HPI flow, break loca-tion, and break size) for use in ECCS/ brittle fracture analyses. The results l
of the investigations are provided below.
l 2.1.1. HPI Flow Effect HPI flow rate has a significant impact on RCS pressure and downcomer temper-
! ature. When HPI flow increases, the RV downcomer water temperature decreases and the RCS pressure increases. Both of these changes increase the potential for reactor 5 asel brittle failure. Therefore, the worst condition from the standpoint of brittle fracture mechanics is the condition . maximum HPI flow into the RCS. Except where operator action is explicitly modeled, the analyses assumed the maximum HPI system flow allowed by the piping configuration with three HPI pumps operating at pressures above %1500 psig and the flow from two -
MPI pumps and two makeup pumps (as on Davis-Besce) for RCS pressures below 2-1 Babcock & Wilcox
%1500 psig (see Figure 2-1 for the pump head curves used in the analyses).
This means that the unthrottled analysis encompasses the maximum HPI flow in-jection capability of all 177-FA plants.
2.1.2. Break Location Effect A small unmitigated LOCA with prolonged total loss of feedwater and RC pump trip will eventually result in a loss of primary loop circulation. Under this condition, mixing of cold HPI water with the water in the RV downcomer is pri-marily dependent on the capability of the vent valves to provide circulation of hot water into the downcomer.
For cold leg breaks, the hot water leaving the core flows (1) through the hot leg, steam generator, and broken cold leg to the break, and (2) through the vent valve, dows. comer, and broken cold leg to the break. The latter path has the least flow resistance and thus allows a large portion of the hot water to enter the downconer for mixing. Furthermore, the diversion of HPI water to the break reduces the total amount of HPI water entering the downcomer. For hot leg or pressurizer breaks, more HPI water is available to enter the downcomer. In addition, less vent valve flow occurs in a hot leg or pressurizer break, thus decreasing the amount of hot water available for downcomer mixing. As a re-sult, the most severe downcomer conditions will result from a break in the hot leg or pressurizer. The analyses that follow use a pressurizer break to eval-uate system conditions.
I 2.1.3. Creak Size Effect Because of the combination of parameters that influence brittle fracture sus-l ceptibility, the worst break size cannot be determined a priori. As the break size increases, the downcomer temperature decreases due to a lower system pres-sure and increased HPI flow. However, the lower system pressure tends to off-i set the ef f ect of the lower temperature. In addition, the initial cooldown rate increases as the break size increases. Because these effects tend to off-set each other, a spectrum of breaks was analyzed to show the effect of the l
break size. Three pressurizer break sizes - 0.007, 0.015, and 0.023 ft -
l l
were analyzed. The 0.007-ft 2corresponds to the PORV orifice area, the 0.023-ft 2break corresponds to that of the safety valve, and 0.015-ft is an inter-mediate-size break. The bounding break sizer were chosen by the following loeic:
2-2 Babcock a Wilcox
- 1. The 0.007-ft2 break was the smallest break size considered for this in-vestigation because the operator is instructed by procedures to open the PORV during a loss-of-feedwater event.
- 2. Break sizes larger than 0.023 ft2 result in rapid depressurizacion to pres-sures at which the LPI system provides makeup (with little or no repres-surization). The transient response of these larger breaks is similar to j that of the large break LOCA being considered under NRC Task Action Plan A-11. Therefore. 0.023-ft2 is the largest break size that was considered in this investigation.
While the discussion above indicates that break sizes can be adequately bounded, the worst break with respect to thermal shock cannot be defined a priori because of the interaction of RPi flow, pressure, and temperature on the brittle failure concern. Therefore, the PORV case was chosen for the detailed 10-hour CRAFT run since it represents the most probable event.
2.2. LOCA Analyses Without Operator Action to Throttle HPI Flow 2.2.1. 0.007-ft2 Pressurizer Break A 10-hour CRAFT analysis was conducted to determine the system response for an z
extended total-loss-of-feedwater accident.1 The break size cho en was 0.007 fe at the top of the pressurizer. This corresponds to an open PORV, which is the most likely small break to accompany a total-loss-of-feedwater event.
2.2.1.1. Model Develooment
' In response to the NRC request to provide an analysis of the thermal-mechanical conditions in the vessel for 10 hours1.157407e-4 days <br />0.00278 hours <br />1.653439e-5 weeks <br />3.805e-6 months <br />, an eight-node CRAFT model, as shown in Figure 2-2, was developed to determine the thermal-hydraulic conditions in the vessel .2 This model was compared with the 22-node model developed for the small break LOCA analysis. A comparison of the results from both models (as shown in the Figures 2-3 anc 2 4) shows that the simplified model can adequately predict thermal-hydraulic conditions in the vessel. A sudden drop in the downcomer temperature is caused by the initiation of HPI. The system noding is described below.
I 2-3 Babcock a Wilcox
, - -~ .. . -
Node 1: Reactor vessel downcomer and lower plenum.
Node 2: Reactor core and upper plenum.
Node 3: Cold legs between RC pumps and reactor vessel.
Node 4: Hot legs.
Node 5: Primary side of steam generators and cold legs between steam generators and RC pumps.
Node 6: Secondary side of steam generators.
Node 7: P'tessurizer. .
Node 8: Containment.
The CRAFT code assumes homogeneous mixing of the liquids in a node and deter-mines its thermodynamic conditions based on the thermal equilibrium between the steam and liquid phases. This assumption will result in completa mixing of the cold and hot fluids entering the downcouer region from the cold legs and vent valves, so the downcomer node temperatures calculated by CRAFT are inixed mean temperatures.
2.2.1.2. Assumptions Used for 0.007-ft2 Break The 0.007-ft pressurizer break size (PORV throat area), without operator ac-tion to throttle RPI flow, was chosen for the 10-hour CRAFT analysis since operator guidelines call for opening the FORV and HPI injection if all feed-water is lost. The following key assumptions were made in this analysis:
- 1. Reactor and RC pumps trip at time zero. Mixing of cold HPI water with the hot fluid in the RCS is minimized when no flow circulation around the primary loop, i.e. RC pump trip and loss of steam generator heat removal capability.
- 2. Loss of main and emergency feedwater is assumed to occur simultaneously at timo zero.
- 3. PORV is opened at 20 minutes by operator action.
- 4. HPI is actuated at 20 minutes -- HPI system flow without operator action is assumed as decribed in section 2.1.1. This HPI flow and the coldest BWST temperature (40F) will promote a colder downcomer temperature.
- 5. Four vent valves are modeled. The vent valve flow enhances mixing. The most severe case would be no vent valve flow since this produces lower downcomer temperatures. However, the system is self-compensating with re-spect to vent valve flow; decreasing the vent valve flow promotes a higher AP across the vent valve, which will increase flow through the valve.- Four vent valves were modeled although all plants except Davis-Bessa have eight.
- 6. A Moody discharge model with a. discharge coefficient of 0.75 was used to calculate flow through the open PORV. This value is based on the normali-zation of the Moody choked flow to the design steam. flow through the PORV.
Babcock & Wilcox
- 2-4
-r m a -e - w e-,
- 7. Initial operating power level is 102% of 2772 MWt.
- 8. Decay heat is based on 1.2 times the ANS standard.
Modeling of the piping and quench tank downstream of the PORV was treated as a part of the containment since flow choking always occurs at the PORV.
For purposes of this analysis, the HPI has been injected directly into the down-comer at the inlet elevation. HPI flow into the cold leg pipe volume (node 3) between the reactor vessel and the RC pumps will reduce the nede temperature below the lower bound of the steam table used in CRAFT. In order to avoid this difficulty, the HPI flow was injected directly into the downcomer.
The vent valve flow area is based on four vent valves in a fully open position.
As the differential pressure across the vent valve falls below 0.25 psi, the valve opening angle decreases. The flow reduction due to the partial opening of the vent valve is accomplished by increasing the flow resistance in accor-dance with the AP across the vent valve as shown in Table 2-1.
2.2.1.3. PORV Relief Line Choking Evaluation An evaluation was performed to determine whether choking flow ever occurs downstream of the PORV. If choked flow occurs in the downstream piping, mass accumulation and pressure buildup in the pipe will result. This may create an unchoking condition of the PORV. The upper and lower boundary conditions in the pressurizer were used to examine the flow characteristics in the down-stream piping to demonstrate that downstream choking will not occur. The cal-culations are provided below.
- 1. Upper Bound Condition The Moody choked flow through the PORV is calculated for an upstream condition of P = 2500 psia and h, = 731.7 Btu /lba as follows:
Mass flux G Moody
= 11,369 lbm/ft2 _,,
Throat pressure P, = 1500 psia, Exit quality x = 15%,
Flow through PORV W ~ A*c = 0.007 x 11,369 = 97.6 lbm/s.
rv Moody Babcock & Wilcox 2-5
Using the throat pressure and the exit quality, the enthalpy of the mixture is 695.4 Btu /lba. The ch6ked flow in the downstrean pipe is calculated using the PORV exit condition, i.e., P, = 1500 psia and h, = 695.4 Stu/lbm:
Mass flux G = 7543 lbm/s-ft2, Moody Throat pressure P = 885 psia, g
Flow rate W pgp, = AxGMoody
= 0.051 x 7543 = 3*.4.7 lbm/s.
- 2. Lower Bound Condition With pressurizer pressure P, = 1400 psia and enthalpy h, = 598.8 Btu /lbm, the choked flows through the PORV and the downstream pipe are determined similarly:
b* = 0.007 x 9064 - 63.4 lbm/s, porv P = 830 psia, x = 5*,
ho = 549.2 Stu/lbm.
The choked flow in the downstream pipe for P = 280 psia and h = 549.2 Beu/lbm is W
p , = 0.051 a 6383.1 = 325 lbm/s.
The Moody discharge model was used to calculate steam and saturated water flow through the PORV. The orifice equation was used to calculate subcooled water flow. The pressure in the quench tank was assumed to reach equilibrium with the containment within 20 minutes; the maximum pressure drop between the PORV and the quench tank will be 200 psi.
The calculations above indicate that the flow in the downstream pipe is always l
l greater than that through the PORV. Therefore, choked flow will occur only i
through the PORV during the transient.
- s-I l
l l
l Babcock a'#ilcox l 2-6 .
L-
i 2.2.1.4. Results 2
The following tabulation is the sequence of events assumed for the 0.007-f t open PORV case.
Time, s Sequence of events Reactor trip, RC pump trip, turbine trip, and loss of all 0.0 feedwater.
420.0 Secondary side boils dry.
RCS repressurizes and exceeds safety valve setpoint pressure 780.0 of 2515 psia, and safety valves open.
1,201.0 PORV is open and HPI is initiated (operator action).
2,300.0 Loop ficw essentially stops.
5,300.0 HP1 flow matches leak flow, and system reaches a subcooled state at approxi=ately 1500 psia.
36,000.0
- End of analysis.
Following reactor tr .', the steam generator provided suf ficient cooling and the system depressurized. The system repressurized and exceeded the safety valve setpoint pressure af ter steam generator cooling was lost at 420 seconds.
7.e loop flow continued until approximately 2300 seconds into the transient.
- f. cop flow was maintained because both the hot and cold legs were filled with
" ster during this period and the density gradient between the cold and hot legs vas enouch to maintain the loop circulation. Figures 2-5 through 2-8 show the l
f liquid levels in cold legs, hot legs, pressurizer, and reactor vessel. The l fluid temperature plots, shown in Figures 2-9 through 2-11, indicate that the The primary system reached a subcooled state at approximately 5300 seconds.
flow rates and qualities as a function of time for the core exit, PORV, vent-valve, and HPI are provided in Figures 2-12 through 2-18. Limited steam flows were observed during the early part of the transient. The water inventory in the primary system is presented in Figure 2-19, and the pressure in the core as a function of time is shown in Figure 2-20. The system pressure stabilized at 1500 psia. The vent valve flow continued for the entire 10 hours1.157407e-4 days <br />0.00278 hours <br />1.653439e-5 weeks <br />3.805e-6 months <br />.
I Babcock a.Wilcox 2-7 L.
2.2.1.5. Conservatisms
- 1. Because this analysis is generic, maximum achievable HPI flow was consid-ered. As indicated in section 2.2, this involved the combination of two different HPI systems.
- 2. No operator action was assumed to reduce the RPI flow.
- 3. HPI flow was assumed to be directly and totally injected into the down-comer. If the actual piping configurations were modeled, whereby the HPI
- is directed into the cold leg pipe, a f raction of the total HPI flow in-jected tends to flow backward through the steam generator, resulting in less HPI flow into the downcomer and a less severe temperature degradation.
- 4. Four vent valves were also used in the analysis to envelop the Davis-Besse raised-loop plant. The lowered-loop plant with eight vent valves will have vent valve flow equal to or greater than that of the raised-loop plant. One of the key factors affecting downcomer temperature is the amount of hot water flowing through the vent valves. . Greater vent valve flow results in a warmer downcomer temperature.
2.2.2. 0.015- and 0.023-ft2 Pressurizer Breaks In addition to the 0.007-ft 2 pressurizer break, 0.015- and 0.023-ft 2 breaks
- were analyzed to determine the effect of break size on the reactor vessel down-comer temperature and system pressure. These additional analyses were not as comprehensive as the 0.007-ft 2 analysis. They used simpler calculational meth-ods and only determined the RV condition.
2.2.2.1. Analvtical Method The eight-node CRAFT model described in section 3.1 was used for the initial blowdown analysis. The 0.015- and 0.023-f t 2 breaks were run for 20 and 45 minutes, respectively. The subsequent transient calculations, performed using a steady-state code, are provided below.
Under the steady-state assumption, the rate of change of mass (MRCS) "" *"~
' ergy (ERCS) in the RCS is zero:
- RCS " (1) dt de I
2-8 Babcock & Wilcox L
Then the core outlet enthalpy is calculated:
+h (2) hh"WHPI MPI where h = c re utlet enthalpy, Btu /lbm, h
bPI = enthalpy of HPI water, Btu /lbm, Wgpy = HPI flow, ihm/s, q = decay heat, Btu /s.
The vent valve flow is determined by performing an energy balance in the down-comer region:
W =WHPI c ~ bPI (3) vv h3 -h where h = cora inlet enthalpy, Btu /lba, W = vent valve flow, ibm /s.
The vent valve flow, which can also be calculated from the elevation pressure drop across the vent valve, is given by 288 x goA a? oh x P h (4)
W = = 96.26 A vv K vv K where = core outlet density, 1bm/ft3, A = vent valve flow area, ft2, K = loss coefficient, SP = elevation pressure drop, pei
= Hx(p - p }/144' e h o = core inlet density, ibm /ft3, H = elevation heat, ft.
Assuming a system pressure and a downcomer water temperature T , the vent valve flow W can be calculated using equations 3 and 4. The downcomer water tem-perature is determined by iterating on the assumed eT until equation 3 and 4 Babcock & Wilcox 2-9
I predict the same vent valve flow. A benchmark study was performed for the stuck-open PORV case (0.007-ft2 break) assuming a system pressure of 1500 psia.
The results indicate that the steady-state code predicts the downcomer tempera-ture approximately 9% above the CRAFT prediction, as shown in Figure 2-21. The 9% deviation in the downcomer temperature was used as an adjustment factor for the two breaks analyzed.
2.2.2.2. Assumptions Used for 0.015- and 0.023-ft 2Breaks The assumptions used in the CRAFT model are the same as those described in section 2.2.1 except for the following:
- 1. The break was initiated at time zero. The 0.015- and 0.023-ft: ,,,11 break transients assumed that a break occurred simul aneously with the 2
loss of feedwater. The 0.007-ft break transient assumed that the break occurred 20 minutes af ter losing feedwater ficw by opening the PORV.
transient did not assume an initiating break as do the 0.015- and 0.023-ft2 small break transients, and HPI 2
was assumed to be operator-initiated at 20 minutes for the 0.007-ft break.
- 3. A discharge coefficient of 1.0 was applied to the Moody discharge model instead of 0.75 because these are considered simple breaks that do not have the complex flow geometry of the PORV. The Moody correlation was used for both saturated and subcooled water.
The steady-state analysis was performed using a constant system pressure based on HPI flow equal to leak flow for a given break size. The steady-state system pressures were calculated as 1000 and 600 psia for the 0.015-and 0.023-ft2 breaks, respectively.
2-10 Babcock & Wilcox
2.2.2.3. Results The downcomer temperature transients predicted by the CRAFT and steady-state codes for the 0.015- and 0.023-ft2 breaks are provided in Figure 2-22. The loop circulation ceased early in the transient as shown in Figure 2-23 be-cause of the loss of steam generator cooling and the RCS voiding.
2.2.2.4. Conse rvat isms The maximum HPI flow indicated atave was used for both the CRAFT and steady-state calculations. No operator action was taken to reduce the HPI flow.
The HPI water temperature was assumed to be 40F.
2.3. LOCA Analyses With Operator Action to Throttle HPI Flow Fracture mechanics analyses performed on the data from the three breaks with-out operator action to throttle HPI flow analyzed above using the techniques of section 5 of this report produce results that do not meet the fracture me-chanics criteria af ter a couple of hours when HPI is not throttled. As a re-2 sult, the 0.007 , 0.015 , and 0.023-ft breaks were analyzed again, now assum-ing that the operator started throttling back the HPI flow race when the core outlet temperature reached 100F subcooled. (HPI flow under these conditions is l independent of the number of HPI pumps operating.) The operator then maintained the core outlet temperature at 100F subcooled for the remainder of the tran-sient. Maintaining the subcooling margin resulted in a slowly decreasing re-actor coolant pressure as the core decay heat dropped.
2.3.1. Assumptions Used The 100F subcooled conditions were used as the basis for a hand calculation to
! determine the downcomer pressure and temperature. The following assunptions were made:
l 1. The HPI water temperature is 40F.
The system is in a steady-state condition; i.e., HPI flow is equal to leak 2.
l flow.
l 3. Leak flow is based on the Moody correlation with a discharge coefficient of 1.0.
- 4. The core outlet temperature is maintained at 100F subcooled.
l
- 5. Decay heat is based on 1.2 times ANS standard.
l i
( ,s
! 2-11 Babcockat VVilcox
2.3.2. Analytical Method The results of the eight-node CRAFT analyses, as described in section 2.2.2.1, were used to determine the RCS conditions until the core outlet became 100F subcooled. Operator action to reduce HPI flow to maintain 100F subcooling is assumed at this time. The remainder of the transient conditions are calcu-lated using a steady-state analysis as described below.
In a steady-state condition, the relationship of HPI flow to leak flow is defined as Wg =W HPI "v (5) where W = leak flow, lbm/s, W = w, m s.
HPI v = specific volume of HPI, f t /lba, 3
HPI v = specific volume of core outlet water, ft /lbm.
8 Assuming a system pressure and 100F subcooling, the leak flow can be cal u-laced by the Moody correlation. Equation 6 is used to determine a HPI flow (Wgpy) required to maintain 100F subcooling. Equation 5 is used as a conver-gence criterion for determining the system pressure. If WHPI and W fail to satisfy the equation, the system pressure is readjusted until the criterion is satisfied.
q=WHPI(hh ~h HPI) . (6) where q = decay heat rate, Btu /s, g = enthalpy of core outlet water based on the 100F subcooled state, Btu /lba, h = enthalpy of HPI water, Btu /lba.
MPI W = HPI flow rate, Iba/s.
HPI The mixed downcomer temperature was calculated to be equal to the saturation temperature minus 150F. This value is based on the assumption of 100F sub-cooled at the core outlet plus SOF core AT.
Babcock & Wilcox 2-12
2.3.3. Results The downcomer temperature and pressure plots are shown in Figures 2-24 and 2-25. Fracture mechanics analyses performed on these data are discussed in section 5 of this report.
Table 2-1. Vent Valve Opening Vs Resistance AP across vent Avg opening angle, Resistance factor (*
valve. psi degrees K
> 0.25 21 4.2 0.25 > AP t 0.2 17.5 5.0 0.20 > AP 2 0.16 12.0 8.8 0.16 > AP 2 0.12 8.13 l' 6. 0 0.12 > AP 2 0.08 4.13 46.0.
=
0.08 > a? O
(*} Resistance factor, K, is defined as K = (288pgA2gp)fg2, l
i l
l h k& cu 2-13
_ _ _ ._ . _ _ . . ~. _. _ _ ,
i I
i l
\
cm e
a N
1 m
a 2
u, A
=
== .
A O
- a. -
u,
=- -
- 3 A
A
w
-w A V, u, =w A w
w
u, .a 2 = u, u,
_ a.
- w-2 a. =>
~ m o -
- a. m wa
= = -
-_ :2 - >a
- o. = <=
a-
.o. w - .a o e, w ~
.r.- w u, u, z -
m
= w o a
> - == 2
= w MA U w -
a w w >-
i 1 W, <A C3 4*
x u, = w a= >a C3 K A A 3 G3 3 = = Q.
C3 C BM e-B O o 53
_. a. a. - - -- o
- - = < w
_ a. a. o m = ww a w w z = =
1 w w a >= == Q
< z z U G3 4 o o Ca 7 -
f GW o a w N w w CE u z C.3 =
o - - a
=
3 -
8
- =
s
/
1
_ .m N
I I I I I o
. . . = = =
e2 a a a a o = <> a o u m - -
3lsd 'aJnssaJd SCH 2-14 Babcock & Wilcox
n m
Mr a.
m I
i
! u ec T O o
=a O
Of.
5 "
o m
o x
o U y y n --
Ea o h
o w *
= g_
J -i z w
@ z z
u - <
m , -
en somoounog z
<.a
.=.e 7 m m
3 -
w "
h w a
u
= a a 8 oo z o-N 1
N 8
> n -
1
.eo
=e s.
l l
apTS UUST.td ,
2o2essua9 mea 2s 3pTg Depuo39$
so2essua9 means 2-15 saw ck wilcox l
Figure 2-3. Downcomer Temperature Vs Time, Comparison of l Multinode and Eight-Node CRAFT Models, Stuck- [
Open PORV, Two HPI Pungs, AFW 4 40 s l I
575 l
570 565 -
5 560 -
SHARP DROP CAUSED BY HPI INITIATION 3
2 j 555 -
3
=
3 4Ut.TI-N00E 550 -
545 - 8 N00E 1 I I I I 540 400 500 600 700 O 100 200 300 Time, Sac l
2-16 Babcock s,Wilcox
Figure 2-4. Downcomer Pressure Vs Time - Comparison of Multi-node and Eight-Node CRAFT Models, Stuck-Open PORV, Two HPI Pumps, AFW @ 40 s 2400 2200 -
2000 -
.2 E.
.i a, 1800 -
U 2
E 1600 -
M N
5 8-N00E 1
1400 -
1200 -
Hl.TI-N00E l
1000 -
, , i '
l 100 200 300 400 500 1
! Time, sec i Babcock & Wilcox 2-17 l
a O
Q o
n C
o O
a n
.:d m f5 a o as u -s o
a - N W 3 C*
3 -
N a w ~
= Q c3 u
'A m O U A ~ a m
'v u -s v -
o
- a. o = N z u.
~
m a . -o -
-e u -
.J t =
~- a o- -s
- o Ow C
~
- w u @
eo u -
.=
0H
- v-
.J2 =
= - a u- Q
.4 3 -
o
-s .=
.e w e-U3 o
e
_ a C
I @
N G
W 3
e
- u. I I I o e o o e e o o o o e Q Q Q Q C.
a,. ~ = o 11 '13A31 D!nbl1 Babcock s.Wilcox 2-18
Figure 2-6. Hot Leg Level, 0.007-f c 2 Break Without HPI Throttling, Node 4 49.640 Hot Leg Full of Liquid 49.635
__- 49.630 -
E 3
a
'E -
Voiding in Hot Leg Occurs Because FORV y is Opened at 1200 seconds 49.625 -
49.620 -
49.615 30 35 10 15 20 25 O 5 Time, Sec x 10 3 l
l l
l l
l I
2-19 Babcock 4.Wilcox l
l
2 Figure 2-7. Pressurizer I.evel 0.007-ft Pressurizer Break Witliou t llPI TlirottIing, Node 7 40.000 -
~
35.000 -
Y E 8 g 30.000 3
25.000 20.000 ' ' ' ' ' '
E O 5000 10000 15000 20000 25000 30000 35000
?
8 Time, sec x-w N
1 1
l 1
l 2
Figure 2-8. RV Liquid Level 0.007-ft Pressurizer Break Without HPI Throttling, Node 7 j
.~-
- Reactor Vessel Full of Liquid
- 38 -
a a
O E -
- 37 2
M E
36 2 4 6 8 10 Time, nrs l
l i
1 Babcock & Wilcox 2-21 .
r e
z i
r u
s s
e r
P t
f 7
0 02 0 e eN
,do -.
m i ,
T g n .
si Vl t
v et 0 r o 0 ur 0 t h 5 aT 1 r
eI pP 1 mi t e
Tt u 0 t o 0 eh 0 l t t i 0 1
uW O
k ea r e or CB 0 0
. 0 5
9 2
e
- r u
g i - - - -
F 0 0 0 0 0 0 0 0 0 0 0 0 0 1 6 5 4 3 2
$ B:" EE"
?. 8R .w oh
Figure 2-10. Downcomer Temperature Vs Time, 0.007-ft 2 Pressurizer Break Without ilP1 Throttling, Node 1
! 700 e
600 -
500 -
Y 5 i
O 3 400 -
E E
300 -
200 -
?
W 8
x e- ' ' ' ' ' '
100
- E 0 5000 10000 15000 20000 25000 30000 35000 i
N Time, sec
l t
i 2
l Figure 2-11. Ilot Leg Temperature Vs Tiuie, 0.007-f t Pressurizer Break Without ilPI Throttling, Path 4 700 l
600 -
S
~ =
i e *~ 400 -
300 -
200 15000 20000 25000 3G000 35000 F 0 5000 10000
?
8 Time, see w
P I
i?
N
2 Figure 2-12. Core Exit Flow Vs Tine. 0.007-ft Pressurizer Break Without HPI Throttling, Path 2 3,Bx10 '
4000 - -
3000 '
1
.~
3 2000 '
)
1000 '
i i ) 1 '
N 0
0 10 20 30 40 50 60 10 Brs Time. Mi" i
i
. 2-25 BabcockL N #
,- we+ y , , - , , - - - - . - , ,-e,- - - - - -
Figure 2-13. Leak Path Flow Vs Time. 0.007-ft2 Pressuri er_
Break Without RPI Throttling, Path 7 700 --
600 --
f afety S
Valve Relief l
I o
500
> 7 l
i 2
400 y 300 -
oscillati ns caused by 200 !i Switching Between Moody l
I ! M fodelandOrificeEquation
[ 1 -
100 -
t 0
15000 20000 25000 30000 35000 0 5000 10000 Time, sec Babcock & %Vilcox 2-26
i Figure 2-14. Vent Valve Flow Vs Time, 0.007-fc# Pressurizer Break Without HPI Throttling, Path 8 1200 1000 Change in Vent Valve Resistances Input at CRAFT Restarts 800 -
O, v
) ~
600 _
=
b V 400 -
200 -
0 -
i i ,
-200 20000 25000 30000 35000 0 5000 10000 15000 Time, sec Babcock & Wilcox 2-27
Figure 2-15. IIPI Flow Vs Time, 0.007-ft2 Pressurizer Break Without IIPI Throttling, Path 9 250 k
200 -
a
) .
150 - k a
n :
100 50 -
1 0
15000 20000 25000 30000 35000 F 0 5000 10000 K
8 Time, sec x-a N
Figure 2-16. Core Outlet Quality, 0.007-ft Pressurizer Break Without HPI Throttling, Path 2
.500
.400 -
.300 -
2 2
o i
E l g- .200 -
.100 -
CORE OUTLET l
QUAll TY = 0.0 l INDICATING LIQUl0 FLOW
' ' I 0.000 O 5000 10000 15000 20000 Time, sec l
l 2-29 Babcock & Wilcox
Figure 2-17. Leak Path Quality (PORV), 0.007-fc 2 Pressurizer Break Without HPI Throttling, Path 7 1.200 1.000 -
Leak Path Flcw Changes From Steam to Water
.800 -
a .600 -
E I
.400 -
t
.200 -
0.000 O 5000 10000 15000 Time, sec
~
l l
Babcock & Wilcox 2-30
i Figure 2-18. Vent Valve Quality, 0.007-ft 2Pressurizer -
Break Without HPI Throttling, Path 8 1.000
.800 -
m E .o00 -
W 8
E I
.400 -
L Vent Valve Quality = 0.0 Indicating Liquid Flow
.200 l
0.000 ' ' !
l 30000 5000 10000 15000 20000 25000 0
Time, sec l
Babcock & Wilcox
,3
Figure 2-19. Primary System Inventory, 0.007-ft 2Pressurizer Break littliout itPI Throttling 1.23 l STEAM FORMATION s SOLl0 SYSTEM i'22 -
RESULTING FRGM OPENING 0F PORV
[
1.21 - AT 1200 SEC.
n[ 1.20 -
% 1.19 -
O g 1.18 -
2
- 1.17 -
';> % STEAM M ,% 1.16 -
FORMATION N RESULTING 1.15 : FROM DEPRESSURIZATION 1.14 -
I ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' '
l 1.13 10 2 103 104 10I Time, sec U
k- ,
9' N
2 Figure 2-20. Core Pressure Vs Time 0.007-ft Pressurizer Break Without HPI Throttling, Node 2 2600 F
2400 -
l 2200 -
2 I 2000 1 si N
m 1800 1600 -
[
1400 -
I 1200 10000 15000 20000 25000 30003 35000 0 5000 Time, see I
2-33 Babcock & Wilcox
i Figure 2-21. Downcomer Temperature Vs Time at 1500 psia, Comparison of Eight-Node CitAFT to Semi-Steady-State Analysis Method d
1
?
600 -
500 -
a w 400 -
Y w
- 5 b
e STEADY STATE 1500 PSIA E 300 -.
5 l
CRAFT 200 F
E 100 8
se
$ l i l l l l 1 I
{
- 4 5 6 7 8 9 10 0 1 2 3 Time l Ilts
l l
Figure 2-22. Downcomer Temperature Vs Time. 0.015- and 0.023-ft2 Pressurizer l
Breaks Without ilP1 Throttling - CRAFT Semi-Steady-State Analysis i
600 500 400 -
"r R
?
us
- 300 -
e 3
o 015 fl 00EAK 200 -
CH0KED FLSW OUT BREAK e 1000 / PSI A CH0KE0 FLOW 00T BREAK e 600/ PSI A g
100 -
/
.,23 n2 00 m g
g 8 i ' i ' i ' i ' i s
- 2 3 4 5 6 7 0 g 10 E
tu , hours N
a
Figure 2-23. RC Loop Flow Vs Time, 0.015- and 0.023-ft 2 Pressurizer Breaks Without HPI Throttling, Flow Path 3 5
10 104 1--
~
\ %
\.
E 10 3 \
I -
g -
'\. 015 FT 2
\
\.
g - ,
c _
\
N s 102 i !
t l
I 023 FT 10I _
l -
t l _
l 0 i e i i i t l 10 0 120 240 360 480 600 720 840 l
l Time, Sec Babcock & Wilcox 2-36 I.
l k
4 2=
~,
u N y > -
nm
.=* O u2 a~ m
- 1. o - .
m eo m t a -
u a.
- n. sm .
's N =
a U3 i - a mu -
N Q
- e. a ooO
=
'u
=
mu a
m a - r-a w =
.o =
E cm 1
~,
-@ ,- m e- m m
- O ~ m oo e e.
U u a -
e
.aa -
i o
~mm o 2 z e CW sn
- Om N e 3 =
C O. . a
. o -. ea oa o
- GM V as
- .2 ) =
i- 2 = -
m W >
g3 v.
>-m ae 5
u
=
w - *F a
w a O a.- O
=
a
=
=u u .C u d j neC u -
o 8 w "n -
t v- . a
= .
=. a. 2 a=o = 8 -
e-g -
w a - m n .= u. O w u-y v' S g e
v2 w-u = 8 o
= mn -
a =
a .x .= -
g M
= d U 2 = = w ~~
N m
_9 m " ~_* ,
o.
e N
i i k N - m -
I u I 3 i =
-e i
f# t ! I e I
o o o e .a o.
a = .
e a w R m -
- 3. dual Jaso:uso0 AH 2-37 Babcock & Wilcox
Figure 2-25. RV Pressure Vs Time, 0.001 , 0.015 , and 0.023-fc2 Pressurizer Breaks With IIPI Throttling at 100F Subcooled Core Outlet Assuming Moody 1 Discharge Flow for 100F Subcooled Water l
3000 -
9 h I t
a E
- 2000 i n ;
d, E 3o
-100* SUBC.
5 I
100*SUSC. BREAK 1000 -
100*SUBC. 7 F
W ' .015 FT 2 BREAK 2 HPI I W1 g .
0 2
.023 FT BREAK ar "
f2 HPl
> , ' ' I i , ,
g '
= 3 2 3 4 5 6 7 8 13.3 47
'h Time (hrs)
A> /4 W'
9+#>G $
-e. .. <e 1,.
sh< '4
.T.ST TARG.T (MT-3) l l
l.0 '! E8 BM
, E!NE I.l l,'2EM l.8 1.25 1.4 1.6 4 6" >
4% 4 'b II$; /
4{!)fk 43
l
'A (.
. TEST TARGET (MT-3) l'0 m nu 821 i
'd m "
7 2
E tu i.i ["2 Ele
.8 1.25 1.4 1.6
- 4 4$ 44
- $$;(b' ,,,
? A f :k?
- 3. REACTOR VESSEL DOWNCOMER MIXING The LOCA analyses described in section 2 determine mixed mean downcomer tem-perature. One node (node 1 in Figure 2-2) is used in the CRAFT model for the RV downcomer; it calculates the RV downcomer temperature assuming complete mixing of all fluids in the downconer. For the extended loss-of-feedwater transients, the fluids entering the downcomer are the cold leg loop flow (only at the very beginning of the event), the EPI flow, and the vent valve flow. After the loop flow stops, only the HPI and vent valve flows remain.
The velocities of the flows are very low (less than 1 fps in the cold leg and downcomer); consequently, complete mixing in the downcomer as determined by CRAFT may not occur.
The fracture toughness of the KV is a function of the temperature of the down-comer fluid next to the RV wall. Therefore, to derive the RV fracture tough-ness, the downcomer fluid mixing must be evaluated to determine the fluid temperature next to the RV wall.
l In addition to HPI flow mixing with the hot water coming from the vent valves, two other mechanisms are available for heating the HPI flow:
- 1. Upstream mixing in the cold leg piping.
- 2. Heating by the reactor vessel valls.
- 3. HP1 pump energy.
- 4. Heating by the cold leg piping.
These effects, however, were conservatively ignored in the analyses. The heat available from items 1 and 2 above is expected to be significant. The hotter fluid from the vent valves is expected to travel into the cold leg piping be-yond the 2 feet that were modeled. The vent valve will mix with and heat the HPI fluid in the cold leg before it enters the downcomer. This is the gravity l effect discussed in section 3.3.
The reactor vessel wall will also provide heat to the downcomer fluid. How-ever, a more important feature of this heating is the inherent tendency to 3-1 Babcock &Wilcox
reduce the wall-to-fluid heat transfer. This is because the fluid next to the vessel wall is heated up locally. The buoyancy force due to the density gradient tends to oppose the devnward flow and as a result, the velocity of the fluid near the vessel wall :ould be slowed or even reversed. This mixed convection phenomenon would tend to reduce heat transfer from the vessel wall into the fluid and maintain the vessel wall temperature higher than predicted by these calculations.
Because the vent valve flow offers the most benefit with regard to downcomer water heatup, an effort was made to analytically predict the mixing. MIX 2, a two-dimensional code, was used to model the region of the downcomer where mixing takes place.
3.1. MIX 2 The computer code used in the downcomer water / water mixing analysis is a two-dimensional thermal-hydraulic code, MIX 2. It solves'the continuity, momentum and energy equations in both space and time for single-phase, compressible water flow. The solution method employed is the implicit-continuous-Eulerian technique. The gravity effect is also included in the analysis to handle natural circulation or mixed convection problems. The inputs to MIX 2 are the system pressure, inlet flow velocities, and temperatures. The turbulence ex-change is modeled by an effective viscosity model used by Argonne National Laboratory. The outputs of MIX 2 a'a the local velocity and temperature fields of the domain being analyzed, l
3.2. Analysis Assumptions l The downcomer and part of the cold leg are modeled by a 14x12 grid system.
HPI flow is modeled by a uniform stream going from right to left, and the vent valve flow is modeled by the stream coming from the top (as shown in Figur,e l 3-1). The two streams of dif ferent temperatures are mixed in the downcomer region. It must be pointed out that MIX 2 is a two-dimensional code, so as-sumptions must be made to account for the flow distribution in the circumferen-l
! tial direction, i.e., how the EPI and downcomer flows spread out in the down-comer annulus.
An obvious assumption that c'an be made is that both the HPI water and the vent valve flow spread out quickly and distribute uniformly in the downcomer annulus.
Calculational results indicated that the fluid temperature in the downcomer 3-2 Babcock & Wilcox i
obtained from this analysis is generally quite high, and it provides an opti-mistic estimate for the vessel wall temperature. A three-dimensional analysis of downconer mixing would be expected to show results tending toward this sit-uation with a significant amount of HPI flow distribution in the circumferen-tial direction. On the other hand, we can assume that the HPI water does not spread around the annulus, while the vent valve flow distributes uniformly.
This is a much more conservative assumption, and the calculational results given herein as an example use that assumption.
3.3. Analysis Performed The 0.023-ft2suall break analysis was used as a base case for the mixing analyses because loss of circulation flow occurs faster for this break than -
for the 0.007- or 0.015-ft breaks 2
(s540 seconds). As such, it would be the controlling break for minimum cooldown times for small breaks, which require heating of the HPI flow by hotter vent valve water. The following parameters were used for the mixing analysis:
Number of cells in X-direction 14 Number of cells in Y-direction 12 6X, ft 0.139 SY, ft 0.467 Turbulent diffusivity vT , 10-" ft 2/s 1.0 System pressure, psia 1100 HPI flow velocity, fps 0.268 HPI temperature, F 40 Vent valve flow velocity, fps 0.658 Vent valve flow temperature, F 540 The results of the MIX 2 calculation are presented in Figure 3-2, in which the fluid temperature profiles in the downcomer immediately below the nozzle are given. These calculations were performed with and without accounting for gravity effects. It is also noted that because of modeling limitations, the mixing boundary condition used 40F water in the cold leg 2 feet from the down-comer. If the length of cold leg pipe from the downcomer to the EPI injection point had been modeled, additional mixing would probably occur. The locations of the HPI nozzles on the cold leg pipes and their distance from the downcomer are shown for lowered- and raised-loop plants in Figures 3-5 and 3-6, respec-tively.
3-3 Babcock & Wilcox
It can be seen that the case with the gravity effect has a more moderare tea-perature gradient in the downcomer. This is because the gravity effect tends to cause the HPI flow to settle and flow along the bottom of the cold leg, creating space for the hotter vent valve flow to come in to the upper part of the cold leg pipes. This phenomenon can be illustrated by the two schematic flow maps appearing as Figures 3-3 and 3-4. As a result of the gravity effect, considerable mixing takes place in the cold leg and the stratification effect is less pronounced.
3.4 Summary Based on the results of the mixing analysis, it is concluded that mixing will occur and that the HPI water will be heated by the hotter vent valve water.
Quantification of the mixing benefit is mort dif ficult. If the circumferential distribution assumptions of the previous section are accepted (uniform vent valve flew distribution, concentrated HPI),- then the analyses show the mixing phenomenon could provide as much as 150F of heatup based upon 540F vent valve water. In addition, the heacup could be greater than 150F if the existing 17 feet of cold leg between the HPI injection point and the downcomer were used in the analysis rather than 2 feet. However, the uncertainty in the circum-ferential distribution of flow and in the analytical predictions (both com-putational and modeling) makes the exact benefit difficult to determine. There-fore, because of these uncertainties, bounding analyses were performed. These are discussed more fully in sections 4 and 5.
3-4 Babcock & Wilcox
Figure 3-1. Nu:nerical Model of MIX 2 Analysis VENT VALVE FLOW HPl FLOW II Y II _
~
i RV INLET +
PIPE ~
e l
l a:
/
l Babcock & Wilcox
Figure 3-2. Downcomer Fluid Temperature Profiles - HPI Temperature 40F, VV Temperature 540F, (g7.g)% / 1.5 600
_ THERMAL SHIELO ,
REACTOR r 500 s VESSEL WALL l
m
=
400 l
2 E '% WITH OENSITY
$ [ EFFECTS N
.2 300
= \
C \
- N
- \
E N 200 N e
NO DENSITY EFFECTS, 100 g 0
2
.023 FT PZR BREAK NO OPERATOR ACTION 1
Babcock & Wilcox 3-6
Figure 3-3. Downcomer/ Cold Leg Velocities Without Density Effects VV FLOW _
F V l Y / .g.__ _
h %
HPI f ll / ~ ~
FLOW f y -- -- .e-- -
.e-
.o --
l l / .e- - m ll -
1 COLD LEG l t i
l I l I l f f f f DOWNCORER t t 1 1 l
l t
3-7 Babcock & Wilcox
Figure 3-4. Downcomer/ Cold Leg Velocities With Density Ef fects VV FLOW i
- +
P 4 y (
~
k D = ^
HPI FLOW
/ f -
\
/ f _
= ~
COLD LEG ,
(
i \\
( \\\ DOWNCONER I\
l i
l l
l l
l Babcock & Wilcox 3-8 w- t- g. y yp e +=m,
Figure 3-5. Locations of HPI Nozzles (One on Each Cold Leg Pipe) - Lowered-Loop 177-FA Plants
^
l 4 \
J ,- \
' \
\
HOT LEG PIPE \
's / ' ,
C0s.D LEG P I PE -
STEAM N', ,
GENERATOR ll.6' PLAN VIEW b.,,
MIGH PRES 3URE INJECTION N0ZZLE CF N0ZZLE r 3 l RC PUMP \
M m
r _
l Li.s' V '
l REACTOR l
[VE3SEL ELEVATION VIEW l
l l
l 1
l J )
l 3-9 Babcock 3.Wilcox l
t I
Figure 3-6. Locations of HPI Nozzles (One on Each Cold Leg Pipe) - Raised-Loop 177-FA Plants k / t /
-g / \ -
/ NOT LES PIPE /
/ _ ,i \
COLD LES PIPE
\_
\ %
STEAN \' - ** g GENERATOR g PLAN VIEW 3*
NIGN PRES 3URE j INJECTION N0ZZLE RC PUNP N- ,
Nw I
~ <
9
_ 2.0' REACTOR I VESSEL-ELEVATION VIEW s
\
Babcock & Wilcox 3-10
1 l
- 4. BOUNDING REACTOR VESSEL CCOLDOWN ANALYSES As previously discussed, uncertainties exist in determining the extent of water mixing in the teactor vessel cowncomer; therefore, to bound the overall concern, some conservative analyses were conducted. Four analyses assuming
. little or no HPI-vent valve flow mixing were performed to determine the RV ther=al gradients versus time. These situations could be perceived as HPI flowing into the RV downcomer and then streaming down along the RV wall with little or no mixing with vent valve flow. For these analyses complete mixing was assumed in the RV downcomer while RC loop flow existed. Once the RC loop flow stopped, the RV downcomer temperature was rapidly dropped over 40 seconds to a very low temperature and then sustained there. In one analysis, the final downcomer temperature was assumed to be 40F, which corresponds to the mini m SWST temperature. In the other three analyses, the final downcomer te=peratures were 90, 120, and 150F, which reflect various amounts of mixing.
Reactor vessel temperatures during a 0.023-ft break with no operator action were calculated. Fluid conditions were taken from CRAFT results. Wall sur-face heat transfer is obtained from the larger of (turbulent) forced and free convective heat transfer. The transient one-dimensional wall energy transfer problem was solved using the explicit Euler technique. Wall temperatures and temperature gradients are obtained for nominal conditions, for varying injec-tion temperatures, and (by separate analysis) with allowances for azimuthal conduction in the wall.
4.1. Assumptions
- 1. Fluid Heating - Af ter loop flow s:agnates, the downcomer bulk fluid tem-perature is arbitrarily set to tesperatures of 40, 90, 120, Or 150F.
- 2. Initial Temperatures - Gamma and neutron flux attenuation in the wall are used to set the initial temperature distribution in the reactor vessel.
- 3. Vessel Outer Surface - The outer surface of the RV is assumed to be perfectly insulated.
' Babcock & Wilcox 4-1
- 4. Film Heat Transfer -- Film heat transfer variations in opposing mixed con-vection are ignored. The film heat transfer coefficient (HTC) is set to the larger of the (pure, turbulent) farced and free HTCs.
4.2. Conditions
- 1. Geometry -- The vessel wall is 8.4373-inch SA-508 class 2 steel, clad on the icner face with 0.1875-inch stainless steel; the inside diameter is 170.o., inches. The inner boundary of the downcomer is the thermal shield, with a 151-inch outer diameter. The heated length is 15 feet (but is of no consequence here, without fluid heating).
- 2. Initial Conditions -- Initial conditions are those of an operating 177-FA plant at full power. Coolant flow is 131 mlbm/h at 555.4F and 2200 psia.
Flux attenuation in the vessel wall generates 24 kBtu/h-ft 3 at the inner surfacc, attenuating approximately exponentially with wall depth (with a linear attenuation coefficient of 8.4/ft). Film heat transfer is by forced convection, and the total temperature rise across the vessel wall is 17F.
4.3. Analyses l 1. Fluid Conditions -- Fluid co'nditions are used as input except that loop flow from CRAFT is added to injection flow af ter HPI initiation at 140 seconds, and HPI flow is multiplied by 4.5 to account for fluid streaming af ter loop flow stagnates at T=540 seconds '(the ratio of downcomer circum-ference to inlet nozzle diameter is 4.5).
- 2. Filn Heat Transf er -- In opposing flow, with forced convection downward along a vertica] heated wall, heat tranfer may differ from either pure forced or pure free convection. In laminar flow, opposing heat transfer is usually degraded from pure (forced or free) convection, but this in-fluence in turoulent flow is unknown. Thus, the film heat transfer herein is estimated by evaluating the pure forced convection HTC and selecting the larger of the two Figuce 4-2. The forced convection HTC 3
is as follows :
H = 0.023 K/D Re o
.a Pr3 *"3(Btu /h-ft2 ,.7) f ed
" *** K = fluid thermal conductivity,' Btu /h-ft *F, Babcock & Wilcox 4-2
D = hydraulic diameter, ft (D % 2W where W = downcomer width),
Re = Reynolds number, the ratio of inertial to viscous forces, Re = VD/v, 3 V = fluid velocity, ips, v = kinematic viscosity, ft 2 /s,
?r = Prandtl number, the ratio of storage to conduction energy transfer, Pr = C u/K, Cp = specific heat, Btu /lba *F, u = dynamic viscosity, Ibm /h-ft.
Because R forced is proportional to V" 5, it decreases abruptly as loop flow sesgnates (Figure 4-2).
The f ree (or natural) convective HTC is" H,ree = 0.094 K/L (Gr Pr)l!3
" "#* L = heated length, ft, Gr = Grashof number, the ratio of buoyant to inertial force, Gr = gSATLs fy2, g = gravitational acceleration, 32 ft/sz ,
S = fluid thermal expansivity, 3 = 1/o 3c/3T .(1/*F),
ST = governing temperature dif ference, wall to fluid , F.
Notice that (turbulent) H g is apparently indepencent of heated length.
Also, as H decreases, the wall-to-fluid temperature difference in-forced creases, as does H g (Figure 4-2).
- 3. Vessel Wall Heat Transfer - The energy equation in the wall:
a2 odp y=K 3
+ ,
with T = T(r,t) and boundary conditions (inner surface I ST i convection), K p (r = rg ) = H(T(r = r ) - Tbu',.], and (insulated outer surface) (r = r 9 )~= 0, is solved by discratization in space and application of the Euler explicit nethod to solve t'e approximate system of ordinary differential equations.
Temperature-dependent properties are employed, s,s as are the HTC model-ing techniques previously described.
Babcock s.Wilcox 4,3
The solution of the energy equation is verified by comparing it to the slab approximation (valid for large inner radius):
- p x The adequacy of the temporal incrementation is verified by doubling the number of time steps and comparing the standard and refined solutions.
- Spatial and temporal increments of or = 1/8 inch and AT = 1/1C second are used.
4.4. Results The resultant film HTC changes from forced to free at 7 minutes into the transient (Figure 4-2). Thus, free convection governs the remainder of the transient and ranges from h = 200 Stu/h-ft2 ,.? at t=10 minutes to h = 100 at t=1 hour, decreasing with wall surface temperature. Wall temperature pro-files respond to downcomer temperatures (Figure 4-3); as loop flow stagnates, the downcomer temperature approaches the injection temperature, and wall sur-face heat transfer increases markedly. By t=10 minutes, the inner surface temperature approaches that of the injected fluid, but the outer wall temper-atures have barely changed.
The assumed downcomer temperature was varied to assess its impact. As ex -
pected, the resultant wall temperature profiles are less sloped with rataed downcomer temperatures, especially at later transient times (Figure 4-3).
Tangential conduction was investigated using FELCON, a transient, two-dimen-sional, finite element conduction code.7 The wall temperature response with-out targential conduction was obtained by setting the entire inner surface to 40F at t > 0. Tangential conduction effects were then introduced by setting 2 feet of the inner surface to 40F, while retaining the remaining 8 feet at 550F, and extracting the wall temperature profiles in the cooled region.
As with increased HPI temperatures, tangential conduction decreases the wall temperature gradient, particularly at later transient times (Figure 4-4).
As discussed earlier, these results are very conservative and present a hounding esse to the thermal shock question. The results of the fracture me-chanics analysis for these cases are discussad in section_5.
Babcock & Wilcox
i Figure 4-1. Downcomer Fluid Temperature.
0.023-ft2 Break 600.-
CORE OUTLET w_
500 -
l I
l i
400 -
1 t DOWNCOMER
", ! (Complete Nixing) ,
e 1 l
! " 300 -
1 e
- 1 I
200 -
l 1
! l i DOWNCOMER (No Mixing, 40F HPI) 100 -
TERMINATION OF LOOP FLOW I f f g 9 l 0 200 400 600 800 1000 1200 Time, see Babcock & Wilcox 4-5 l
i
Figure 4-2. Heat Transfer Coefficient Vs Time WALL CONVECTIVE HEAT TRANSFER COEFFICIENT (n)
VERSUS TRANSIENT TINE.
- 0. 8 Pr0 .4 nforced = 0.023 Ky Reo nfrea = 0.094 t (Grt Pr)1/3 023 FT 2 PZR BREAK NO OPERATOR ACTION n
2.". 2000 -
R 1000 - FORCE 0 5
,f 500 -
.S a -
0 200 _
,\
- FREE ,/
5 "~ \
100 -
" ~ ~~ g FREE 3
3 50 -
b
- FORCED 0
20 -
5 :: j c.s P e g g 2 5 10 20 50 100 (5 Seconds) 1 Transient Time, Min 4-6 Babcock s.Wilcox
Figure 4-3. Transient Wall Temperature Profiles, 0.023-ft 2 Pressurizer Break Without HPI Throttling FOR 40 (- ), 90 ( -)
12 0 (-- - - -), t.N O 15 0 (-- -- - ) Transient Time, Downcomer Fluid Temperature, F minutes 0 5 l ~
l 550 -
f l
/
d A 500 -
i
/
/ r j l j /
' /
400 -
/
~
i ll l
/
/7 f
/9;V ,-
350 - !
/ / /
15
/-M s~
/ '/l! ll'f/%'
l /
/
300 -
ll '
o , '
/40 Downcomer Fluid Temperature, F 250 -
/
/ /
/
T_ , , , , _
0 1 2 3 4 5 6 7 8 Wall Position, in-hes I
Babcock & Wilcox 4,7
F1gure 4-4. Tange3tial Conduction TRANSIENT WALL 'EMPERATURE ;
PROFILES WITHOUT ( )
AND WITH (----) TANGENTI AL CONOUCTION Transient Time, Minutes 2
550 - .023 FT BREAK NO OPERATOR ACTION 10 500 -
20
/
s'
/
u.
$ 400 -
, z 2
u -
^
5 /
/ 30
/
/
/ -
300 -- / / j #04 _
/
/
/
/ /f -
/ / 50
/ /
/ / ~
[ / /
/ 60 200 - / / / 60 g: ,
/ , /, , , /, ,
1 2 3 4 5 6 7 Wall Position, feet l
l t
I 4-8 Babcock 3. Wilcox l
. . . _ . . _ . . . . . . . _ . . . . . . _ _ , . . _ . . -.- _ _ . . _ _ ~ _ - . _ _ . . _ _ _ _
- 5. FRACTURE MECHANICS ANALYSES The specific transients discussed here represene two characteristically dif-ferent fracture situations requiring fracture mechanics analyses to quantify the margins of safety ag:tinst vessel failure.
The first type of trans1(nt has a relatively slow cooldown rate, so that RV t thermal shock does not occur. This type includes 'he following transients:
Case 1: 0.007-ft 2pressurizer break with no HPI throttling by the operator.
Case 2: 0.007 , 0.015 , and 0.023-ft pre-2 urizer breaks with RV downcomer mixing owing to HPI and vent valve flow uniformity in the azimuthal direction with operator action to throttle back HPI flow. ;
However, for the transients abova, brittle fracture acceptance criteria may be violated after several hours if the following events occur simultaneously:
( (1) the RC water is cold, (2) the vessel is below the brittle-to-ductile tran-sition temperature, and (3) the RC pressure is high enough that the applied j stress intensity factor :xceeds the static crack initiation toughness. In this situation, cracks may propagate without arrest. These transients require the plant operator to take appropriate action to restore RC loop flow and/or .
depressurize the plant in a safe and stable manner. i The second type of transient is characterized as thermal shock of the vessel l
where the inside surface is subjected to cold RC water instantaneously. The inside surface of the vessel cools down rapidly, developing thermal gradients through the wall thickness, resulting in high thermal stresses. These tensile, thermal stresses on the inside surface combine with the tensile stresses caused' by the RC pressure and cause high applied stress intensity factors for small l flaws postulated to be present on the inside surface of the vessel. Thes transients include the hypothetical bounding case below.
l
! Case 3: 0.023-ft 2pressurizer break with no HPI throttling and rapid RV cooling.
5-1 Babcock & Wilcox
l 1
These transients must be evaluated to determine whether the temperature at the crack tip is high enough that the corresponding static crack initiation frac.ture toughness is ebove the applied stress intensity factor so that : rack initiation is not predicted. Subsequent crack ( Test and the possibility of warm prestressing are also evaluated.
5.2. Methodology The linear elastic fracture mechanics (LEFM) analytical technique has been used to evaluate both types of transients discussed above. The validity of LEFM for predicting crack initiation and arrest has been demonstrated by the thermal shock experiments conducted in the Heavy Section Steel Technology Program at the Oak Ridge National Laboratory.
5.1.1. Thermal Stress Intensity Factors The thermal stresses due to the temperature gradient through the thickness of the vessel are computed from the following general relationship:
T 2 , a r C 2 Trdr - Trdr ,
r " I '" b2 O g= "_ h [*
a Trdr +
<a Trdr - Tr2 aE 2 0 # #~
2 "1-v b2 - a2 a L ,
" *** = = coefficient of thermal expansion, E = elastic modulus, v = Poisson's ratio, G ,
Og,G = stress components, r, e, z = reactor vessel axes. ,
The temperature distribution through the wall can be assumed to be approximate-ly parobolic. The stress equations are evaluated by assuming the following general representation of the temperature distribution through the vessel.
T = A(b - r)2 5-2 Babcock & Wilcox
l T = temperature, r = vessel radial direction, A,b = constants.
The stress intensity factors are then computed by the following generalized relationship:
y= NFg2+ NF NF1+
g NF23+ 3g c = crack depth, radially, N'N,N,N3 0 1 2 = coefficients of crack opening stress polynomial, F,F,F,F 1 2 3 g = geometry magnification factors depending on crack depth.
5.1.2. Pressure Stress Intensity Factors The components of stress in the RV beltline region due to pressure are com-puted as follows:
o = Pi at r = a, a 2 pt C * '
z bi - a2 O g"b gi lY
" "#* = radial, axial, and hoop stresses, a,o,o g r, z, 6 = reactor vessel coordinats axes.
The stress intensity factors are computed from the same generalized relation-ship presented in section 5.1.1.
5.1.3. Welding Residual Stress Intensity Factors The residual stresses due to welding are computed on the basis of the evalua-tion of residual stresses in heavy weldments conducted by Ferril, Juhl, and Miller.3 The residual stress distribution through the vessel thickness is given below for the three types of weld geometry shown in Figure 5-1.
Longitudinal Welds: a(x)/S = 0.12 - 0.36 x +0.18x2 Circumferential Welds:
Case 1: a(x)/S = -0.06 + 0.18x2 (Single-V) 5-3 Babcock & Wilcox
Case 2: o(x)/S = 0.12 - 0.72 x +0.72x 2 (Double-V) where o(x) = residual stress distribution, S = yield stress, x = a/t (t = thickness), o s a s t.
The stress intensity factors are computed from the generalized relationship presented in 5.1.1.
5.1.4. Material Fracture Toughness Data The material fracture toughness data were obtained.from the reference curves of the ASME Boiler & Pressure Vessel Code,Section XI, Appendix A.' Figure A-4200-1 gives lower bound static crack initiation toughness, KIC, a% crack arrest toughness, K73, as functions of metal temperature and material refer-i ence temperature, RTg.
The material reference temperature, RTNDT, is adjusted to account for irradia-tion embrittlement effects. The amount of adjustment to be added to the bai-tial reference temperature is computed from USNRC Regulatory Guide 1.99. The adj ustment is a function of the material's we % of copper and phosporus and i the accumulated neutron fluence, n/cm2 . The accumulated neutron fluence cor-responding to 6.0 EFPY* was the basis of the analysis described herein. (As of January 1, 1980, the lead B&W plant had accumulated 4.0 EFPY. The peak l
neutron fluence for the beltline region on the vessel inner surface correspond-ing to 6.0 EFPY is 3.99 x 1018 n/cm 2 . This value is adjusted to account for specific weld locations axially and circumferential1y. The neutron fluence attenuation through the vessel thickness is also taken into account.
Taking these factors into consideration, the controlling material was found l
to be the longitudinal weld seam,.WF-70, b1 the lower'shell of the Rancho Seco vessel. The computed adjusted reference temperature on the inner surface was 246F. The properties of this material have been used as the base case for which all results have been quantified. The applicability of the results to reactor vessels other than that of the Rarcho Seco plant and with lower 1r-radiation levels is discussed in section 5.4.
- EFPY: effective full power year.
5-4 A "'
5.2. Flaw Parameter Assumptions The reactor vessels in question bve not operated long enough to have been sub-jected to an inservice inspection. Based on the shop inspections and the ASME Section XI baseline inspections, there is no evidence of flaws in any of these vessels. However, in order to perform the fracture mechanics analysis, the existence, location, orientation, and size of flaws were assumed. Surface flaws with the major axis oriented longitudinally and the minor axis oriented radially were postulated in the controlling weld metal. While the critical flaw size in the radial direction was a product of the fracture mechanics anal-ysis, the aspect ratio for the initial flaw was assumed to be 6:1 as recommended by Section III, Appendix G, of the ASME Code, and the aspect ratio for arrest i and subsequent initiations was assumed to be infinitely long. This assumption I
is consistent with the crack propagation results from the thermal snock ex-periments conducted in the HSST Program.
i l 5.3. Results l 5. 3.1. Fracture Mechanics Evaluation Criteria The transients evaluated here are considered to be accident conditions. There-fore, vessel integrity must be maintained to facilitate safe reactor shutdown.
Crack initiation can be allowed provided the cracks can be arrested. The cri-terien for precluding crack initiation is as follows:
"U ' "" * ** *1' KIT + KIP + KIW < KIC and cracks are arrested provided KIT + KIP
- KIW < Kg at flaw size a 2 where K = applied stress intensity factor due to thermals, K 7p = applied stress intensity factor due to pressure, K
77
= applied stress intensity factor due to residual stresses, K = static crack initiation toug u ss, IC Kg = crack arrest coughness.
The existence and applicability of warm prestressing is also evaluated. "Jarm prestressing exists provided crack initiation does not occur prior to or at 5-5
. _ . . . ~ _ _ . - . - . _ _ _ _ . _ _ _ _ _ __
the maximum applied load. If the load decreases continuously from the point of maximum load, subsequent predictions of crack initiation by LEFM are conservative for the following three reasons:
- 1. The introduction of compressive residual stresses (_ the crack tip due to unloading.
- 2. Work-hardening in the plastic zone around the crack tip.
- 3. Blunting of the crack tip by plastic flow.
5.3.2. LEFM Pesults Three transient cases vera analyzed for break sizes of 0.007 f t2 (stuck open FORV), 0.015 ft2, and 0.023 ft2 (stuck-open safety relief valve). These tran-sients were further analyzed for three conditions of mixing. The first case is complete, perfect mixing, which uses the downcomer reactor coolant tempera-ture transient directly from CRAFr. The second case considers complete circum-ferential distribution of the HPI and vent valve flows and predicts the mixed temperature at radial and axial locations in the downcomer. The third case assumes little or no HPI vent valve flow mixing. The vessel is subjected to a thermal shock produced by cold HPI water.
5.3.2.1. 0.007-ft2 Pressurizer Break Without HPI Throttling (Case I) _
< These transients are considered to be acceptable because crack initiation is i
not predicted before several hours into the event. The applied K will even-tually exceed K for flaws greater than 0.5 inch deep. However, a warm IC prestressing situation clearly exists. The operator should take action to depressurize the plant since the applied K exceeding KIC cann t be tolerated j indefinitely.
5.3.2.2. 0.007, 0.015, and 0.023-ft 2 Breaks With HPI Throttling and Downcomer Miring With HPI and Vent Valve Flow Uniformly in Azimuthal (9) Direction (Case II)
The temperature transient'is a little more severe than that of Case I. As in Case I, fiswa deeper than 0.5 inch in the generi~c limiting longitudinal weld would be predicted to initiate several hours into the event without operator action to throttle HPI and depressurize the system. _ The time for initiation would be sooner for Case II than for Case I. Again, a warm prestressing sit-untion clearly exists. Operator action is required to depressurize the plant, 5-6 kW Mm -
f
,.- - - - . - - , . ,,,.,+,,a ,,.-,-e. ,-se,- m w- r<w . - , , , a -,---e.-s-. . . - - , , - - v--- marm- , -.e v
l so that the actual system pressure is below the critical pressure for crack initiation.
5.3.2.3. 0.023-ft 2Break -- Without EPI Throttling --
Bounding RV Cooldown Analyses (Case TII)
In order to evaluate the required amount of water mixing in the downcomer, a series of thermal shock transients was analyzed. These temperature transients are 550-40, 550-90, 550-120, and 550-150F; the results are tabulsted in Table 5-1. Comparison of the results to the actual transient pressures reveals that the mixed tcuperatures of 120-150F in the downcomer present an acceptable situation for an extended time. The critical times shown in Table 5-1 cor-
- respond to the critical pressure at which crack initiation without arrest is predicted by LEFM provided the actual transient pressure is higher than the critical pressure. For cases where warm prestressing exists, the critical pressures and rimes are conservatively low.
i l
l The information shown in Table 5-1 has been incorporated in Figure 5-2 to de-i fine the permissible operating region for the bounding material and bounding instantaneous temperature reductions. LEFM analyses on circumferential welds, on less restrictive weld material, (i.e. . Davis-Besse) or on realistic tran-sients with non-instantaneous downcomer temperature reductions would result in expanded permissible operating regions.
The results presented in Table 5-1 have been generated specifically for tha bounding 0.023-f t2 break. Comparing the results for the 550-150F case to tha I actual transient pressure, it is seen that the critical pressure is hi her F than the actesi transient pressure throughout the first hour. The operator
! actions required for Case II with partial mixing are also applicable for this transient. For the 550-120F transient, operstor action to restore loop flow or depressurize the plant would be necessary within 40 minutes of the initia-l j tion of the event. This is because beyond 40 minutes, the critical pressure for the 0.5-inch flaw is below the actual transient pressure.
For breaks smaller than 0.023 ft2, the actual transient pressures are somewhat higher. However, it takes longer before loop flow completely stops. There-fore, the times at which the critical pressure for crack initiation without arrest exceeds the actual pressure are longer for breaks smaller than the
0.0 23-ft
break.
5-7 Babcock & Wilcox
5.4. Applicability of Base Case Because of the inherent differences between flaws oriented in longitudinal and circumferential velds, the base analysis is applicable to plants with longitudinal weld seams. The base case is very conservative for circumferen-tial welds, which have half the pressure stress of longitudinal welds. As indicated in Table 5-2, these plants are Oconee 1, IMI-1, TMI-22 Crystal River 3 Arkansas Nuclear One (ANO-1), and Rancho Seco. The potential for cold water at the veld location would be most likely to exist only on plants with welds under or near cold' leg nozzles. The locations of the longitudinal welds with respect to the cold leg nozzles are shown in Figures 5-3 through 5-8. In ad-dition, the 3 ) cations and dimensions of core flood nozzles, vent valves, and hot and cold leg pipes and nozzles are provided in Figures 5-9 and 5-10. As can be seen from these figures, the only plant with longitudinal welds under or near the cold leg nozzles are Oconee 1. ANO-1, and Rancho Seco. Welds for the other plants would be subjected to substantially higher water temperatures.
Hence, the base analysis is very conservative for the other units with longi-tudinal welds.
In summary, significant variations in veld material, weld types, and irradia-tion times exist between plants, thus making the bounding analyses very con-servative for some plants.
5.5. Conservatisms It is felt that the fracture mechanics analysis described above has a number of inherent conservatisms. Without elaboration or quantification, these con-servatisms are listed below.
- 1. Flaw size.. shape, orientation, and location.
- 2. KIC
""d KIA 1 wer bound toughness curves.
- 3. Adjusted RTNDT fr m Upper bound of Regulatory Guide 1.99.
4 Applicability of LEFM to stresses abcve yield as in the case of severe thermal shock with pressure.
5-8 Babcock & Wilcox
Table 5-1. Results of Generic Bounding Analysis With Instantaneous Cooldown in Downcomer Coolant Temderature Downcomer Critical Allowable Critical Presence temp change, flaw size, pressure, time,(a) of warm F in, psig 0 TC min. prestressing?
550-40 0.21 0 15 No 550-90 0.21 430 35 Marginal 550-120 0.50 771 40 Yes 550-150 0.50 930 50 Yes (a)The critical time (Tc) is defined as the time at which crack inicia-tion is predicted to occur without subsequent crack arrest.
1 1
l l
l l
5-9 . Babcock & Wilcox l
i Table 5-2. Comparison of Reactor Vessel Materials (*
Adjustment NDT' Controlling Lon itudinal Welds Oconee 1 (5.998 EFPY) SA-1493 184(D)
Ocone, 2 (5.528 EFPY) NA(" NA Oconee 3 (5.393 EFPY) NA NA TMI-1 (5.516 EFPY) SA-1526 198 TMI-2 SA-1493 160 Crystal River 3 WF-18/8 144 ANO-1 WF-18 160 Rancho Seco WF-70 222(b)
Davis-Besse 1 NA NA Midland 1 NA NA Midland 2 BZB-243 39(d.e)
Controlling Circumferential Welds Oconee 1 SA-1229 171 Oconee 2 WF-25 223 Oconee 3 WF-67 173 T'.1 -1 WF-25 220 TMI-2 WF-193 158 Crystal River 3 WF-70 184 ANO-1 WF-112 195 Rancho Seco WF-154 187 Davis-Besse 1 WF-233 128 Midland 1 WF-70 151 Midland 2 BZB-243 39(d)
(*)
Reference:
January 1, 1980 plv_s 2 EFPY adjust-ment to RT fr a Regulatory Guide 1.99.
ET These longitudinal welds are in the upper shell and underneath an inlet nozzle.
(*)NA: not applicable.
(d)The material listed is upper shell material, which is controlling over the weld material.
(*}Not a longitudinal weld.
5-10 Babcock & Wilcox -
Figure 5-1. Types of Weld Orientations LONGlTUDINAL 4ELO INSIDE SURFACE .
STRESS DIRECTION d
l CIRCUMFERENTI AL WELO l
l r
CASE I CASE 2 4
~
d d INSIDE =
d SURFACE h INSIDE ~
SURFACE p
777 9
4 d
d l'
+ a/w + a/w
_~
i Babcock & Wilcox 5-11
Figure 5-2. Allowable Pressure Vs Downcomer Temperature Note: P-T combinations are independent of the rate of temperature change, dT/dt, i.e., transient-independent.
2500 Unrestricted operation is allowed to the right 2000 - ar.d below the solid line for first 35 minutes of any rapid cooldown tran-sient.
$ 1500 -
5 2
E u
3 1000 -
a O"'
cm
! 500 l
t 0 i i e i i e i i i i 0 100 200 300 400 500 600 Reactor Coolant Downcomer Temperature.*F l (Temperature at RV Surface at Location of Controlling Material) l l
l 1
l l
l l
5-12 Babcock s.Wilcox l
1 l
Figure 5-5. ntI-2 Inside Surf ace Reactor Vessel - 11 eld 1.ocations of Interest i
Iiiiiiiiril 10 FT 0 30 60 90 120 150 180 210 240 270 300 330 360 Degress
- I 1 i e i 1 a i e i i e i RV FLANCE f -
- 1
! 2.7'
l 1 6@ (
l I I v _
QQ) - -
i b 5.4' l .
== 1.s.
j _ _
'6.l*
! , 7 . l ' _,
i - 29.5' __
i 1 1 2 -= 12.s* =
I --
35.;. =
6.i' 8
w 9'
- y > 44.r =!
s j
Figure 5-6 Crystal River 3 Inside Surface Reactor Vessel - Weld 1.ocations of Interest 1
!I I I i iI i i 10 FT I O 30 60 90 120 150 180 210 240 270 300 330 360 Degrees A E E R I i R R e a 3 A A
't
- RV Fl.ANGE g
- r 2.7'
! o h h O CFh h h h CF h 1 d < 5.4-L- L L L L L _ ,
y V ~ ~
v v E
5.4*
\ t h
= 9.l' _
O'I 31.4' _
U h
- F = 13.6' =
X 6.1 8 = 36.0 z x
P (I
l I --
n '
2 = 44.7' =
l Figure 5-7. ANO-1 Inside Surf ace Heactor Vessel - Held 1.ocations of Interest Iiie iii2_i_I_j j 10 FT C 30 60 90 120 150 180 210 240 270 300 330 360 Degrees I i I .1 8 i i i i i i i i i
RV FLANGE
! I ,7 2 CF 0 0 0 0 CJ 0 0 0 0 f b Q & Q Q (y '
&v - - &_
- ; _ _ _.s- 3.4 i l 1
= ,4.4 = I e as.a. -
s.i-I o 17.6-g , =
g, R _
as.s- ;_
R x ,
P '
W
.= le 44.7' =1
@)
M 1
l l
- , ! i s
e e
r ' '
g ' ' ' _
e 7 4 4 I l D 2 5 5 6 6 .
n L o o.l.
l n.
I I
]
0I (
6 =
3 =
0 h s d
l 3
3 I
7 E
e G l W
l
- 0 0
3 I
N A
L F
h 7 e V s R s
V t
e r
o 0
7 2
i -
n-o c l a 0 e 4 R
e c
l a
a 4 i 2
h 4 9 2
4 a
f a r
S s ut n T
0 1
i h
F 2 '
e er l 0
7 d e n 1 4
it 4 s n 4 Fd
_ i nI 0
_ 81 i C _
e
_ f 1 oo a
=l I c
e s S n oi o
l 0 a h 3- = =
t it c a 5
1 (
nc KIa o.
8 0
2 1
I h
5 e '
i F
r u
g 0 0
t - e '
8 5
0 2
1 2
0i 6 h 0n 3
F = - = _-
C 0g
_, g gg* s. NsE
. I; i ll 4 !, ;i: '
!;, iikl ' iI ' '
Figure 5-9. Reactor Vessel Nozzle 1ocations - Inside Surface.
Typical 177-FA Lowered-1.oop Plant 3.7FT
- - l>iiil i e il '
10 FT 30 60 90 120 150 180 210 240 270 300 330 360 Degrees i
I i ,1 I i, i g
i g
i
- l I 4; I I
' RV FLANGE 199l, 237 l 303 19 57 i i13 i 161 341 f
' ' ' f n i I i I l : j I 3.5*
l j, I i l l l l 1 CF " & y,,,
e' g l
=
O
,0 C- 6 _Ili6- @-
_ _ ,0.. _ . _l..
+ = - 3 r' = = 45" W
8 x
- g. il N
w 44.7' =
Q i=
Figure 5-10. Reactor Vessel Nozzle 1.ocations - Inside Surf ace, Typical 177-FA Haised-l.oop Plant (Davis-Besse -)
- 3 1 Fi L '
10 FT l-- 'l Degress 180 210 240 270 300 330 360 30 60 90 120 150 0 I '
t i 3 I 8 I '
t ,t , i I fg 135 16 189 2h5 315 3'41 RV FLANGE 4h 7
- ; i i j l l i i i " h j ' i 8 8 a 1 3.5 FT i i , l l l VV ., VV h CF h VV l 8 VV h 7.0 FT IN OH l
m O Our 0
CF g w o
I Q
I O ,, i V
+ > -
,i N g
y 28" =- = = 36"
~ *- 11 1/2"
$ 25.7 FT 37n _ = - = 45" - - 14" l
ir 8
x v e
g 44.7 FT =
M
- 6. SCMMARY AND CONCLUSIONS The investigations and analyses described in thiu report were performed to evaluate the concern of brittle fracture following a small LOCA in response to the NRC's information request of July 12, 1979. In addition, an extended loss of feedwater was assumed to occur.
t 6.1. Analyses Performed LOCA analyses were performed for break sizes of 0.007, 0.015, and 0.023 ft 2 assuming no feedwater to the steam generators. Each break size was analyzed assuming (1) no operator action and (2) the operator throttles HPI flow to
! =aintain a subcooling margin of approximately 100F at the core outlet. The LOCA analyses were performed to determine the HPI flow rate, vent valve flow rate and teaperatura, and RCS pressure.
l t
A very conservative evaluation was developed to define the worst-case, bounding
! downconer temperature conditions used for the LEFM analyscs. The bounding case l
assumed that the HPI flow into the downco=er streamlines down along the RV wall with little or no mixing with the vent valve fluid. Resulting towncomer fluid bulk temperatures of 40, 90,120, and 150F were assumed for 18 tuts to the worst-l case ~~.fM calculations.
Other - more realistic - downcomer mixing analyses were performed which in-cluded combinations of some of the following factors: gravity effects, cir-cumferential distribution of the vent valve flow and mixing with the concen-trated EPI flow stream, circumferential distributions, and mixing of both HPI and vent valva flows.
Linear elastic fracture mechanics analyses were performed for each break size, j with and without operator action to throttle HPI flow, for a range of mixing l assumptions.
i 0.2. General Conclusions .
, 1. Vent valve flow occurs during a small-break LOCA.
1 i
6-1 Babcock & Wilcox
(
- 2. As long as forced flow exists, the incoming HPI water will mix with water returing from tha e' _a generators, and no reactor vessel thermal shock concern exists.
- 3. In the case la "kich no loop flow is present, vessel downcomer local wall temperatuces 121 depend on the flow rate and temperature of the HPI water, the flow rate and temperature of the vent valve return flow, and the de-gree of mixing between them.
6.3. Specific Conclusions
- 1. Generic analysis results show RV thermal shock is not predicted to occur for the transients identified below. While brittle fracture failure cri-teria may be violated af ter several hours if certain conditions occur si-multaneously as discussed in section 5, these transients are considered acceptable because crack initiation is not predicted before several hours into the event. Operator action is necessary to depressurize the plant since the applied K exceeding KIC cann e e tolerated indefinitely.
2 0.007 , 0.015 , and 0.023-f t pressurizer breaks with RV downcomer mixing owing to HPI and vent valve flow uniformity in the azimuthal direction with operator action to throttle EPI flow.
2
- 2. The results of the conservative generic bounding analysis, a 0.023-ft pressurizer break with no HPI throttling and rapid RV cooling, show that mixed temperatures of 120-150F in the downcomer present an acceptable con-dition for an extended etze. Operator action would be necessary to ensure the existence of loop flow or depressurize the plant within 40 minutes of event initiation.
Welds most likely to experience rapid cooldown are those veritcally below l
the cold leg nozzles into which HPI water is injected. As discussed in l
section 5, the welds of concern with respect to brittle failure of the vessel are longitudinal welds. Only in the Oconee 1. ANO-1, and Rancho Seco reactor vessels de longitudinal welds exist under or near the leg 1
nozzles.
- 3. State-of-the-art methods do not support very accurate analytical predic-tions of three-dimensional fluid mixing in the downconer. Using the de- .
velopmental model (MIX 2) discussed in section 3, wall temperatures low l 2 enough to raise a brittle fracture concern for the 0.023-f t break are 6-2 Babcock & Wilcox
predicted when compared to the bounding-type limits derived from Figure 5-2 from instantaneous cooldowns and worst-case vessel irradiation to 6 EF"Y.
4 Investigations indicate that no brittle fracture concern is predicted when the downcomer temperatures obtained using the developmental mixing model are compared to the generic bounding-temperature results and the operator takes action to throttle ..'t now. '"essel irradiation through June 9, 1980.)
- 5. Because of th0 analytical limitations, reactor vessel wall fluid tempera-tures cannot be accurately quantified at this tLae. Therefore, a generic bounding analysis has been performed using the following conservatisms which, if realistically modeled, should show that no immediate fracture toughness problem exists:
' fluid mixing and heating in the cold leg piping. Maximum HPI capabil-icy was assumed.
- b. Streamlined HPI flow directly down the inside reactor vessel vall --
No circumferential spreading of the HPI stream was assumed.
l c. Instantaneous reduction of the downcomer fluid temperature next to the i vessel wall at the weld location - No credit was taken for fluid mix-I *ng in the downcomer or for warm prestressing of the vessel materials.
- 6. The analytical results and conclusions in this report are based on a ge-I neric bounding analysis intended to envelop all B&W 177-FA plants. The operational and geometric assumptions used to develop the generic bounding model do not necessarily represent the actual situation for any specific unit. For example, the bounding analysis assumes that the worst-case lon-gitudinal weld exists directly below the f alet nozzle. In fact, only Rancho Seco, Oconee 1, and ANO- have longitudinal velds near the cold leg nozzles. Other plants (Davis Besse 1. Midland, Oconee 2 and 3) do not have longitudinal welds at all. Also, the generic analysis only as-sumes four vent valves rather than the eight which exist on all vessels i
! except Davis Besse. Modelling eight vent valves would result in more vent valve flow to the downcomer. Thus, the actual existence or extent of a brittle fracture concern is highly plant-specific and should be assessed on a plant-by-plant basis.
6-3 Babcock & Wilcox
- 7. REFERENCES 1
CRAFT 2 -- Fortran Program for Digital Simulation of a Multinode R actor Plant During LOCA, BAW-10092, Babcock & Wilcox, Lynchburg, Virginia, April 1975.
D. F. Ross (NRC) to J. H. Taylor (B&W), letter, "Information Request on Reactor Vessel Brittle Fracture," July 12, 1979.
8 F. ~4. Dittus and L. M. K. Boelter, University Publications in Engineering, Vol 2 (1930), p. 443.
G. C. Vliet and D. C. Ross, " Turbulent Natural Convection .. . ," ASME J.
Heat Transfer, November 1975, pp 549-555.
5 Nuclear Systems Material Handbook, Vol 1 (TID-26666), Hanford Engineering Development Laboratory, Richland, Washington (1978).
8 ASMI 3 oiler and Pressure Vessel Code,Section III, Appendixes (1977).
D. T. Suchanan and Y. 9. Esii, FELCON -- General Purpose Program for Solving Ther=al Conduction Ptoblems, NPGD-TM-268. Rev. 1,- Babcock & Wilcox, Lynch-burg, Virginia, February 1979.
8 D. A. Ferril, P. B. Juhl, and D. R. Miller, " Measurement of Residual Stresses in a Heavy Weldment," Welding Journal, WRG Suppitzent, Vol AS, No. 11, November 1966.
ASME Boiler & Pt.asure vessel Code,Section II, Appendix A.
7-1 Babcock & Wilcox
= . - , .
3 Docket No. 50-346 License No. NFF-3 Serial No. 685 January 30, 1980 Attachment B REPORT ON POWER-0PERATED RELIEF VALVE ,
OPENING FR08 ABILITY AND JUSTIFICATION FOR PRESENT SYSTEM AND SETPOINTS
- Submitted to Satisfy Requirements of NUREG-0737, items II.K.3.2 and II.K.3.7 Document No. 12-1122779 - Rev. 1 January 1981 -
i .
J t
I Table of Contents 1.0 Introduction & Sumnary 2.0 Discussion 2.1 Evaluation of PORY Opening Probability During Overpressure Transient 2.2 Evaluation of PORY and Safety Valve Reliability 2.2.1 Safety Valve Failure Rate History 2.2.2 Evaluation of Small Break LOCA Probability /Need for PORY Isolation System 3.0 Conclusions s
e
1.0 INTRODUCTION
AND SU' MARY NUREG-0737, " Clarification of TMI Action Plan Rcquiret.ents," November 1980, required that a report be submitted which provites the information identified in Items II.K.3.2 and II.K.3.7. Specifically, NUREG-0737 requested the following information/ justifications:
- 1. II.K.3.2 Compile operational data regarding pressurizer safety valves to e
detennine safety valve f ailure rates e Perform a probability analysis to determine whether the modifica-tions already implenented have reduced the probability of a small break 1.0CA due to a stuck-open PORY or safety valve a sufficient amount to satisfy the criterion (<10-3 per reactor year), or whether the automatic PORV isolation system specified in Task Item II.K.3.1 is necessary.
- 2. II.K.3.7
! e Perform an analysis to assure that the frequency of PORV openings
)
is less than 5% of the total number of overpressure transients.
l This report is submitted in compliance with NUREG-0737 and demonstrates that the requirements of NUREG-0737 are met with the existing Power-0perated Relief Valve (PORV), Safety Valves and High Pressure Trip Setpoints and that no automatic isolation system is required.
2.1 Evaluation of PORV Opening Probability During an Overpressure Transient An evaluation of the probability of PCdV opening has been performed. Two separate analyses have been performed. The first is an analytical estimate, the second is an analysis based upon operating experience.
2.1.1 PORY 90enino Probability Based Upon Analyses A series of calculations have been completed using best estimate numbers to estimate the probability of PORV opening. Wherever possible, these calculations were based on operating plant data ir, an attempt to provide realistic estimates for the analyzed events. The following paragraphs stenarize the results and calculational basis for the analysis.
[ The probability of the PORY lif ting during a loss of feedwater (LOFW) or turbine trip is approximately 3.9x10-6/Rx-Yr for plants with a PORV setpoint of 2450 psig and 3.9x10-3/Rx-Yr for plants with a PORV setpoint of 2400 psig. The latter setpoint is presently applicable only to Davis desse 1. These probabilities are based on the assumptions that the high pressure trip setpoint is 2300 psig with a standard deviation of 1.4 psi and that the actual setpoint at which reactor trip occurs is a random variable which is normally distributed. The small standard deviation is based on the fact that the PORY and RPS actuation points are not completely independent; i.e., they share a common source; i.e., sensor anc instrunent string. Thus, these parts of the string errors are perfectly correlated and cancel one another in the analysis. Other parts of the relev, ant string error are not correlated and it is upon these that the 1.4 psi standard deviations are based. In a similar fashion, the
_2
actual opening setpoint of the PORV is also assuned to be a random variable with a normal distribution. The assumption of normality for the actuation of either the high pressure trip or the PORV is just an assumption; no data is available to justify or deny the validity. The RCS pressure rise above the RPS high pressure trip setpoint (hence referred to as " pressure rollover") during a LOFW or turbine trip was determined by a combination of plant data and eng.i.aering analysis. Pressure rollover data from the operating plants (Table 2.1-1) was compiled from available data. However, these data points represent situations in which the PORV could open, thus decreasing the amount of pressure overshoot. Tnerefore, it was necessary to correct for the PORV opening, since we are interested in the situation in which it remains closed. This was accomplished by benchmarking the CADD code to a tra.'sient in which the PORV was isolated. After satisfactory duplication of this transient, the code was rerun modeling proper functioning of the PORV. The resulting pressure correction to the rollover data was 17.4 psi. The rollover data itself was tested and is statistically acceptable as normally distributed.
It has a mean of 9.2 and a standara deviation of 27.52 psi. The presence of negative values in this data set indicates that the RPS trip setpoints.
have frequently been set low. Since the data reflects actual operating experience, the use of the negative values can be justified in the l
analysis.
l .
Using the above data and assunptions, a Monte Carlo simulation of the relation l PORY - RPS - EXCESS - BIAS = SAMPLE
. a nam -
v-
was conducted. The terms in the above relation are defined as follows:
PORY - PORY setpoint, a normally distributed random variable RPS - High pressure trip setpoint, also a normally distributed random variable .
EXCESS - Pressure rollover, a randomly distributed normal variable BIAS - A c5nstant (17.4 ps51 defined by analysis which compensates the rollover cata for the fact that the PORY will remain closed.
Six thousand sample values of the above alogrittun expression were calculated using the SAMPLE code. A negative value of the above expression implies the PORY opens. In the computer trials, no negative values in 6000 instances were observed.
It was then assuned that the random variables described above are independent in the probabilistic sense, so an analytic approach was applied. The sun or difference of several independent normal distributions is also a nomal distribution with mean equal to the algebraic sun of the means and standard deviation equal to the square root of tne sun of variances. In this case, the mean is 2450 - 2300 - 9.23 - 17.4 = 123.37 (except DS-1, = 73.37) and standard deviation is (1.4)2 + (1,4)2 + (27.52)2 = 27.59 (for 08-1,= 27.59)
The probability that the PORV will open during an overpressure transient The is 3.9X10-6/Rx-Yr (for DB-1 this value is 3.9X10-3/Rx-Yr).
statistics show that we can be 99% confident that at least 99.99% of all LOFW and turbine trip high pressure transients will not open the PORV for tt e PORV set at 2450 psig. For a setpoint of 2400 psig, the statistics indicate a 99% confidence that more than 99.4% of the overpressure transients will not result in opening the PORV.
2.1.2 PORV Opening Probability Based Upon Operational Data NUREG-0667, " Final Report of the B&W Reactor Transient Response Task Force," contained a listing of reactor trips (148) with PORV actuations prior to the TMI-2 accident. Since the accident at TMI-2 approximately 59 trips have occurred on B&W designed plants. Approximately 42 of these trips would have lif ted the PORV with the old setpoints. Of the 190 trips that would have lif ted the PORY with old setpoints, three of these events would have lif ted the PORV with the new setpoints. In addition the i
modifications that have been made to the plants since those transients would have precluded PORV actuation given the same initiating events on those plants and the new setpoints. Based on these data, it is estimated that the present PORV opening probability is less than 1.6% for an over-pressure transient. The 5% probability stated in II.K.3.7 of NUREG-0373, though irrelevant and inappropriate, is currently achieved. The relevant probability is not that of opening, but that of 'an unmitigated sticking open, which is obviously very low though not quantified.
1 l
l l
t
TABLE 2.1-1 PRESSURE ROLLOVER DATA Trip # Power, % Peak Pressure, psig Rollover, psig 1 95 2355 0 2 90 2385 +30 3 25 2400 +45 4 20 2385 +30 5 90 2390 +40 6 32 2345 -10 7 40 2360 +5 8 40 2352 -5
~
9 92 2375 +20 10 15 2365 +10 11 35 2400 +45 12 13 2370 +15 13 14 2355 0 14 38 2380 +25 15 98 2410 +55 16 72 2400 +45 l
17 100 2340 -15 18 100 2340 -15 t 19 100 2390 +35 20 100 2330 -25 21 98 2325 -30 22 15 2355 0 23 9 2370 +15 24 30 2345- -10 25 99 2350 -5 26 16 2295 -60 O
e*-== e r< a c. -- . ~ , . . , ,.7 _ , ,
l 2.2 Evaluation of PORV and Safety Valve Reliability 2.2.1 Safety Valve Failure Rate History There have been three cases where pressurizer safety valves were lif ted on B&W plants. None of these cases resulted in failure of the safety valve to reseat. Because of the few data points, no estimate was made of the safety valve failure rates.
2.2.2 Evaluation of Small Break LOCA Probabilities /Need for PORY Isolation Systen The contribution to the probability of a 58 LOCA from an open PORV was estimated by two methods. The first was an analysis effort, the second was based strictly upon operational data. The results are discussed below:
2.2.2.1 Small Break LOCA Probability Calculations The probability of a stuck open PORV is the product of the probability of being demanded open times the probability of failing open on demand.
The raising of the PORV setpoint has reduced the number of demands and thus the probability of being in the stuck open state. The point f
estimate for PORY SB LOCA probability (variation not estimated) is calculated to be 5.04 x 10-4 per reactor year (5.48 x 10-4 for Davis Besse) which is less than the II.K.3.2 requirement of 1 x 10-3 per reactor year. The initiators of PORY actuations have been grouped into five categories along the associated frequency of each category.
Details on how the values are calculated are contained in Table 2.2.2-1.
- 3. PORV opening on operator action under 1.58 x 10-2/Rx-Yr ATOG guidelines
- 4. PORV opening due to instrunenta, tion 5 x 10-3/Rx-Yr control faults
- 5. PORV opening from additional 1.8 x 10-3/Rx-Yr consideration from II.K.3.7 TOTALS 2.40 x 10-2/Rx-Yr 2.61 x 10 /Rx-Yr(08)
This total is then multiplied by the probability of the PORV sticking open on demand.
Note that all plants except Davis Besse (Crosby PORV) have Dresser valves; however, the entire B&W operating plant experience was used to arrive at a generic PORV sticking open probability as follows: There have been ten stuck open PORY events, five of which could be classified as mechanical failure of the PORY (the other five were basically installation errors).
Using all these five failures in determination of future frequency is considered conservative since two of the failures (OC-3,6/13/75 and CR-3, f
11/75) were rectified by design changes, another (TMI-2, 3/28/79) cause is unknown. OC-2, 11/6/73 could be considered as a burn-in failure and the l
I I
DB-1, 10/13/77 event is a Crosby valve. Using five failures in 250 demands results in a value of 2 x 10-2 to fail to reclose on demand. This value is considered conservative not only due to the inclusion of all five failures but also the nunber of demands is probably much higher than 250. There have been 148 documented PORV openings on reactor trips; however, there is not a listing of PORV demands when the reactor did not trip (e.g., ICS runback) nor is consideration given to transients that cc !d have actuated the PORV f '
nunerous times during an event. The value of 250 demands is conservatively used here. An analysis was also performed to include j values for other than mechanical failure that keep the PORV open. The results of this analysis is sunmed with the mechanical contributor (2 x 10-2/d) to arrive at the value for failure to reclose on demand (2.1 x 10-2/d).
Probability of PORY small break LOCA equals:
(2.4 x 10-2)(2.1 x 10-2/d) = 5.04 x 10-4/Rx-Yr j (2.61 x 10-2)(2.1 x 10-2/d) = 5.48 x 10-4/Rx-Yr(08) 2.2.2.2 Small Break LOCA Probability Based Upon Operational Data
.As discussed in Section 2.1.2, there have been three events which with the revised setpoints would have actuated the PORY. However, the plants have been reconfigured (e.g., upgrades on aux. feedwater, control circuitry of PORV, NNI power sources, AC power sources) so as to reduce the probability of these PORY actuations. Conservatively estimating that one event could occur in the 45 years of B&W plant
- operation, yields a probability of occurrence of 2.22 x 10-2/Rx-Yr.
l I
L
The previous section gave a PORV f ailure probability of 2.1x10-2/d.
Therefore the probability of a PORV snall break LOCA equals:
(2.22x10-2d/Rx-Yr)(2.1x10-2fd) , 4,7xio-4/ Rx-Yr which is less than the 1.0x10-3/Rx-Yr criterion.
3.0 C0t;CLUSIONS Both the analytical prediction and the estimate based on historical data result in values for a stuck open FORY from all causes which meet the requirements given in II.K.3.2. Note that no credit has been assigned for the operator closing the block valve given an open PORY. A1alytical predictions (given proper auxiliary feedwater response) result in a value less than .01% of PORV openings for overpressure transients (taking into account the most limiting non-anticip'atory trips) and historical data shows the frequency to be less than 1.6% which satisfies the criterion (less than 5%) specified in II.K.3.7.
Since the requirements of II.K.3.2 and II.K.3.7 are met with the current PORY configuration and set point it is not necessary to address the requirement for an automatic block valve closure system per II.K.3.1.
l
[
l i
l ,
1 l
Table 2.2.2-1
- 1. The probability of a PORV opening on an overpressure transient from Section 2.1.1 for plants with PORY setpoint of 2450 3.9 x 10-6/Rx-Yr for plants with PORV setpoint of 2400 (08) 3.9 x 10-3/ Rx-Yr
- 2. The PORV opening probability in a transient with delayed aux. Teeo A value of 1.0 was assigned for PORV 1.4 x 10-3/ Rx-Yr opening probability if aux. feedwater was not supplied. A value of 1.4 x 10-3/P.x-Yr for loss of all feedwater was referenced from a 8&W calculation which used average unavailability as calculated in the generic aux. feedwater reliability studies (BAW-1584) in conjunction with generic EPRI data on loss of mai- feedwater frequency and loss of offsite power frequency.
l l
On completion of the ongoing aux, feedwater reliabilityanalysis(AP&L,SMUD,FPC)more i
specific values can be applied to those plants.
- 3. The PORV opening proba~'lity on operator action under ATOG guioellnes There are 3 events that call for l operator opening of the PORY: a) Loss of All Feedwater. This contribution is already counted i in 2 above; b) Small LOCA. Not applicaole to l
l l
l l
l l
i
'I Table 2.2.2-1 (Cont'd) this calculation since the plant is already in a small LOCA; c) Steam Generator Tube Rupture (considered smaller than small LOCA as defined in II.K.3.2 so argunent of .
b) does not hold): The demand on the PORV given a tube rupture varies depending on l i
! whether offsite power is available or lost.
If offsite power (Reactor Coolant Punps) is available, only one PORV opening is required,
)
whereas in the loss of offsite power scenario as many as 23 PORV openings are required.
The value calculated assumes that the probability of Steam Generator Tube Rupture considered with a LOOP event is small (no j causal effect of LOOP or Steam Generator .
I Tube Rupture) and therefore, the WASH-1400 of I x 10-3 for a LOOP given a reactor trip is used in the calculations. There have not been any tube ruptures in the cunulative 8&W experience, (45 Rx-Yrs) due to the limited nunber of years experience. A Chi-square 50% confidence value with 0 failures is rather high (1.54 x 10-2 Rx-Yr).
i l
l
- . ~ . - . .
Table 2.2.2-1 (Cont'd) 1.54 x 10-2/Rx-Yr x 1 demand (off *te ,
power available) 1.54 x 10-'/Rx-Yr 1.54 1 /Rx-Yr x 10 3Of0-2 fsite Power Loss / Event x 23 demands (offsite power lost) 3.54 x 10-4/Rx-Yr 1.58 x 10-2/Rx-Yr In the final calculation of probability to reclose, it should be noted that no adverse effects of the 23 demands in the loss of offsite power case on PORV .
operability is assuned.
- 4. PORV openina due to instrumentation control faults This has been estimated at 5 x 10-3/ 5 x 10-3/Rx-Yr .
reactor year. This value assunes that power supply faults and other control deficiencies have been corrected.by each u tility.
- 5. PORV openino probability from additional consioerations from II.K.3.7 T%ra *ra overcooling transients that initiate HPI and operator failure to throttle or terminate flow before the PORV setpoint is reached. There have been 8 overcooling transients that initiated
, .y- <
T[ble 2.2.2-1 (Cont'd)
HPI in 392 reactor trips. The current frequency of reactor trips is 6 trips /
Rx-Yr per plant. In this event sequence, the operator hat approximately 4 minutes from time of HPI initiation until PORV setpoint is reached. The operator
- failure rate to terminate or throttle HPI flow is based on having ATOG in place (1.5x10-2/d - based on NUREG-CR-1278 with moderately high stress). The overall proba'uility of tais seoaence is therefore estimated to be 6 trips /Rx-Yr x 8/392 overcooling events / trip x 1.5x10-2 = 1.8 x 10-3 Rx-Yr N.A. for DB TOTALS 2.40 x 10-2/Rx-Yr 2.61 x 10-2/Rx-Yr (08) l l
Note that these values are dominated by the conservative analysis of steam I
gensrator tube rupture. Analytical studies could be performed to obtain a more realistic value. Also note that the calculation for category 4 did not include operator or maintenance induced faults, such as the DB event of 10/27/80.
l l
i .
l i
r - , , --