ML20235N447
ML20235N447 | |
Person / Time | |
---|---|
Site: | Vermont Yankee ![]() |
Issue date: | 06/30/1987 |
From: | Jackie Jones, Wheatley P EG&G IDAHO, INC., IDAHO NATIONAL ENGINEERING & ENVIRONMENTAL LABORATORY |
To: | NRC |
Shared Package | |
ML20235N403 | List: |
References | |
CON-FIN-D-6022 EGG-RTH-7506, NUDOCS 8707200073 | |
Download: ML20235N447 (85) | |
Text
, . - . -- .. .
' ~
f .' 'd .; . '
f s
.g - <
[l. ll '
- ., EGG-RTH-7506 June 1987 i
a' INFORMAL REPORT
.r t
[ 4 /(
- idaho. TECHNICAL EVALUATION REPORT
- REVIEW AND ENational EVALUATION OF THE RELAP5YA COMPUTER CODE AND THE yL ,
Engineering" , VERMONT YANKEE LOCA LICENSING ANALYSIS MODEL FOR' USE IN SMALL AND LARGE BREAK BWR LOCAS
. Laboratory
" i 4 .t P ' U- Managed: '
. J. L. Jones
'. .by the U.S.: P. D. Wheatley V. " Department .;
?il
?
.of Energy . ,
)-,
9 i
.s4
'.l.
.E
,\
1:
li1:
p 1
p.;-; ?-
),e
'i .
h EGzG, .u.
- No'j"ci,","l,': j Under DOE Contract No. DE-AC07-761001570 No. DE-AC07-76/D0tS70 :
FIN No. 06022 87'o7poo673 070630 fDR ADOCK obocopyj PDR
..*.----.-.__2. u._m___m.a. . - ____m____ _ _ _ . , _ _ _ _
~
Y J'
l i' ' EGG-RTH-7506 1
4 .
7 je . -
TECHNICAL EVALUATION. REPORT: ,
e REVIEW. AND EVALUATION OF/THE' RELAP5YA COMPUTER CODE-AND THE VERMONT YANKEE LOCA ANALVSIS MODEL FOR USE IN SMALL AND LARGE BREAK BWR LOCAs
.J. L. Jones P. D. Wheatley Published June 1987 Idaho National Engineering Laboratory Idaho Falls, Idaho 83415 Prepared for the U.S. Nuclear Regulatory Commission Washington, D.C. 20555 Under DOE Contract No. DE-AC07-76ID01570 FIN No. 06022
l a. ,
I e
ABSTRACT W -A review was: completed of th'e RELAP5YA computer code to determine its capabilities for performing..l.icensing analyses. The review was limited to .
- Boiling Water . Reactor (BWR) reactor applications. 'In addition, a
- Loss-Of-Coolant. Accident '(LOCA) licensing analysis model, using the:
-RELAP5YA computer code, wasLreviewed. This model was' reviewed for.
.' application"to.the, Vermont Yankee Nuclear Po'wer Station to perform. full break; spectra'LOCA and fuel cycle independent analyses ( The review of the
.RELAP5YA' code consisted of an evaluation of all, Yankee Atomic Electric
~
Company.-(YAEC) incorporated modifications to the RELAP5/M001 Cycle 18 Lcomputer code from' which the-YAEC version' of the code originated.
7- LQualifying. separate and integra1' effects assessment calculations were reviewed to evaluate the' validity and proper implementation of the various added models. Ths LOCA-licensing method-was assessed by reviewing two RELAP5YA' system input models and evaluating-several small and large break qualifying-transient' calculations. The review of the RELAP5YA code
. modifications and their. assessments, as well as the' submitted LOCA input model,.is discussed and the results of the review are provided.
e ii a'
n . .
i i
SUMMARY
This reportLeovers the review and evaluation of the Yankee Atomic
' Electric. Company-('YAEC)' computer code RELAPSYA and a full break spectra l ,
LOCA' licensing analysis method applicable to the Vermont Yankee Nuclear Power. Station. RELAP5YA'is a computer program capable of analyzing.
. steady-state and transient thermal-hydraulic behavior in light-water /
reactor systems. The code:has features that allow compliance with requirements in 10CFR50.46 :and Appendix K. The code was developed from the RELAP5/M001 computer code originally developed and released by the Idaho National' Engineering Laboratory' (INEL). The code and the LOCA analysis licensing model were submitted to the Office of Nuclear Reactor Regulation (NRR)'within;the Nuclear Regulatory Commission (NRC) by YAEC for review and-approval. NRR. requested assistance from INEL in this review. Assistance intthis review was limited to.those~ aspects pertaining to BWR related steady-state.and transient' applications.
The review consist'd e of first reviewing the RELAP5YA computer code for BWR related applications. This review concentrated on the modifications YAEC made to'the RELAP5/M001 Cycle 18 computer code and their implementation of these modifications. Each modification is separately discussed and assessed. The code was reviewed to assure that known updates and corrections made by INEL to RELAPS/ MOD 1 Cycle 18 during the development of RELAP5YA were incorporated or appropriately justified if some were excluded. ' Finally, the code is reviewed for compliance with NRC requirements. RELAP5YA was reviewed both'for compliance with Appendix K required models and for use for best estimate calculations with additions for calculational uncertainties under SECY-83-472.
- As a result 6f this RELAP5YA review, it is recommended that RELAP5YA be accepted as a best estimate code for performing BWR small and large break LOCA and fuel cycle independent analyses on the basis that suggested conditions and requirements are followed. RELAP5YA also includes many of the models or capabilities required by Appendix K.
iii
y - .
Based on the recommendation of RELAP5YA, a review of a LOCA-licensing :
analysis model, applicableto the' Vermont Yankee Nuclear. Power. Station was 7
undertaken. LThe model actually uses two'RELAP5YA input'models: a full Nuclear Steam Supply Steam (NSSS);model and a specific. Hot Channel (HC)
.model. .The;NSSS'model includes the main system components, such as, the reactor vessel'with internals and core, two recirculation-loops with jet pumps,;feedwater l'ines, main steam lines, ECCS systems, trip'.and control logic systems',- and point reactor kinetics. The NSSS model. includes' a.
representative double-ended-guillotine (DEG) large break modeling anL approximately 28' inch diameter break in the discharge pipe of one.
recirculation loop.just. upstream of the header pipe. T.he important large break' accident' assumptions include a coincident loss'of normal. auxiliary power, failure of the.LPCI in both recirculation' loops, and failure of the recirculation lo'op discharge valve to close on demand due to its proximity to the break. Only HPCI and two LPCS systems are available-to mitigate
~
large break accidents. To represent small breaks, the NSSS model-also includes'a representative three. inch diameter break.at the.same location
. postulated to' fail and only.two'LPCS and two LPCI systems are available to i mitigate'the LOCA accident. Several large and small break transient calculations were presented by YAEC and reviewed. Based on this review, it.
is recommended that the NRC-NRR approve the model for use with RELAP5YA on '!
the basis that suggested conditions are met.
I l ..
I i
iv
j {. .
4 L " '
CONTENTS ABSTRACT:... ..........................................................
. ii.
?
SUMMARY
b................................................................'iii-
- 1. INTRODUCTION ....................... ........................ ..... 'l '
- ~2.- RELAP5YA CODE' DESCRIPTION AND MODIFICATION ASSESSMENT . . '. . . . . . . . . . 4-
.2.1 . General' Code 0verview ........................................ 4-2 .' 2 Model Description and Assessment ............................. 6 2.2.1' . Interphase Drag Models. for Vertical Flow; Regime . .. . 6 2.2.2 C r i t i c a l Fl ow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 2.2.3 Jet Pump:..................... ... ................. 14 2.2.4< Nucleate Boiling..................................... 20 2.2.5 ' Critical Heat Flux ................................. 21 2.2.6 .Rewet and Quench ................................... 22 2.2.7 Multiple Surface Radiation ......................... 25 2.2.8 Heat Transfer Logic Options ........................ 26 2.2.9 Fuel Behavior Models ............................... 27 2.3 Phenomena Models Important to BWR Systems .................. 30
-2.3.1 Steam Separator.Model .............................. -30 2.3.2 Channel Box......................................... 31 2.3.3 Counter-Current' Flow Limitation .................... 31 2.3.4 Condensation and Vaporization ...................... 31 L2.4' Integral, System Calculations ............................... 33 2.4.1 THTF Calculations .................................. '34 2.4.2 T LTA C a l c u l a t i o n s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
- 3. RELAP5YA CODE UPDATE REVIEW'...................................... 44
- 4. NSSS BWR LOCA LICENSING ANALYSIS MODEL . .. . . . . . . . . . . . . . . . . . . . . . . . . 46 4.1 Vermont Yankee Vessel Model ................................ 48 i i
i
, 4.2' Core Power.................................................. 50. j l
'4.3 Hot Channel Model .......................................... 51 4.4 LOCA. Transients ............................................. '53 y^
4.4.1 Large Break LOCA Transients .... ................... 54 4.4.2 Small Break LOCA Transients ........................ 60 l
1 i
v b _ . _ _ _ _ _ _ _ _ _ _ _
E' s f'
n 15 .' COMPLIANCE WITH NRC' REQUIREMENTS ................................. 63
.6. REVIEW:0F UTILITY QUESTION RESPONSES ............................. 69
'7. RECOMMENDATIONS AND CONCLUSIONS .................................. 70
- 8. REFERENCES .... .............................................. .... 74 FIGURES
- 1. TLTA Test 6425/2 upper plenum fl uid ma s s . . . . . . . . . . . . . . . . . . . . . . . . 10
- 2. .TLTA Test 6425/2 bundle fluid mass ............................... 11-
- 3. TLTA. Test 6425/2 lower plenum fluid mass ......................... 12
- 4. Jet pump mode 1..................................................... 17 ;
t t 5. TLTA Test 6425/2 broken loop jet pump mass flow rate ............. 18
- 6. TLTA Test 6425/2 intact loop jet pump mass flow rate . . . . . . . . . . . . .
19
- 7. Comparison of calculated and measured critical qualities ......... 23
- 8. TLTA Test 6425/2 lower plenum pressure response ............ ..... 37-
- 9. .TLTA Test 6426/1' lower plenum pressure response ... .............. 39
- 10. TLTA Test 6425/2 drive line break flow ........................... 40
- 11. TLTA Te st 6426/1 d ri ve l i ne brea k fl ow . . . . . . . . . . . . . . .. . . . . . . . . . . . . 41
- 13. Vermont Yankee hot channel (HC) model nodalization ............... 49
- 14. Vermont Yankee =one-eighth rod bundle reduction layout for HC model ..................................................... 52 15 . High power as sembly fl uid mas s ( LBLOCA-EB) . . . . . . . . . . . . . . . . . . . . . . .
. 56
- 16. High power as sembly fl uid mas s ( LBLOCA-EA) . . . . . . . . . . . . . . . . . . . . . . 56
- 17. L Long-te rm brea k fl ow rate s . ( LBLOCA-EA) . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 .
L18. Long-te rm brea k fl ow rate s ( LB LOCA-BA) . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
- 19. Bundle heat transfer coefficient comparisons at elevations 71-79 in, for TLTA Test 6422/3 data and Test 6425/2 predictions ...................................................... 65 vi
- 20. Bundle heat transfer coefficient comparisons at elevations79-100 in, for TLTA Test 6422/3 data and Test 6425/2 predictions .................................................... . 66
- 21. Bundle heat transfer coefficient comparisons at elevations 101-122 in. for TLTA Test 6422/3 data and Test 6425/2
. predictions .............. ............... ....................... 67.
TABLES i i
1 1., Summary of Vermont Yankee large break accident assumptions. . . . . . . . 55
- 2. Summary of Vermont Yankee small break accident assumptions ....... 58 O
l vii l
L' _ _ _ _ _ _ _ _ . _ _ _ _ _ _ _ _
e, ,
~
6 '
-TECHNICAL EVALUATION REPORT:
RE' VIEW AND EVALUATION 0F THE RELAP5YA COMPUTER CODE'-
l
~ AND THE VERMONT' YANKEE LOCA ANALYSIS MODEL FOR USE IN SMALL~AND LARGE BREAK BWR LOCAs
'1. INTRODUCTION a '
)
.RELAP5YAI '-is a computer program for Light-Water Reactor (LWR) System thermal-hydraulic' analysis which.provides integral analysis capability of-theLsystem and' core steady-state and transient response to normal and-2 off normal events. .RELAP5YA was adapted from RELAPS/M001 Cycle 18 by Yankee Atomic Electric Company-(YAEC) for, use in Loss-of-Ccolant Accident
- (LOCA) analyses. ; The RELAP5YA code was submitted to the Nuclear Regulatory Commission (NRC) by YAEC.for review and acceptance for licensing; applications as,a method to' analyze the entire Bolling Water Reactor (BWR) break spectrum in a manner that conforms to U.S. Nuclear Regulatory Commission (USNRC) requirements.
-YAEC'also submitted a BWR LOCA input mcdel for review and approval by the NRC. This model is applicable to the Vermont' Yankee Nuclear Power
. Station to perform full break spectra LOCA and fuel. cycle independent.
analyses that comply with. USNRC regulations. The Vermont Yankee licensing
. analysis mode 1Luses the RELAP5YA computer code and two base input models.
The first model'is the main Vermont Yankee Nuclear Steam Supply System
'(NSSS) model used to perform full break spectra analyses. The second model is the Vermont Yankee Hot Channel (HC) model which is used to calculate the
' detailed hot channel response using boundary conditions obtained from the full NSSS model. With,the results from both of these models, YAEC expects to determine the Design Basis Accident for the Vermont Yankee Nuclear Power Station. '
I a- _=-
f
- The-Office of Nuclear l Reactor'. Regulation (NRR) is responsible.for the evaluation.'and review'of computer codes and their proposed applications.
NRR. requested the Idaho'. National Engineering Laboratory (INEL) provide Lassistance:inLthe: review.of the.YAEC RELAP5YA' computer code with-respect to LBWR'related applications. Specifical'y, the request for assistance included: -
6! /
~ "
- 1. Evaluation of RELAP5YA as a' method to analyze the entire BWR
' break. spectrum.
. i'
- 2. Assurance that the code corrections since the release'of
. RELAP5/M001 Cycle 18 (which is the basis for RELAP5YA) were
' incorporated in RELAP5YA.
- 3. l Review of the Vermont Yankee BWR LOCA analysis model 'as described in YAEC-1547.3 f 4. Evaluation of compliance with requirements contained in 10CFR50.46 and Appendix K.
- 5. Evaluation of RELAPSYA suitability for best estimate calculations within the NRC guidelines of.best estimate codes with uncertainties (SECY-83-472).4 Related to the above reviews, the NRR also requested that INEL review and evaluate the utility responses to the NRC questions regarding the LOCA modeling and applications. Three sets of questions were reviewed: general PWR'and'BWR quessions,5-9 specific BWR related questions,10 and the final set of PWR and BWR related questions 11,12' generated by this review.
This report contains the results of the RELAP5YA review and assessment i for performing'ful'1. spectra break BWR LOCA analyses as well as a review of .
the submitted Vermont Yankee licensing analysis model. Section 2 preser.ts ,
a brief overview of the history of RELAP5YA and its development from the 2
-- -- - ___-_b
RELAP5/ MODI Cycle.18 computer code developed at the INEL and publicly released. This section also discusses and assesses the specific BWR related and licensing-related modifications incorporated into RELAP5/ MODI Cycle 18 resulting in RELAP5YA as a licensing analysis code. Section 3
, reviews the implementation status in RELAP5YA of all documented modifications and/or correction updates generated by the RELAPS code development group at the INEL. Section 4 presents the review of the actual Vermont Yankee LOCA licensing analysis model. Section 5 reviews the code and the LOCA analysis method for compliance with NRC requirements defined in 10CFR50.46', Appendix K, and other NRC requirements. Section 6 presents a general cverview of the responses to all question sets submitted to the NRC durine this review process. Section 7 presents the recommendations and conclusions reached from this review with Section 8 listing the references.
f O
e 3
l j
'y l
+ l
{ 2. RELAP5YA CODE DESCRIPTION AND MODIFICATION ASSESSMENT i
This section will present a brief description of the RELAP5YA code and
'in:'ipate its relationship to the publicly released RELAPS/ MODI code. Then tach.of YAEC's specific BWR related modifications and assessments 13 gg)) ,
ce ciscussed and reviewed to evaluate their validity and acceptability.
Finally, RELAP5YA will be assessed against applicable integral tests.
t 2.1 General Code Overview The RELAPSYA code was developed from the RELAP5/ MOD 1 Cycle 18 code which was originally developed and released by the INEL under USNRC sponsorship. YAEC retained the same basic formulation of the differential equations for the thermal-hydraulic models, code architecture, principle solution techniques, and user convenient features as in RELAP5/ MODI with some changes to the constitutive relations. Therefore, m eept for the nine areas noted below and the specific PWR related modifications,1 RELAP5YA and RELAP5/H001 Cycle 18 are very similar.
The application of RELAPS/ MODI at the INEL to a wide variety of thermal-hydraulic problems has shown the formulation of the differential equat9ns and solution techniques provides numerically stable solutions.
With RELiPBYA maintaining the same basic approach and with a detailed asressment of all YAEC's modifications, it can be concluded that it will also provide numerically stable solutions. It is recommended that the general code structure be accepted for BWR licensing analyses.
RELAP5YA is a LWR system transient simulation code. The code is based on a one-dimensional, two-fluid, nonequilibrium model which includes the ,
necessa y thermodynamic state relations and constitutive ecuations to descr'.be two phase flow. To minimize the number of constitutive relations .
required for mathematical model closure, the least massive phase was assumed to be at the local saturated condition. The code formulation of the hydrodynamic components, power sources, heat structures, trips, and f'i control systems provides a flexible method of modeling LWR systems.
l 4
1
RELAPSYA includes many general component models from which general systems can be simulated. The component models include pumps, valves, pipes, heat structures, reactor point kinetics, and control system components. Special process models are included to account for form losses, abrupt area
, changes, branches, and choked flow.
RELAP5/ MOD 1 has been modified by YAEC to provide improved code simulation capabilities and to provide teatures compatible with requirements in 10CFR50.46 and Appendix K. The BWR related modifications reflected changes in the following areas:
- 1. New interphase drag models for vertical flow
- 2. Addition of the Moody two phase critical flow model
- 3. Addition of a jet pump model
- 4. Modification of the nucleate boiling algorithm
- 5. Addition af a new critical heat flux (CHF) option
- 6. Addition of a rewet and quench model for reflooding and spray cooling
- 7. Addition of a multiple surface radiation heat transfer model
- 8. Addition of return-to-nucleate boiling and transition boiling lockout logic
- 9. Addition of new fuel rod behavior models from the T00DEE2-EM code.14 The modifications for er.ch of these specific areas and their assessment are presented in more detail in Section 2.2.
5
, , t
.j
- g. .X n 2.2 tModel' Description'and Assessment t ,
This section:will describe and assess each of the YAEC's specific BWR related' modifications to. evaluate their acceptability to perform LOCA i
1'
. analysis.; --
-2 2.1 ' Interphase Draa Models for Vertical Flow'Reaime The interphase drag force provides the primary mechanism for phase separation in two phase' flow. :At high drag force the' phases travel
'together, while a low drag. force results in the. phases traveling at different: velocities. At very low values of interfacial drag,. the phases scan also. travel'in different directions resulting'in counter-current flow.
This;1atter phenomenon istvery important for the geometric arrangement of
'BWR reactor cores Lsince Low-Pressure Cooling Spray (LPCS) liquid is injected into the' upper plenum of a BWR an'd drains down into-the: fuel bundles and bypass-: region; 'The counter-current flow limitation-(CCFL) phenomenon 'can significantly reduce coolant in a fuel bundle due.to holdup of.eme'rgency coolant. CCFL can also occur' at the side entry orifice (SEO) of aBWR fuel bundle. and. perhaps at a jet pump's ' drive' nozzle from Low Pressure Coolant Injection (LPCI) during 'a blowdown. Since RELAP5YA does not' provide any special. treatment of the CCFL phenomenon due to its' complex
- f. low. geometry dependence, YAEC assessed this phenomenon relative to their
~
vertical interphase drag models.
YAEC's preliminary assessment of the interphase drag models in RELAP5YA revealed that new vertical flow regime models were required in the mass flux range.of 0-2000 kg/s-m2 , which represents the most prevalent flow' range during normal and off-normal LWR conditions. This range was
. further subdivided into less than 100 kg/s-m2' for very low flows, between
.100-150 kg/s-m2 for low flows, and between 150-2000 kg/s-m2 for .
' intermediate flows. In these flow ranges, YAEC recognized two problem
[- ) areas they felt needed correcting. First, for low flow rate tests in the Two-Loop Test Facility (TLTA),15 the code was not able to predict the two phase level transient in the bundle. This was because the calculated
- l. void fraction at some locations in the TLTA heated bundle showed large 6
l oscillations due to numerical instabilities at low flow rates. Second, void fractions at intermediate flow rates were underpredicted in the FRIGG 16 tests by up to 25's.
The numerical instability at low flows was traced to the interphase velocities. Old time values for the relative velocity were used to
- calculate the interphase drag, which was then used to calculate new time values for the relative velocity. YAEC felt this solution technique resulted in the numerical instabilities and, therefore, they decoupled the interphase drag calculation from the old time value of the relative velocity at low flow rates. The new interphase drag model is restricted to low flow rates where reasonable models for the relative velocity in vertical flow could be found. To improve the calculated void fraction at intermediate flow rates, a review of available literature indicated the model for interphase drag could be improved to provide a better phenomenological description. As a result, a new flow regime map for vertical flow was developed.
To evaluate RELAP5YA's ability to adequately predict interfacial drag and, hence, address the counter-current flow phenomena in two pnase systems such as in BWRs at steady-state and transient conditions, calculations were compared to the FRIGG IO and GE swell level I7 tests, respectively. Code calculations of the void fractions were compared to experimental void fractions since direct measurement of interfacial drag is not possible and since the void distribution at low mass fluxes depends strongly on the interphase drag force.
Calculated comparisons to the steady-state FRIGG test data (test numbers 313006, 313007, 313017, and 313020) indicate that for mass fluxes 2
between 729-1464 kg/s-m the interfacial drag is adequately predicted except in the low void fraction region. In this region the lack of a subcooled boiling model in RELAP5YA results in data discrepancies. While these discrepancies did exist, exit void fractions were adequately predicted for the test cases studied. Since the exit void fraction will depend on the slip between the phases which in turn depends on the interphase drag model in RELAPSYA, good prediction of exit void fraction 7
7 l~
l L
L indicates an adequate l calculation of the bundle response by the interphase model forl steady-state tests. While RELAPSYA-does provide reasonable exit v'oid fraction. predictions'.at steady-state, the code cannot always guarantee prediction of. the' steady-state ~ axial void distributions due to the lack of
.a subcooled boiling'model. However, any difference'in axial void fraction -
distribution-will only. modify the time of peak clad temperature (PCT) since most'of the high power bundle undergoes CHF anyway. Hence, it was judged -
- that the lack of a subcooled boiling model does not have a major impact for BWR-LOCA analyses.
For the low flow (0-100 kg/s m2 ) transient conditions, code calculations'were compared to GE Vessel Blowdown Test No. 1004-3, which represented BWR-type conditions. For all transient cases evaluated, the calculated void fractions were always below that observed in the data in the lower parts of the vessel. Exit void fractions always reached near unity at a: lower elevation than in the test, These results indicate that the transient calculation of the interphase drag was representative of the data.
Another assessment of the drag model at low mass fluxes (0-20 kg/s-m 2) was.in the Two-Loop-Test' Apparatus 15 (TLTA) Boil-Off.
Test 6441/6 (Section 2.4.2). This test simulated conditions which would
' occur during a'small break LOCA if no emergency core cooling systems, including automatic depressurization system (ADS), were available. The transient resulted in a slow boil-off at constant pressure and constant bundle power. RELAP5YA results for various locations in the bundle showed that void fractions either showed acceptable comparisons or showed conservative results with the code predicting higher void fractions at earlier times in the transient. ~
1 After reviewing the interphase drag assessment test cases, it was .,
found that no assessment was presented for mass fluxes in the range of 2
100 to 729 kg/s-m . -In response to this deficiency (response A1.11 of Reference 12), YAEC presented additional FRIGG loop data for assessing the interphase drag models for these intermediate flows. These tests, labeled 8 I
L Test Section FT-36c,. Runs.613124 and 613120, were run at a pressure of 2
30 bar and with mass fluxes at 472 and 483 kg/s-m , respectively.
Comparison of RELAP5YA predicted steady-state void fraction distributions to that of .the test data showed very acceptable comparisons for both mass
., fluxes. presented in this intermediate range.
V Finally, to supply additional evidence that the current vertical flow
. regime model provides reasonable estimates of bundle responses, YAEC presented comparisons of the BWR-type TLTA Test 6425/2 data (see
'Section 2.4.2) to RELAPSYA predictions.10 These comparisons illustrated RELAP5YA's ability to predict CCFL at the upper tie plate and the SEO. As shown in Figure 1, these results clearly indicate the code overpredicts upper plenum holdup, and overpredicts bundle fluid mass draining rates as
-seen in Figures 2 and 3. These overpredictions, however, would result in higher PCTs during the LOCA transient calculations.
Based on these steady-state and transient assessment calculations, the new interphase drag models in RELAP5YA are judged to be properly implemented, to provide an acceptable simulation of the interphase drag related phenomena, and to yield conservative results whenever best estimate e results could not-be obtained. Further details of the code development effort for the drag models are described-in Section 3.1 of Reference 1.
2.2.2 Critical Flow RELAP5YA contains two critical flow models which can be selected by the user. The first is the Moody Critical Flow model and the second a best estimate critical flow model already in RELAP5/ MOD 1. The Moody model wil' be discussed first followed by a short discussion of the best estimate critical flow model.
Appendix K to 10CFR50 requires the Moody Critical Flow model be used at all break locations to calculate two phase critical flow during a LOCA.
YAEC added the Moody model to RELAPSYA as a user option. The model was
. incorporated as a table of mass flux versus stagnation enthalpy and pressure. The critical flow table may also be entered with static pressure 9
] i 1i\
- 5 -
e e e s < s e
i
. s M
, U N
E L
P
(
a R
E P
P U
/
e s
a S
S 2 M H
N U .
R s -
s i
p a -
A m -
d i
5 u 2 a.
l f
L e !
4 e es s
i i
t e m
_. r T u n
T ' e S l E _
p T _
. r e
p p
_ u 2
/
5 2
4 e
r 6
t s
e T
A Y A .
5 T P L A
I T
E R .
1 e -
r -
u l s i g
e n -
i F e = - -
e=
o
' .! ll l
. = * " : ,
- ~ -
h E
L D
N U
B [ ( .f l{
I S
p r S
2A f M J N \ y U g R
qfI .
s s
a m
5 2
g .t l
t t
s e
d i
u 4
t h . s e l e f 6 T t l e
s d n
E u I
b 2
/
5 2
4 6
t s
. e A T Y -
5 A
P ' _ T A = L L T
- E R
2
- e
', r
* u g
i F
0; 0
1
=
I ; !l 1
l .
l1 It 9
o s . N O 6
N U
N E
L P
n i
8 3
20 N .
s U
R a'M s
a m
d i
u l
5 f 2 ,
m 4
6 4 =
1 n
u T
S E
4 l e
p T . r e
1 w l
o i _
Ww 2
/
5 2
4 6
t A s Y
u 5 e .
P T ._
A I A _
E T R L
- T 4 3
4 e _
r u
g
. e
_ i e-ni
_ I! _
!! l
and enthalpy from either the donor cell or a user-designated cell. These options were included in RELAP5YA because of calculational difficulties which arose when the Moody model was applied using stagnation conditions from the donor cell. The difficulties arose because of differences between the Moody calculation of the slip ratio and the slip calculated from the basic equations in RELAP5YA. Code assessment work by YAEC 13 showed these
- difficulties can be overcome in most cases by using the static conditions from the donor cell and this is the option strongly recommended by YAEC.
When the Moody option is selected, the best estimate RELAP5YA (i .e. , '
RELAP5/ MODI critical flow model) is used until a void fraction of 0.05 is -
reached. For void fractions greater than 0.05, the Moody model is used. r Due to the large drop in critical velocity when the fluid changes from subcooled to two phase, a transition region, using an underrelaxation __
scheme, is used in the void range of 0.01 to 0.1 for both critical flow models. When the Moody model is used, the code uses an artificially large interphase drag coefficient to prevent numerical oscillations since the calculated slip ratio is very dependent on the interphase drag. Hence, with the Moody model, the code inherently uses slip ratios of near unity instead of values calculated from Moody's theoretical slip model. YAEC noted, in response to question Q.IX.17,6 that RELAP5YA calculates lower slip ratios than Moody's model resulting in a conservative estimate of the system mass loss since more liquid and less vapor goes out the break. It also decreases the system depressurization rate, compared to using the Moody theoretical slip ratio, and thereby refills the system more slowly because of the reduced ECC flow rates. This change is judged to be implemented properly.
The RELAPS/ MOD 1 best estimate critical flow model is available as a user option in RELAP5YA. The critical flow model is that developed by Ramsom and Trapp primarily for calculation of the mass discharge from
~
system at breaks or nozzles. The model can also be used to calculate the existence of choked flow a internal points. The RELAP5 critical flow model has under gone considerable assessment as part of the RELAP5/ MODI and -MOD 2 assessment. Work at both Sandia National Laboratory and INEL has shown that the critical flow model yields good results.
13
1 YAEC presented two phase critical flow assessment calculation
-comparisons:for the Marviken Test Full-Scale Critical Flow Tests I8 using the recommended , Moody model and'the best' estimate RELAP5YA critical flow ll model. Due to inconsistent RELAP5YA predicted results, addressed in Reference 7 (' Question 0.IX.9), YAEC presented rerun prediction calculations -
using the Moody model. The results showed the> calculated Moody break . flow exceeded the measured flow.by approximately 15%,.the vessel pressure decreased more rapidly.than the test data, and that_the system voided more
-rapidly than observed in the test. The rerun results, using the best-estimate critical flow model, showed good predictions of the test data.
The Lspecial case-of break flow with stratified flow in the upstream volumes'was addressed.in Reference 7-(response to Question 0.VI.7). .For '
this type of -flow, the break flow void fraction is modified to reflect the relative position'of the stratified liquid level in the upstream volume ard the mechanism of vapor pull-through.or liquid entrainment. This' approach
'is identical to the model used in RELAP5/M001.
The review-of the Moody break flow model and the supporting assessment calculations indicate that the model has been properly implemented in RELAP5YA. Using the recommended static pressure and enthalpy option,'where donor cell velocities are' low, provides conservative break flow calculations relative to the Marviken test data. For the same reason, the results also confirm that the use of the recommended option of static pressure and enthalpy with the Moody model provide conservative'results.
Additional'information on the implementation of the Moody model in RELAP5YA is presented in Section 3.2 of Reference 1.
2 2.3 Jet Pump The RELAP5YA jet pump model is intended to predict jet pump behavior
} .
l during normal and abnormal operating conditions including LOCAs. The model was developed because the components available in the base-code, RELAP5/ MOD 1 could not predict such jet pump behavior as momentum mixing and flow dependent energy loss coefficients, which is essential in the 14
analysis of BWR reactors. The model developed is specifically intended to q represent the hydraulic behavior in the throat region of the jet pump by properly accounting for the momentum mixing effect for normal operation (i.e., positive drive and suction flow) and by developing a means of
. describing mechanical energy loss coefficients during off-normal conditions. The diffuser and tailpipe regions are modeled by other components available.in the code.
~For normal operation, momentum mixing.of the drive and suction streams is accounted for by applying integral mass and momentum conservation equations to the-throat region. For off-normal conditions, the integral mass and mechanical energy conservation equations are used to identify the appropriate mechanical energy loss coefficients (i.e., flow independent and flow dependent). The complete details of the jet pump model development
'are presented in Section 3.3 in Reference 1.
3
.The new RELAP5~YA jet. pump model was assessed against the one-sixth-scale INEL jet pump tests performed in the LOFT Test Support Facility. Both steady-state and transient test data were compared.to calculated predictions. The steady-state tests allowed the comparison of differential pressures between various locations and varied subcooling.
The transient results allowed comparisons of two phase jet pump behavior.
To compare against the numerous steady-state INEL jet pump data, RELAP5YA results were obtained over a wide range of M values (ratio of suction to drive flow) by simulating a slow transient in which the drive flow was held constant and the suction flow varied. The inputted loss coefficients for the modeled jet pump are given in Tables 2.3-2 and 2.3-3 in Reference 1. Calculated predictions for both positive and negative drive flow rates showed generally good results which predicted the data.
Yet, during the review of the jet pump model by the NRC in Reference 10, a question (Q.II.1) concerning the suction area corresponding to the assumed suction stream velocity was submitted to YAEC. YAEC stated that the suction stream velocity was " based upon the horizontal projection of the 1
15
.l-area of the ' frustum of a cone. indicated as 'S' on Figures 3.3-2 (in
- Reference,1)','l see Figure 4. '_During the review of the model, YAEC determined that the area.of theLsuction junction was better. represented by the area of the frustum of the cone. This resulted in a new set' of loss
.' coefficients (Tables 11.1.1 and II.1.2 of Reference 10) and a reevaluation of the jet pump model. The new results indicated an even better match to the; experimental data.
Two INEL transient blowdown tests using BWR-type conditions were used to compare against RELAP5YA calculated transient results. One. test (Test 1)= simulated the blowdown of an intact loop = jet pump response while the other (Test'2) simulated a broken loop jet pump blowdown response.
These calculated test predictions showed larger than expected deviations from the test: data, especially for Test 1. Specific areas of concern were reflected in Questions Q.III.4, 5, 6, 7, and 8 i Reference 10. In response.to these questions, YAEC attributed the poor performance not to
~
the. jet pump model but rather on the system piping model. In this piping model, YAEC did not include heat structures, which affected the total energy of the system, did not feel confident that test boundary conditions could be represented' accurately based on the uncertainty of the available data, and did not feel they had optimized the piping nodalization to handle the momentum terms correctly in the complex piping connections. Based on this utility response, these calculations do not qualify as a basis for assessment review and, as such, only show trends and areas of potential input modeling concerns which must be avoided in LOCA analysis models.
A better assessment of the jet pump model behavior under transient conditions is given by TLTA Test 6425/2. This test represents integral large break. system behavior with average core power and ECC availability.
Comparison of jet pump behavior for this integral test 10 , shown in Figures 5 and 6, indicate good agreement between the predicted jet pump ,
behavior and the data.
16 l
L_
I p.
1 i
\
g s
n h- '
h .
m r l n y i m
JETPMP h
e __
e r v v m
c SNGL VOL ;
I i
O h SNGL VOL l Figure 4. Jet pump model.
)
17 j l
I il i\ jl i! i aA 1:
fi,'
e s
e s e " ,.
g e -
t a
r w
l o
f i
l s
O s L a 2
F 2
%' p m
P " u m
N U
J p R t l e j
p t
e o 5
s u o k
I
.tM E l
4
~- I n 6 T e k
T S -
- b o
r E ..
T .
2
/
g 5
. 2 4
- 6 A t Y
5 P
A W .
. T s
e L A E .
T R L T
p V V
5 Ii I ieEI I li\
11 I I t I I 1I I 'l
, . e 7
- r u
g V i .
F Q~ " .
I!a s
, III1
j 1!i l, J__
_g1 gi s .
s . ,e s
, i,
+
- o. * -
f N.
u . _
a e _
u" t a
c.
r .
o w
n o
, l _
s f i
s _
t t - t o -
s 2
l F
(
y,-
a m
p 1
j l-l m _
'e I P
il l s u p
N J )i U A ,
,Y t Y
R S i.I e I P f j A
I p
E o R l r l o
u u
5 \
e t
t 2 !;1 t f c V e 4
l a
6 / l)[
I.5; T t n
T S
i
, 2 E ,
/
T ,E" 5 2
4 l
l 6
\
\
t s
', W T e
4' r
A T
j4 ' a
' L
',1 T 6
e r
' ] u g
I f.
x o i
. F
. s i
= _
11;E G
l1I1l!ll ,
lJ Based on the overall good agreement of the subcooled steady-state INEL l test data.and two phase transient test predictions of TLTA test data, the- ]
~
Jet pump model. has been properly implemented in RELAP5YA.
{
2.2.4: N_ucleate Boiling Initially, RELAP5YA used the Chen' correlation 20 gg. calculate forced-convective or' nucleate boiling heat transfer coefficients'and the modified Chen' correlation.to calculate subcooled nucleate boiling heat transfer c' coefficients. YAEC's assessment of the Chen correlation indicated that the heat transfer coefficient was underpredicted at low quality and high mass flux. YAEC reviewed the data used to develop the Chen correlation and found that the data of=Schrock and Grossman 21 was more applicable to PWR applications. This data was taken at low pressures (60 to 380 psia)'and high void fractions. Conditions in BWR reactors include subcooled and low te
. quality saturated conditions up to 1000 psia. To improve.the nucleate boiling heat transfer. coefficient predictions, YAEC added the Thom correlation 22 to RELAP5YA. -Th'is correlation is used for void fractions less than 0.80 and the Chen correlation is used for void fractions above 0.90. For' void-fractions between.these values, an interpolation between'the two correlations is used. The new nucleate' boiling heat transferL algorithm is described in Section 4.1 of Reference- 1.
-Thissnew nucleate boiling model was assessed 13 by comparing the results from RELAP5YA to Bennett single tube tests 23 at 1000 psia, using a wide range of mass fluxes (0.2E6 to 1.0E6 lbm/hr-ft2 ), various inlet subcooled conditions (49 to 78 Btu /lbm), and a wide range of void fractions (0.0;to 0.99). These comparisons showed that the measured temperatures were more accurately predicted at low void fractions using the new model '
than when.the Chen' correlation was used by itself. At higher void
~
fractions the'RELAp5YA forced convection boiling model (Thom-Chen) overpredicted.the wall temperatures.
i 20
Based on the. good low void fraction-(i.e., less than 0.8) temperature i
predictions of the RELAP5YA nucleate boiling model and the conservative predictions at higher voiding, the model provides reasonable results in RELAP5YA.
-2.2.5 Critical Heat Flux As a' result'of YAEC's code assessment work, they felt more accurate critical heat flux (CHF) predictions were needed. The new CHF user option uses'two correlations to cover the conditions expected in a BWR core during a LOCA. At high mass fluxes, a modified form of the_Biasi correlation.is used. The modified Biasi correlation 24 was the result of extending the use of the.Biasi correlation'to bundle geometries. At low mass fluxes, the Griffith-Zuber correlation 25 .is used. For: intermediate mass fluxes, a linear. interpolation between the two correlations is used to obtain the critical heat flux value. For both correlations, a minimum CHF value of 2
1000 W/m is imposed. In addition, a critical void fraction criterion is used so that wall dryout is assumed to occur when the void fraction exceeds the' critical void. ' Details on the new CHF option are found in Section 4.2 of Reference 1.
YAEC assessed the new critical heat flux option against steady-state CHF tests performed at the Columbia University Chemical Engineering Research Laboratory,26 General Electric,15 and Oak Ridge National Laboratory. The tests chosen for assessment were selected to be representative of BWR fuel rod bundles and to cover a variety of thermal-hydraulic conditions with different axial and radial power distributions. YAEC stated (response to Q.VII.11 of Reference 6) that the new' correlations in RELAP5YA were best estimate CHF correlations.
Five tests from the Columbia test program were used in the assessment. The RELAP5YA results showed the CHF was calculated within one computational cell of the measured location for all the tests. For two of !
the tests, the calculated CHF location and quality were higher than the 21
_.._._.-_..________o
1
, :n ir
. measured, data. YAEC attributed this difference to:the fact that the high and low-power; rods in the test section were represented by one average rod: )
.in.the RELAP5YA model. They stated that CHF would have been calculated at a lower quality. and l'ocation had a high power rod been modeled.
~
-YAEC also assessed the new CHF option in RELAP5YA against data from THTF at 0ak Ridge National Laboratory.27 Comparison of the calculated *
.and measured CHF locations and qual _ities showed the calculated location and 1
quality.were lower than the measured data in all three assessment runs. )
i The GE CHF tests-were performed in the Nine Rod Test Section. For the four tests analyzed, the calculated CHF location was higher than the measured location in three of them. In the fourth test, the calculated CHF i location was' lower than the measured location. In terms of the calculated and measured quality at CHF, the CHF quelity was accurately predicted in two of the tests, was low by.5% in one test, and high by 16% in the last- )
test.
1
.The new CHF correlations in RELAP5YA were assessed against a wide.
variety of test conditions. The results, summarized in Figure 7, showed that the correlations did a reasonable job of calculating CHF for the tests analyzed 1(within data uncertainties). It is concluded that the new CHF 4 model is reasonable for LOCA analyses. j ,
i 2.2.6. Rewet and Quench j YAEC's review of RELAPS indicated that it lacked the models necessary
]
to accurately calculate the reflood portion of a LOCA in a LWR, As part of its code development effort, YAEC added a rewet and quench model to j RELAP5YA as a user option. The rewet and quench model consists of four submodels: rewet and quench initialization, quench front velocity, heat ,
. transfer enhancement, and quench front advancement. Each of these submodels is discussed in detail in Section 4.3 of Reference 1.
I 22
1.0" ,
0.8 ,,.
~
a* .
.D.
'e 0.6 ,
o -
1
,5 .
w u
1 0.4 - -
1 ?
5 8 u ,,
O
. .W.. j 0.2 - O Columbia -
A GI A C A a l
exst ,
0.0 ., ! , t , I . I '
~
0.2 0.4 0.6 0.8 1.0 r.rperimental crt:1 cal Qulity Figure 7. Comparison of calculated and measured critical qualities.
23
The new reflood models in RELAP5YA have been checked against two THTF ;
tests for bottom quench'and.'an integral TLTA test which had both top'and bottom rewet. .The THTF. tests will be discussed here. The assessment of-the.the reflood models based on the TLTA test data will be' discussed in' .
Section'2.4,2 ,
The two tests used to assess the reflood model were from the THTF at Oak Ridge National. Laboratory, Tests 3.09.100 and!3.09.100 The initial conditions for these tests were pressures of approximately 565 psia,-a reflood rate of 4.8 in./s and a linear heat rate of 0.62 kW/ft for Test 3.09.100'and a re<1ood rate of'2.3 in./s and a linear heat rate of 0.31 kW/ft for Test 3.09.100 The results'of-the assessment calculation for Test'3.09.100 showed the code.did.a. good job predicting the system void fraction profile as a
~
l'
. function of time. The calculated bundle mass inventory, the core collapsed' liquid level, and the quench front in the core compared well'to the test data. When the calculated results differed from the test data'for these parameters, they were in the conservative. direction (low). Rod temperature L profiles at four different points in time were compared to the test data.
L ' This comparison showed.that, except for the effect of the grid spacers, the code did a good job of calculating the temperature profiles. A comparison of the calculated and measured quench times showed that the code calculated rod quench was approximately the average of the quench times at a given core elevation.
t For Test 3.09.10Q, the' comparison of the bundle mass inventery, the core collapsed level, and the quench front were not.quite as good for this i test as for Test 3.09.100 but the differences were still within the experimental uncertainties. ,
RELAP5YA did a reasonable job of predicting the bottom reflooo tests at THTF. Top down reflood, discussed in Section 2.4.2, was also predicted in'the integral TLTA test. For the important parameters, where differences l 24 1
existed between code calculated results and test data, the code tended to predict' conservative results. Therefore, the new reflood models in RELAP5YA are recommended for use in LOCA calculations.
2.2.7 Multiple Surface Radiation
- Because radiation heat transfer can play an important roll during certain portions of a LOCA event, and since multiple surface radiation heat transfer was not included in the development of the RELAPS/ MODI code, YAEC included it in the RELAP5YA code. Any heat structure in a RELAP5YA input model may be specified to participate in radiation heat transfer by the user and the fluid is assumed to be transparent to the radiation process.
The radiation model uses a lumped-system approximation for gray diffuse surfaces forming or contained in an enclosure. The basic assumptions used in this medel are: the fluid does not participate in the radiation process; reflectance is not a function of incident or reflected direction; surface temperature, reflectance, and radiosity are uniform over each surface; and view factors and emissivities are constant. The details on the implementation of these assumptions in RELAP5YA are discussed in Section 4.4 of Reference 1.
YAEC performed several assessment calculations to ensure that the radiation heat transfer faodel was properly implemented in the code. Two calculations, representing relatively simple problems, were performed and.
compared to known analytical solutions. Two additional calculations were performed with and without radiation heat transfer to examine the effect of radiation on the temperatures within a simulated fuel bundle.
For the simple example problems, comparison of the RELAP5YA calculated results and the analytical solutions showed the two agreed exactly. For the two calculations run with and without radiation heat transfer, the model represented a BWR-type fuel bundle experiencing a cooldown and a hentup. The calculated results showed the fuel rod temperatures were lower and the channel wall temperatures higher with radiation than without it.
This type of difference was expected with this model.
25
The res'ults of the limited number of YAEC radiation heat transfer assessment calculations' indicate the'model adequately calculates the radiation heat fluxes based on'a given view factor matrix representing the .
input model: and that..it is properly implemented in the code. Use of-this model is contingent on the justification of the view factor matrix -
calculating code'(if.one is used), YAEC currently proposes to use HUXY to
~
calculate the view factors for BWR LOCA. analysis.
2.2,8 . Heat Transfer' Logic Options Appendix K to 10CFR50 requires that a code must prohibit or lockout
~ return to nucleate boiling heat transfer once a CHF is predicted at an axial fuel rod location during blowdown. This requirement forces the use of a degraded: heat transfer calculation during the blowdown phase of a LOCA' even-though local conditions may allow a rewet and a return to nucleate -
boiling. A return'to nucleate' boiling' lockout option was added to RELAPSYA to meet this Appendix K requirement. The approach taken in RELAP5YA was to have the user input the appropriate flag to activate this option and, at i the 'same time, to supply a multiplier (XMNB) with a value between 0.0 and 1.0, XMNB multiplies the calculated nucleate boiling heat transfer coefficient, thus, degrading the heat transfer coefficient used. The use of a factor of 1.0 will result in using the nucleate boiling heat transfer correlation,directly.
The return to nucleate boiling lockout is activated by having the CHF exceeded at a particular heat structure surface.- If the heat transfer logic subsequently determines that a return to nucleate boiling is possible, the code multiplies the normally resulting nucleate boiling heat transfer. coefficient by the factor XMNB. This calculation results in a ,
. degraded heat transfer coefficient. The nucleate boiling lockout can be overridden via the quench model determining that the node has quenched or .
by manually deleting the option on a problem restart. Use of 0.05 for XMNB is suggested ~ to meet the Appendix K requirement for lockout of return to nucleate boiling. The selection of a larger value for XMNB should be justified in any licensing submittals.
26
. - - = _ - - _ _ - _ _ _ - _ _ - _ _ _ - _ _
Appendix K also requires that a code must' lockout return to transition boiling during the blowdown phase of a LOCA once the clad superheat
'3-
-exceeds 300*F. YAEC has added the appropriate heat transfer' logic to RELAP5YA to allow the user to meet this Appendix K requirement. The user
.; _ sets an input flag to activate this option. Once the code has calculated a clad superheat.in excess of 300 F, the code does not aliow a transition
- boiling coefficient to be applied at that heat structure surface until the transition boiling lockout is manually deleted on code restart cr the quench model calculates a node has quenched. 'Once the lockout is calculated te occur, only film boiling heat. transfer coefficients are applied at that heat structure surface. If local conditions indicate that either the transition or nucleate boiling modes should be used, then the logic will-extrapolate the film boiling correlations into these regions to yield a degraded heat transfer coefficient. Reference 1 discusses both of those lockout options. in more detail in Section 4.5.
To assess that the lockout logic was properly implemented in the code, YAEC performed a number of checkout calculations. These are described in Section 3.4 of Reference 13. While it is not necessary to discuss the details'of'these calculational results, it will be noted that the results '
indicate.the return to nucleate boiling lockout option and the return to transition boiling lockout option work properly and are correctly
' implemented.in the code.
2.2.9 Fuel Behavior Models Fuel behavior models have been added to RELAP5YA to enable the fuel rod behavior to be calculated during a LOCA. These models include a fuel rod internal pressure model, fuel rod deformation model, fuel rod gap heat
. transfer model, and zircaloy-water reaction model. The zircaloy-water reaction model is based on the Baker-Just model as required by Appendix K.
~
The fuel behavior models can be used for either best estimate or evaluation model analysis with the modification of user inputs. All fuel behavior models were basically adopted from the T00DEE2-EM computer code which is part of the Yankee Atomic Generic Water Reactor Evaluation model.14 l
27
._ . . _ _ . . - - _ = . . _ _ - _ _ - _ - - _ _ _ - - _ - _ _ - - _
j S11ght' modifications were made to- the cladding deformation, internal gas
~
pressure, Land. rupture models to reflect the models proposed in j 20 andto'easeimplementationinRELAPSYA[
~
NUREG-0630 Section 5.1 of Reference.1 presents the detailed. fuel behavior model description.
The internal gas pressure model calculates the time-dependent pressure
-in the fuel rod plenum and gap region. The model assumes the gas
- temperature to be the tea.perature of the adjacent coolant.plus an offset.
Per Reference 9 (response to Question Q.II.2), Yankee Atomic has proposed a 2'F offset for LOCA calculations to be consistent with T00DEE2-EM.
LHowever, based on their improved fuel red internal pressure'model in-REU.P5YA, a 10*F. offset II was chosen to be consistent;with the 29 30 FROSSTEY and FRAP-T1 codes and to provide additional conservatism.
The higher' offset margins will provide higher temperature calculations.
The effect of the offset'is ' considered small but due to a lack of assessment of varying offset margins and since this offset margin is an
' input parameter licensing applicable justification should be provided supporting any' offset less than 2*F.
The fuel rod deformation and rupture model consists of a. fuel pellet thermal expansion model, cladding elastic deformation model, clad rupture temperature model, and a cladding plastic strain and flow blockage model.
The fuel pellet ~and cladding thermal expansion along with the cladding l pressure induced strain model were cotained directly from T00DEE2-EM. The cladding rupture temperature correlation was obtained from NUREG-0630 and the cladding plastic strain is identical to that used in T00DEE2-EM. The flow blockage caused by cladding rupture is consistent with NUREG-0630 l
specifications for licensing analyses or can be user supplied.
L 1 .
L The gap heat transfer is identical to T00DEE2-EM with the exception of the modification to neglect conduction through contact between cladding and ,
fuel. This modification results in higher fuel rod temperatures.
28
I,'
L y -
Finally, the metal-water reaction, being almost identical to that in T000EE2-EM, calculates zircaloy-water. reaction on both clad surfaces. In
' addition, once this. option is selected, metal-water' reaction is calculated throughout the transient since no metal-water reaction threshold
. temperature is included.in the model. In,accordance with Appendix K, the 1 rea'ction'model is assumed never to be steam limited. One of the original
- ., ' assumptions in the fuel behavior models included neglecting material l changes in the cladding as oxide was produced. After reexamination of the model (response to Q.VII.40,. Reference 7), YAEC agreed to include the oxide conductivity'and heat capacity.when cladding is oxidized.
To assess the RELAP5YA fuel behavior models and.to ensure that they were correctly implemented, YAEC compared RELAP5YA calculated results to
- T00DEE2-EM calculated results. The first sample calculation represented
.the adiabatic'heatup'of a single PWR-type fuel rod and was executed to assess.all of the fuel behavior models simultaneously. This comparison showed the codes produced: essentially the same~results. The differences between the calculations were not significant and could be explained by~
known differences between the two codes and by differences.in matching the input description for each code. The major difference between RELAP5YA and T00DEE2-EM was in the calcu!ated metal-water reaction rate at high cladding temperatures with RELAP5YA results showing a slightly higher heating rate than those from T00DEE2-EM.
To ensure that the algorithm for calculating the zirconium-water reaction was implemented properly, YAEC compared RELAP5YA calculated results to hand calculations. The two results compared exactly, indicating the model was implemented properly in the code.
Because most of the RELAP5YA fuel behavior models came almost directly
. from the currently approved T00DEE2-EM licensing code (some with slight YAEC modifications), and since RELAP5YA assessment calculations yield essentially the same results as T00DEE2-EM, the models are recommended for LOCA analyses.
29
]I '- ,
,, i e
=
0y
,E ,
2.3 phenomena Models Important to BWR Systems Areas of important;modeling cor.cerns for BWR application include the steam;separatorsk the.channe1Lboxes surrounding the fuel elements, the .iet
.i . pumps. (the RELAP5YA:model was' aiready' described and asse-ssed in. --
6 Section 2.2.3), the counter-current flow phenomena, and the prediction of condensation ~and vaporization.
- 2.3.1..Re'amSeparatorModel While.a' steam:sepirator model (a SEPARATOR component) is included in
.the RELA P 5YA code and is described in Section 3.2.8 in Volume 2 of Reference 2, YAEC chose not to use this model. -Instead the steam separators are modeled as a combination of several components: BRANCH,
. control valves, control logic, and time-dependent . junctions and. volumes.
.This multicomponent model was used in modeling the 129 steam separators in the Vermont Ya'nkee'LOCA licensing method model and in the' assessment-calculations' for the Two-Loop Test Apparatus (TLTA). The multicomponent.
. model, when using all the components. of the model, represents a perfect steam' separator and is1used to calculate realistic steady state conditions within the vessel'. 'Upon transient initiation, the perfect steam separator modelis isolated, resulting in two phase flow passing from the separator BRANCH component volume into the intermediate and upper plenum. This latter modeling of the separator is referred to as the BRANCH separator model in this report. Concerns as to what effect this perfect steam separator isola'. ion has on a 'LOCA transient were addressed by YAEC in two small break studies (SBLOCA-EZ and -E50), which are described in Section 4.4.2. These two analyses showed that the multicomponent separator produced satisfactory results and is recommended for BWR LOCA analysis.
E Since no applications or any assessment of the SEPARATOR component model were presented for NRC Peview, the SEPARATOR component model in RELAP5YA ,
should be restricted and not used in any LOCA analyses.
30
1 1
1 l
2.3.2' Channel Box- l A typical BWR core consists of many fuel rods surrounded by channel boxes which are not present.in the open core PWR design. No special a
. modeling modifications were required to model these channel boxes since the j RELAP5YA' code can already model them by using heat structure components l
- - thermally connected to hydraulic flow channels. Based on YAEC's j modifications, the code can also calculate the effect of thermal radiation
)
on a multiple heat structure configuration (see Section 2.2.7). Hence, it l can readily account for rod to channel. box radiative heat transfer as well !
as hydraulic heat transfer (see Section 4.3). ]
l 2.3.3 Counter-Current Flow Limitation !
I The CCFL phenomena can_significantly affect the thermal-hydraulic behavior of BWR fuel bundles, especially for quenching transients. Since ,
the LpCS liquid is sprayed directly in the upper plenum of a BWR and must )
-drain into the fuel bundle and bypass region, the CCFL phenomena can- I considerably reduce the coolability of the bundle via holdup. Other areas where this phenomenon may be important are at the SED and perhaps at the
. throat of the jet pump due to LPCI injection.
.RELAP5YA does'not specifically provide any special treatment of CCFL due to its complex flow geometry dependence but YAEC provided evidence that the current vertical, low' flow regime drag force calculation model provides conservative estimates for this type of bundle response. The interphase drag models for the vertical flow are discussed in Section 2.2.1, which !
indicates that the new interphase drag models will predict more holdup in the upper plenum as well as allowing more draining from the bundle / bypass. l Though'not truly a best estimate behavior, both lead to higher PCTs. j t
2.3.4 Condensation and Vaporization 4
L This section will review the RELAP5YA condensation and vaporization l models'. 'One of the areas of concern in modeling small break LOCAs is the i ability to calculate condensation heat transfer rates accurately as !
1 31 l
identified in NUREG-0737,31 Item II.K.3.30. Since this review applies to full spectra break applications, the need to confirm this' feature of the small break model against applicable experimental data is recognized. 3 The condensation / vaporization models in RELAP5YA are identical to the models used.in RELAP5/ MODI with the exception of a liquid superheat value ,
of 2 K.for incipient vaporization in RELAP5/ MODI and 0 K for RELAP5YA.
These models are discussed in Section 2.1.3.1 of Reference 1. Both the '
' condensation and vaporization models are empirical and are based on the i
work of Jones'and Saha.32 The empirical constants in the vaporization model were determined mainly from depressurization experirient . The large depressurization rates in these. experiments resulted in relatively high vaporization rates that are reflected in the empirical constants. While the model has shown reasonable interphase drag (Section 2.2.1) and critical heat flux (Section 2.2.5) responses, the use of these empirical constants, which were based on depressurization data, raises a question concerning the applicability of this mccel to processes dominated by wall heat transfer.
YAEC addressed this issue during its code assessment work. In most cases
.the model worked reasonably well; however, at low flows and high quality, representative of nonequilibrium film boiling heat transfer, the vaporization was overpredicted. However, in this film boiling heat transfer regime, RELM3YA also consistently underpredicts heat transfer coefficients resulting in overpredicted surface temperatures (see Section 2.4.1).
The high vaporization rates also affect RELAP5YA's ability to calculate vapor superheat. YAEC noted in Reference 9 (response to .
questions 0.VII.25 and 26) that due to the high mass transfer coefficient, the code is not able to calculate vapor superheat in a control volume -
unless the volume void fraction is 1.0. For this to happen, the nodalization must be detailed enough to show a two phase level and the heat i
4 transfer above the mixture level must be sufficient to vaporize all 32 l
entrained droplets. However, inability.to calculate superheated vapor is only'important for, constant pressure, boil-off transients, which are not present in most BWR LOCA transients.
, The effect of.nodalization on the condensation calculated'during ECC injection was discussed in Reference 9-(response to question Q.II.5).
. While no.-specific BWR related nodalization study was performed, YAEC presented the results of a nodalization sensitivity calculation for the accumulator injection in LOFT Test L3-1.12 Based on ECC applications, the nodalization sensitivity study was found relevant to BWR related applications. The study was performed by repeating the LOFT calculation.
with the nodes upstream and downstream of the injection point divided in half relative to the base calculation. The results of the sensitivity calculation were essentially the same as the base calculation indicating the nodalization used in the base calculation was sufficient to adequately represent the condensation associated with ECC injection. Results of the TLTA Test 6425/2 also support this conclusion.
Based on the information provided by YAEC, the condensation and vaporization models in RELAp5YA'are adequate for modeling BWR LOCA
-accidents. The failure of RELAP5YA to calculate superheated vapor under slow boil-off conditions is not deemed important for LOCA calculations and therefore, the condensation / vaporization model is recommended for LOCA analysis.
2.4 Integral System Calculations RELAP5YA was assessed against several integral tests. These tests J provided information on the integral system behavior under the influence of many interacting thermal-hyraulic phenomena. These experiments included two different fac111tiesi the Thermal-Hydraulic Test Facility (THTF) at j the Oak Ridge National Laboratory 27 and the Two-Loop Test Apparatus (TLTA) at General Electric in San Jose, California.33 l:
33 !
\
l l
i
, 2.4.1R THTF Calculations While.the'THTF was built to investigate system blowdown response under .
large and small break PWR'LOCA conditions, many BWR related phenomena were
. studied. These. phenomena include two phase mixture levels, CHF and -
. post-CHF heat transfer, and rewet and quench front propagation during
~ reflood. A' total of seven' tests were used in the assessment: three- '
steady-state, one transient, one boil-off, and two reflood tests.
The:three' steady-state' THTF tests (3.07.98, K, X) assessed the
' post-CHF heat transfer and the new CHF option. Comparing reported dryout qualities against calculated dryout quality predictions supplied in the response to question Q.VII.13,7 the code did a reasonable job in predicting the CHF . location 'and, in the post-CHF regime, the calculated l-results consistently overpredicted the surface temperature. This result
-was expected since the code uses the Condie-Bengston correlation in the
~
high flow film boiling regime and it has been shown to underpredict heat transfer coefficients.
While the THTF transient film boiling test (3.08.6C) assessed the
. power excursion and subsequent dryout for PWR-type conditions, the assessment of the RELAP5YA CHF model and post-CHF temperature excursions are of practical importance in BWR appl'ications. The RELAP5YA code provided a reasonable simulation of CHF at lower bundle elevations and it predicted the higher elevation temperature excursions quite well.
The THTF boil-off test (3.09.10I) attempted to assess the interphase drag, the CHF option, and the post-CHF correlations at a system pressure of
'650 psia. While RELAP5YA again predicted the occurrence of CHF dryout at a lower elevation, it overpredicted the interphase drag values (based on test section voiding response data). It should be noted that while these ,
boil-off test results were reviewed and support results of the film boiling test, the lack-of knowledge about the test power history and the uncertainty of how it was to be modeled (see question Q.VII.18 Reference 7) must exclude it from being considered as an assessment test case. It was viewed, therefore, only as an example test case.
34
k L
i-The, review of the RELAP5YA predictions of the THTF reflood
. tests (3.09.100 and 3.09.100) was already, presented in Section 2.2.6.
Based on the predicted results of the THTF tests, RELAp5YA provided
. reasonable results involving the effects of many interacting thermal-hydraulic phenomena.
2.4.2 - TLTA Calculations The TLTA facility was built to study BWR-type blowdown responses.
- l. 'YAEC' chose four tests of the~TLTA-5 test series, which approximated a scaled version of a 218 in. BWR/6, for assessment purposes due to the wide range'of phenomena present. These tests included Test 6425/2--large break with average power and ECC availability, Test 6426/1--large break with average power and no ECC, Test 6432/1--small break with degraded ECC and ,
delayed Automatic Depressurization System activation, and Test 6441/6--a system boil-off with final feedwater injection.
l
~Due to the inadequate steady-state capacity of TLTA, the initial conditions for Test 6425/2 using RELAPSYA required establishing the system temperature until steady-state was reached and then ramping the power to the required 5.06 MW. This initialization resulted in an approximate matching of actual initial test conditions. Upon activation of the blowdown valves, the recirculation pumps were tripped and the broken loop pump wasLisolated. The JETpVMP input component was used to model the throat region of the jet pumps and the empirical constants and loss coefficients needed in the input were calculated using guidelines specified for this component. Also, in a desire to run at large time steps, the throat mixing section was lengthened to avoid Courant limitations. Yet, the overall jet pump length was still physically correct. After reviewing.
several questions (Q.IV.1-6, Reference 10) concerning discrepancies between predicted calculations and the test data, YAEC reevaluated this test.
35
- _ - _ - - _ ___.-__--_____e_--_-___----__
?
The_ reevaluation of' Test 6425/2 required renodalization and modification of initial conditions. The renodalizations included correct-i modeling of the. physical size of the jet pump throat,- improved input loss coefficients based on an' improved jet pump model, and a-relocation of a-
- valve closer to the suction side of-the broken loop. recirculation pump'. In-
-addition,-the-initial conditions were modified to better match the power history'of the actual test.
Recalculated results10'for' Test 6425/2 showed that the new initial con'itions d are within the experimental uncertainty of the reported data and' that the -jet pump modifications resulted in an improved prediction of the
, experimental data. No impact of the renodalization to relocate a valve closer to the broken 1oop recirculation pump was observed in the new results.
P The system depressurization response (see Figure 8) was still-
- overpredicted by RELAP5YA. It is believed that this pressure response is a result of not modeling upper tie plate heat structures to correctly predict
~
vaporization of injected ECC. YAEC stated, in response to Q1.13 of Reference 11, that they performed a hand analysis which showed that had
^ these heat structures been included, the release of the stored energy would have accounted for the difference between the RELAPSYA calculation and the
- measured pressure history.
~ Comparisons of the upper plenum, bypass, and lower plenum fluid mass results (see Figures 1 through 3) clearly show that the ECCS liquid tends to be held up in the upper plenum and then drains down through the bypass into the lower plenum. Hence, while the code does not contain a counter-current flow model, it is clear that the interphase drag calculation leads to higher bundle void fractions. ,
Details of the' improved results for Test 6425/2 and the comparisons to .
the earlier assessment results (Reference 13), are presented in Appendix A.IV.1 of Reference 10. It is also noted that the RELAP5YA heat transfer algorithm predicts best estimate cladding temperatures which bound the PCT for this test and that the quench of the upper cladding occurs due 36
lji lI' l1!j
_as a?arJ:*
LNN _
HEE _
OLL
. DPP ,
NRR OEE
. EWP I OP SLU s e s
n o
., p 2 s e
r l 2i 2 P P P'
- e r
u s
s N e U . r R p i23 m u
n e
, l p
5 2 *
. . r 4
_ . e 6' = .
_ w
- 4 o T
S E
T f
If 1
w l 2
/
5 A_
. i
' 2 4
S 6
~
U ,
A T
L T
j - = 8 A e Y r 5 u P g A A i Y L f S E P R A
L E
R b
c 6 ,
a n m =
i 1A 531i>2I l 1lllL
to fluid penetration from the upper plenum. This indicates that RELAP5YA is capable of calculating the cladding thermal response from both bottom reflood and top down quenching.
Additional EM calculations 13 , demonstrating Appendix K required models, were performed for Test 6425/2 using an increased bundle power ,
factor of 1.02, the ANS decay power curves plus 20%, the Moody critical flow model, activated the lockout for return to nucleate boiling and transition boiling, used a departure'from nucleate boiling multiplier (XMNB) of 0.5 (see Section 2.2.8) and only allowed rewet and quench after ECC injection. All results showed expected conservative trends, such as higher rod temperatures, larger break flows, lower bundle mass, etc., when compared to the actual test data.
Code predictions were made against Test 6426/1, which basically represented Test 6425/2 without ECCS availability. The predicted system response for this test compared quite well with the experimental data. For example, as seen in Figure 9, the lower plenum pressure response was virtually identical to the test data. Thus, this comparison seems to support the conclusion reached earlier that the poor system depressurization comparison in Test 6425/2 using ECC indicated that inadequate vaporization was modeled. While the suction side break response for this test compared favorably with Test 6425/2, the drive side brear.
flow response for these two tests was quite different even before the activation of ECC in Test 6425/2. The reason for this test data inconsistency (shown in Figures 10 and 11) is unknown. Since the code predicted break flows well for the suction side and the earlier separate effects tests, no modeling problems are suspected in this assessment.
Detailed discussions of these test discrepancies are presented in +
l Reference 10 (responses to questions 0.IV.7-9).
Test results for Test 6432/1 investigated small break scenarios of BWRs. This test showed that RELAp5YA predicted CCFL breakdown at the side-entry orifice to occur about 100 s earlier than in the actual test.
This early prediction allowed an early refill of the lower plenum and a 38 i
L -. - - - - - - - - - - - - - - - - - - - - - - - --------- --- _
=
m W2x Eww O _J a a
o c. c,,,
E ce cc a Cww w 2 c.
- o c.
mJD a e o
m o c m o N &
- cv - = E
~ L C. E o L
3 m
m 2 2 3 c.
ct; a
~Nm 5 n c u o C c-v2
- L us = O c) nJ c ta
\
- r 3 W s.,
m e N
w a
- o 5a A
- 9 G
~ .
cn o
e 5
~ .m i/ u m
.c
<H Q
Q f,
.< o I
w i
C o e e o o c N = <
==9 (UTsd) IdnSSIEd 39
' 0 A 1 3
h J, .
C C
S
/
i t
u L
L O m.
3 3
F 2 2 E m.
w V o l
I f R k N
U R
D
- b a
e r
u 1
e n
A i l
5 e 2
1 8
+ 0 5
1
)
S c
e i
v r
6 ( d 2
E T
S E
Ay H l
T
/
5 2
4 6
T A
T A L Y T 5
P h
I A 0 E 1 R
e
' 0 r
u A 7 g
T i F
A D _
x _
N
,N jl< i u ki' l
I' l
0
- 1 0 s o 1 5
~
m ::us uGv d' I l 8
t l{l1<l',l! !l lli\! !)
l'
!l I j!lI 0
1 3
a C'
E S
/
H a B
L O 0 . .
L 3 w F 2 l
o n f i
E k V a I e R' r b
N U
D e R a i
n i
l
) e c v e i S r 6 ( d 2 i
!1 0 _
I 4
5 E 1 t 1 M /
6 I 6 T 2 4
T 6 S A E . T I L T
A .
Y 1 5 . 1 P
A e L r -
E u R
- n-,.
0 7
i F
g W _
k A _
T _
A D
y I _
- I _
'- .nll l Lf l
l ' _
l '
1 -
0 .
1 "I
, 0 5 Ok!U !.ng :n3m
>~
{
/
~
E)?
n ..
reduced bundle fluid inventory. This test also showed that the code f ,
overpredictedithe system pressure response due to the overprediction.of.
~
l-condensation effects.1 Similar to the results discussed for Test 6425/2 l
l l: .where the upper' tie plate was not modeled. Results using an artificially l Limposed'ECC liquid temperature (i.e., closer to saturation condition) of-
200 F instead of the actual 90 F yielded better predictions. Hence, these l
'~
'results. indicated.that the. lack of appropriate. heat structure modeling of the test's upper tie. plate, which results in'anunderprediction of the ECC vapor.ization, can be compensated by artificially reducing condensation- .
1 effects.
Compariso'ns'of' calculated results to the TLTA boil-off test 6441/6 assessed the code for heat transfer in a partially covered bundle at decay ;
power levels and : low, natural ' circulation flows. This test again required an initial- transient to obtain the desired initial conditions from which the te'st was initiated. Accounting for the uncertainty.in the initial conditions, as described in the response to question Q.IV.13,10 and the fact that the trends were well predicted, it is concluded that RELAPSYA can ;
adequately predict the theraal-hydraulic behavior of this test.
Predictions of the :hannel void fraction' data also indicate an acceptable j drag model since, at the~1ow flow rates'in this test, gravity and !
interphase drag determines the two phase response. However, excessive clad temperature underpredictions in the post-dryout regime were obtained 'due to !
the code's inability to predict significant vapor superheat until a quality ;
1 of unity is reached. Yet, the inability of RELAP5YA to predict vapor l superheat is only important' for constant pressure, boil-off transients l which are not present in most BWR LOCA transients. Therefore, it is l recommended that RELAP5YA not be used for any transient having a constant j i
pressure, boil-off response. ,
' The above sections addressed the limited number of concerns about the . l TLTA assessment study. Overall the TLTA assessment studies indicated that RELAP5YA was capable of simulating the TLTA thermal-hydraulic responses.
Such items as break mass flow rate, system inventory, and fluid temperatures were well calculated. The system pressure in Test 6426/1 was 42
well predicted and if the upper tie plate had been included in the model, the pressure response calculated for Test 6425/2 would have matched the measured pressure almost exactly. The rod cladding temperatures calculated by RELAP5YA bound those measured in all of the LOCA simulations.
Based on the overall review of the integral system test data
, comparisons, it is concluded that the various models and components in RELAP5YA worked together to provide a reasonable simulation of the integral experiments analyzed. Therefore, the RELAP5YA code is recommended for integral systems BWR LOCA licensing analyses, based on the limitations summarized in the conclusions.
All of the TLTA tests used for assessment were of the BWR/6 ,
configuration. Simulation of these tests confirmed the ability of RELAP5YA to calculate phenomena present during LOCAs in BWRs in general. It is recommended that YAEC include the effects of differences in the BWR/4 and BWR/6 design features in their estimates of code uncertainty for licensing calculations.
l 43 L
- 3. RELAP5YA CODE UPDATE REVIEW m
E, RELAP5YA was developed from the. publicly released RELAPS/ MODI Cycle 18 f
coue. .'During the: RELAP5YA code development,' the RELAP5/ MODI code was
. updated lseveral ' times ~to correct errors or to improve, areas.resulting in >
cycles 19 t'o 29. Cycle 29 was the final version of'the released code.
. Approval offthe RELAP5YA code must include an assessment of the status'of i
1 these. updates.
f To~ assess the status of these updates, YAEC was asked (01.20 of Reference 12) to specify for each update whether'it was (a) incorporated as-
. written by RELAP5 development, (b) incorporated, but modified slightly, or (c)'notzincluded in RELAP5YA. Also included in this question was.a brief description of each update. All updates which were directly incorporated by YAEC in RELAPSYA are not discussed further. Only th'e .latter two categories present areas requiring further discussion.
For those updates which were slightly modified before incorporation into.RELAP5YA, YAEC indicated that these modifications were either required
.to make the correction ~ compatible with the development'of RELAP5YA or
' contained known errors, based on hindsight, which were corrected in later updates.
Those updates not incorporated by YAEC fell into several classifications. First, -YAEC decided to exclude those updates which were
. clearly improvement modifications, such as error diagnostic modifications, attempts to reduce mass error calculations, and printout edit modifications. Second, those updates with known errors and which were corrected in later updates YAEC chose not to include in RELAP5YA. Third, ,
.YAEC did not include correction updates which, by the very nature of the RELAP5YA development, had already been provided or were not applicable. .
,This included exclusion of corrections to CHF correlations not used in RELAP5YA and coding errors already encountered and corrected in preliminary calculations. Finally, YAEC excluded certain correction upcates since they 44
p l
l
' felt that those'model specific applications were not going to be.used in j BWR-LOCA calculations. Therefore, the exclusion of these specific model correction's will restrict certain applications of the code. These specific.
.modeling areas'are:
'1. Cannot use control . variables as power input to a heat structure
- 2. Cannot use reactor kinetics when no power is generated from gamma decay heating i
e 3. Cannot use VALVE component with form loss coefficients-1
- 4. ,Cannot use control variables with reactor kinetic feedback.
With the exception 'of those specific correction updates which YAEC '
l elected not to include and, hence, which will restrict certain applications l
of the code, RELAP5YA contains the intended correction modifications from l
RELAp5/M001 Cycle 18 through 29 which ensure that known errors were !
. corrected.
l l 1
1 1
\
l 45 i
p.
L, E i
+ 4; NSSSL BWR LOCA LICENSING ANALYSIS MODEL-YAEC submitte'd"a BWR'LOCA licensing analysis method3 for.NRC approval. 'This method is applicable'for the Vermont. Yankee Nuclear Power ,
Station to perform full break spectra and cycle-independent analyses. The
- method uses 'two RELAPSYA input models. .
i The: Vermont Yankee Nuclear Power Station is a BWR/4,with a Mark I
. containment that bdgan commercial operation in 1972. .The reactor has a rated power of.1593 MWt and. design-power of 1664 MWt. The plant produces 540 MWe'at rated conditions. The reactor core includes 368 8 x 8 fuel assemblies, Leach having -144 in. offactive height, containing enriched.
002. Compared to other BWR/4 type. reactors, Vermont Yankee has a' larger core bypass region due to the relatively large vessel; diameter (205-in. ID) and the 'small number of fuel assemblies.
A RELAP5YA base input model was created by YAEC specifically for Vermont Yankee to perform Yankee NSSS.LOCA licensing analyses. This model, shown. in ' Figure 12, was developed using the various hydrodynamic and heat structure components available in the' code. The model includes the reactor vessel with. internals and core, feedwater lines, main steam lines, ECC systems, trip and control logic" systems, point reactor kinetics, and each of two recirculation loops including an associated jet pump representing a bank of.' ten pumps. The modeled jet pump forwerd and reverse loss coefficients were determined from the procedure outline in response to question Q.II.9 of Reference 10. A comparison 3 of the Vermont Yankee-licensing coefficients to the scaled EG&G jet pump and the TLTA jet pump coefficients was found acceptable. All ECC systems are modeled and have setpoints'with realistic or conservative trip values. Since all important internal structures are modeled with heat structures in the licensing
~
model, the. effects of vaporization and condensation will be adequately addressed by using realistic temperatures for ECC injection. For example, the high pressure injection system modeled actual conditions of 14.7 psia and 100 F. Using higher liquid temperatures (i.e., nearer saturation) will i
46
[ ".'.
$61 M 661 .l 659 l-
$$7 g-IA0 @ 637 )
655 l 3 653 234 F~7 '
- $=l 8 222 8 223 bl
==- - m WyLe.
2 I j d8 _f!.I 63 l L _ ,W ^ - N' p-% ,
T o
,t- f su C.
' I-
- . 256 22s 256
, r-I226':
, "T- T ~. 7
$ $71
{260 l 220 l 260
- t F4 l 570 %
t *== 9 .
264, E 264 h $50 q,
$49 3- 4 . _ _
4 ?.
266 r M. %j I 2:6 _
266 : : :
721 _.
- s
- 6 720 . Ws & 6 __
l268) 343 g I l 268 j 9 9 9 >*>
270 J270 J 2 2 q,
274 y [ [ [- [* 7 274 4
- a <. / 602 1 g,
,s ~ ~
1 't FnLIb
~ "'
g 1 is lE 27*
I I 3 60 (120l14G(160j {
? l 'Cl! _] 7
- =
- O , -
5 1 g I l 006 j $ ~~
I E T L
~T T e
+ 380-1 }"",
N .
4 330~; 4 Q 7 00*
I 006 / 3 2,
XYhf s 4
?
ss2 d*'
S S
&l '
758
?
e tao T A
? i g -
t
+-
- E
?
a g a
360 362
~ 312 sto ass
.sca [
Figure 12. Vermont Yankee NSSS model nodalization diagram.
47
reduce the effects of condensation and result in more conservative results. .The: low pressure ECC systems model conditions of 14.7 psia and 7 -165 F. The condition ~of 165 F was chosen to' account for the possible accident scenario of released vapor heating of the pressure suppression pool. Where applicable,'the largest delay times are imposed on the ECC -
system trips to provide conservative response behavior.
The NSSS model provides for the modeling of large break LOCAs via a
-multiple-component model representing a double-ended guillotine break.
YAEC presented several representative large break LOCA calculations using a 28-in. diameter break in the discharge pipe' of one recirculation-loop just upstream of the header pipe. For small break responses, YAEC presented several. typical LOCA studies for.an approximately 3-in. diameter break at the same location. Guidelines specified in Section 2.2.2 for break flow
~
modeling were included in the NSSS break flow model. To increase computational efficiency, the NSSS LOCA.model does not account for radiation heat trarsfer in the fuel rod assemblies but the effect of this heat transfer is addressed in detail in the hot channel (HC) model, shown in Figure .13, along with other important details.
4.1' Vermont Yankee Vessel Model The Vermont Yankee reactor vessel model consists of.a combination of hydrodynamic and heat transfer structures. The pressure vessel includes the lower plenum; control rod guide tubes; core bypass; three lumped fuel assemblies representing low, average, and high power levels; upper plenum; standpipes and separators; intermediate steam regions; steam dryer assembly; steam dome; and a downcomer. Based on previous model assessments, it is concluded that an appropriate number of nodes were employed to adequately model each of the above regions. Heat transfer structures are used to model the lower plenum mass, internal core hardware, ,
downcomer, actual vessel walls, control rod guide tubes, channel boxes, fuel rods and cladding, standpipes, steam separator regions, upper plenum mass, and jet pumps. A special multicomponent steam separator model is included in the NSSS model and is described in Section 2.3.1. The vessel walls are represented by heat structure components which are adequately 48 I
i UPPER PLENUM i 1
TV 207 TV 206 103 163 104 164 109 169 4-m
-09
- 01 -09 -09 x . . .
q
-08 ' 08 y -08 -08
~.. .
y E E *
-07 02 s -07 -07 s.
1 t v
- u g -
f
-06 'Of -06 -06 8 8 . - e -
0 -05 o 03 o -05 ?o -05 i H s b e g
o _'
e e
g -
g- -04 .L -04 o -04 ? -04 4 a m .m n
~ 2 * , ,,
-03 -03 -03 -03'
-02 -02 -02 -02
-01 -01 -01 -01 TJ 101 TJ 161 BYPASS H-CHANNEL INLET . INLET TV 100 TV 160 Figure 13. Vermont Yankee hot channel (HC) model noda11zation.
49
__. - __ =_ -.
nodalized to. reflect their correct thermal response. All vessel wall structures.have an imposed adiabatic boundary condition which adds additional conservatism'.
It is recommended that the Vermont. Yankee'NSSS.nodalization be .
accepted. The model'provides sufficient detail to captere LOCA phenomena.
No attempt was made. to verify the input numbers for the Vermont Yankee -
model and should be' reviewed with future submittals.
4.2 Core Power ;
The total core power is determined by the point kinetics model in RELAPSYA. . Conservative as well as best estimate input data can be entered in this model in order to compute the fission power and decay heat per 10CFR50.46 and Appendix K. These include use of an initial power equal to 1.02 times the licensed power level and the use of the 1971 ANS decay curves plus 20%. To further provide for a conservative calculation, all the core power is assumed to be generated in the fuel and no1e is deposited in the cladding, moderator, or any passive heat structures. YAEC added another self-imposed conservative condition via the use of the design power, instead of the licensed rated power, in all their licedsing example calculations.
The core is represented ~ by three lumped assemblies which model low, average, and high power assemblies. The low power assembly represents 116 peripheral icw power assemblies; the average power assembly represents 248 central average power assemblies; and the high power assembly represents four central high power assemblies. Based on'a detailed review of the NSSS model and the code assessment, it is concluded that the assembly model nodalizations are adequate to correctly predict expected thermal response behavior. For cases involving metal-water reaction and ,
clad rupture, renodalizations may be required to ensure Appendix K requirements pertaining to'the minimum amount of nodal cladding involved in the reaction are met. YAEC agrees to perform this renodalization if required. The power distribution within the three modeled assemblies is 50
I dictated by the product of a core region bundle fraction, a core region radial power factor, and a core region axial power factor. Values for these factors for each modeled assembly were determined from data generated 35 by the SIMULATE computer code for Vermont Yankee Cycles 9, 10, 11.
l
, While bounding cycle independent values were obtained by choosing conservative values for each factor based on these three cycles, any future
, licensing method application should provide justification that the current cycle is indeed represented by these three bounding cycle groups.
4.3 Hot Channel Model To evaluate the single high power bundle response in the Vermont NSSS model, YAEC developed a single hot channel (HC) RELAP5YA model which utilizes the hydraulic boundary conditions at the inlet and outlet of the high power bundle region in the NSSS model under the same accident conditions. With this model YAEC can arovide greater detail for the HC during postulated LOCA events.
The HC model, Figure 13, represents two nine-node parallel channels: one representing one-eighth of a hot assembly and the other representing the corresponding bypass region for a hot channel. Heat structures representing the channel box are ir.cluded to realistically couple the thermal-hydraulics in the two channels. Further simplification in the hot channel model is included by lumping the actual rods present in a typical one-eighth bundle section into two rods (see Figure 14). The first representing an actual hot rod (with only a half rod actually modeled) and the se:ond representing the remaining rods. Hence, the hot channel model includes three heat structure componente the channel box, a composite rod in the hot channel, and an actual hot rod. The required
^
radiation view factors for this heat transfer calculation are obtained from 36 the HUXY code for the original one-eighth fuel bundle configuration and
~
then these factors are collapsed (using standard applications of continuity and reciprocity relations) to represent the three structure model with the hot rod "seeing" only the composite rod and the composite rod "seeing" both the channel wall and the hot rod.
51
_________m
V l .
- t C .
[ -
mhS . .'
L L
A 2
- r,.hM ld
=$ ee
~ 'm W ?
dl
)
~
oe x = ~ md o . o B Cm l
l
'(. : s L ra E , f) o N b N t(
A I u
) o ,
D - b yd
- ( ae A l t n
ne
, os ie t r cp
, ue s d r e
rs Y d R eo l r 8T
/E d 1 M nf M uo l . , (
- Y S
\
b do oi n
rt a
8c 0y - / o
=9.
1-
_ - 1l
. -h l .e3 i m'
- E
>=
e, 4 - -- .
L l 0.0
( E.0 (Wg d [M j 'l A9ji) 12.0 TIrt tSEC) 1m.0 3@.0 3(m.0 Figure 15. High power assembly fluid mass (LBLOCA-EB).
I I f
-e gg ..
m b9 l -
R,R 5 j. - . - - - - . -
c.o L so.o :0.0 h)hpj, 1sc.0 se.0 h
son.o Figure 16. High power assembly fluid mass (LBLOCA-EA).
56
bypass for 'the three fuel assembly channels modeled and the different hydrodynamic history for the two cases. Results indicated that the CCFL condition had dropped down one node from the top of the bundle allowing liquid influx which only affected the two topmost nodes of the assembly.
l ,
This-phenomenon'will be self-limiting since the more liquid influx, the greater the potential for CCFL.
An additional area of concern in the'large break LOCA transients involved the large, long term break flow oscillations present in both Case
-EA and -EB. These oscillations resulted from an improper specification of the drywell' containment conditions (16.4 psia,.212 F) which allowed subcooled water to be drawn into the vessel instead of steam. YAEC stated that this problem, which did not significantly affect transient results since it occurred after LPCI injection, would be eliminated in future models'by specifying saturated steam conditions. (Note: the small break transients presented below did not have a similar improper specification of the drywell conditions, as seen in Table 2.)
To provide further assessment of the NSSS LOCA model, YAEC provided an additional " conservative best estimate" calculation (LBLOCA-BA)34 for the large break accident based on the major accident assumptions'of case LBLOCA-EA. The transient was termed " conservative best estimate" because many of the EM options were not activated. This transient was initiated at 1631 MWt (compared to 1664 MWt in -EA), had deactivated the return to nucleate boiling and transitions boiling lockout options thus allowing rewetting and quenching when local conditions permitted, and used the 1971 ANS Decay Curves with a multiplier of one. The calculation resulted in a lower PCT, as expected, with a peak clad temperature of 918 F compared to 1067 F for LBLOCA-EA. In addition, this transient included the proper specification of the drywell containment conditions (saturated steam at 16.4 psia) which resulted in a much improved break flow rate response
.(Figures 17'and 18). j 1
1 57
TABLE 2.
SUMMARY
OF VERMONT YANKEE SMALL BREAK ACCIDENT ASSUMPTIONS
- 1. Small recirculation discharge break (0.05 ft ) 2at 4.0E-6 s.
- 2. Loss of auxiliary power occurs at 4.0E-6 s.
67B-EOC is used.
- 4. Feedwater coasts down to 0.0 lbm/s at 5.0 s.
- 5. MSIVs close in 10.0 s after isolation signal plus 0.5-s delay.
- 6. Recirculation pumps in A and B loops coast down with decreasing power from loss of MG sets.
{
- 7. ADS may actuate if appropriate signals exist. Thereafter, ADS cycles open/close at 12 psid between steam line and drywell anytime ADS criteria are currently met.
- 8. HPCI steam turbine admission valve fails to open on demand. Thus, HPCI fails to inject. (This is the single failure.)
- 9. No credit for RCIC operation.
- 10. Two LPCS Systems inject on demand.
- 11. LPCI-A injection valve opens upon demand.
- 12. LPCI-B .frjection valve opens upon demand.
- 13. Drywell pressure and quality are assumed constant at 16.4 psia and 1.0 for fluid sink conditions. High drywell pressure is conservatively estimated to occur at 18.4 s for this case by a containment calculation.
- 14. Wetwell pressure and temperature are assumed constant at 14.7 psia and 165 F for fluid source and sink conditions.
- 15. EM point reactor kinetics initially at 1,664 MWth.
- 16. EM core heat transfer.
- 17. Passive heat structures are included.
- 18. Moody two phase critical flow model used at the break location. .
- 19. 1971 ANS Decay Heat Standard plus 20%.
58
O.
h o t 8 m. ,
. j m5 s
E ![,, I l . !! I h
-I g 1
- l
- 0. _ . . _
5i<p; t ut_ c l ,.
-ez_ - , . _ . m 3
i i ,. i a !!-
.d *d - .
1 l
'3 J
6 A. ,
t g e o-e UPSTREAM BREAK ; .
8 e--e DWNSTRERM BRERK
- I
g ,
gg_
E' o.o sa.o taa.o teo.o am.o n.o TIFI (SEC)
Figure 17. Long tem break flow rates (LBLOCA-EA).
$, e i' -
W. .I s9 '.
8 i
- L1F a .
g k- 1 y '*-* 1 1 ,: "' * ..f.~~..".'f.,^-- @ L T 4 s^ f u s a-I
. d EI 8 a-e UPSTRCAM BREAK
+- g, ***e CWNSTRERM BRCAK Sh a-.
0.0 80.0 130.0 180.0 240.0 300.s TIPI (SEC)
Figure 18. Long tem break flow rates (LBLOCA-BA).
59
"I 4.4.2 Small 2-aak LOCA Transients The small break transient (SBLOCA-EY)3' represents an approximately 3-in. diameter break in the discharge pipe of one recirculation loop. As in the large heck accidents J a loss of power was assumed at transient -
initiation combined with a tailure of the auxiliary power supply. To
- satisfy the single failure criterion, the HPCI system was assumed to be
- inoperative. The only systems available to mitigate the accident were the two LPCS systems and the LPCI systems. All major accident assumptions for this, case are presented in Table 2.
I n A review of the SBLOCA-EY results revealed a large oscillatory br*.ak
. flow after LPCI initiation. The problem in calculating the proper break flow response resulted from the improper use of the connecting pipe area,
' instead of the actual break area, in the momentum solutions when determining the critical' flow velocities wii.h the use of the abrupt area change option at the break. (Note: the large break transicats, which assumed DEG breaks, were unaffected by the above problem since the break areasandconnecti79 pipe'areaswereidentical.) This problem was-resolved l by using the smooth ' area change option at the break and specifying a sorre$pondingformlosscoefficient(i.e., forward =1.0and reverse = 0.5). However, when the analysis was rerun, another problem was discovered wnich involved the calculation of critical fic s with time-dependent volumes during periods of reversed flow. Tnis problem was resolved by updating the code to correctly integrate the momentum equations, which define the critical velocities, and to use volume-acciared quantities instead of junction-donored in the momentum equations during reversed flow. A detailed discussion of these updates can be found in Reference 11 (response to Question Q1.9). The smallibreak LOCA was rerun as SBLOCA-EW.11 Results of thjs transient followed the same general trends of the earlier transiend but the improved break flow modeling ,
calculation reduced the depressurization response and, hence, delayed CHF and LPCS injection. The peak cladding temperature was found to be about 150 F higher and delayec about;130 s compared to SBLOCA E(;
l 50
.1
,1
( 4 1 . _ _ _ _ _ .
l l
To address concerns in using the multicomponent steam separator model j in the Vermont Yankee NSSS model, YAEC presented two additional small break transients (SBLOCA EZ and -E50). Because of the uncertain performance of the SEPARATOR model component in RELAP5YA, a multicomponent perfect steam separator model was used to achieve the steady-state condition from which a LOCA transient is initiated. Since this multicomponent perfect separator
.- is isolated in the model upon initiation of the LOCA transient, concerns about transients being superimposed upon the LOCA transient arose.
Since superpositioned effects would be more obvious in the slow response small break transient, YAEC compared results of Case EZ to Case EW. Case EZ is identical to Case EW except that the multicomponent perfect steam separator is not isolated at time zero and the transient
. proceeds. Results of this comparison showed some minor differences but overall the calculations yielded similar responses. The main difference between these cases seems to result from the numerical donoring of different steam qualities in the steam dome.
YAEC also performed a steady-state " transient" (Case ES0), where the perfect steam eeparator is deactivated at 10 s into the " transient" and the normal BRANCH component steam separator model (see Section 2.3.1) is activated. No assumed SBLOCA initiating events are activated during this
" null" transient. The results indicate that the two separator models are different. The perfect separator model can simulate.the correct steady-state performance of the actual separators since it eliminates any carryover of liquid or carryunder of steam. The BRANCH component separator results in a gravity induced separation resulting in excessive carryunder and carryover at steady-state. As a result of this different performance, different steady-state conditions resulted.
Another difference between these two cases re& ts from switching the l
separator models. The switching process induces a momentum transient which results in a temporarily (approximately 5 s) reduced steam line flow rate and an increased core flow. This switching transient takes about 30 s to completely die out.
61
l Results using the two separator models during steady state indicate that the multicomponent perfect steam separator can more adequately address l' the desired steady-state performance of the steam separator process by eliminating the carryunder of steam and carryover of liquid, but does not adequately. address transient response behavior. Also, the steady-state ,
perturbation shift between these models seems to be limited to about 30 s, which is a time frame that will be dominated by LOCA responses. For the -
longer term transient response comparisons, both separator models show similar responses. Hence, the multicomponent perfect steam separator model is recommended as a means of determining a steady-state operating condition and the small transient imposed by the switching to the BRANCH' component separator during LOCA initiation has no significant effect on the final predicted '0CA response.
t 62
m H
- 5. COMPLIANCE WITH NRC REQUIREMENTS l n
I
. Appendix K to 10CFR Part- 50 specifies the- required and acceptabl.e.
features'of any model for LOCA licensing application. - Additional NUREG 7 requirements are also presented whenever applicable. YAEC addressed each l LAppendix K BWR'related requirement individually in Reference 1. No attempt H
.. will be made to comment on all of-the YAEC's responses in this report but' those' responses having bearing on'this Llicensing assessment or showing potential areas of concern will be addressed.
Requirements' pertaining to the metal-water reaction were all met.
YAEC also stated that renodalization will be performed, if necessary, in ' 'l the event-'of a cladding rupture to assure that the ruptured node will not
'be more than 3 in. in length. 'YAEC also incorporates-cladding material ,
l property changes as oxide. forms from the reaction process. l J
The fuel rod. requirement was met by using the fuel rod behavior models-from the. licensing approved'T00DEE2-EM I and the cladding rupture and !
flow blockage tables.from NUREG-0630.20 l l
The Moody ' critical flow model was incorporated in RELAP5YA and is ;
- applied to the break location in'all licensing calculations involving a i two phase discharge fluid. YAEC intends to meet Appendix K requirements j for using break discharge coefficients ranging from 0.6 to 1.0 for'all BWR l large break analyses.
l Concerning post-CHF heat transfer requirements, YAEC included an transition boiling lockout algorithm which is activated if any clad surface I exceeds 300 F superheat. Based on a subsequent return to transition or l nucleate boiling modes, the heat transfer algorithm yields an degraded heat transfer coefficient since it will extrapolate the film boiling j correlations into these regions. )
l
.l l
i 63 L
1 After CHE is'.first. predicted during a blowdown, the.. Appendix K 1
' licensing requirements specify.that the calculations will-not.use a nucleate-boiling heat transfer correlation until the quench portion of the
- LOCA. YAEC ' incorporated.a' return-to-nueleate boiling lockout option' to -
respond to this~ requirement. Once CHF is reached,. the algorithm lockouts .
l .the'~ calculation of nucleate boiling heat transfer coefficients until ,
quenching is. allowed by the rewet and_quencn front model'(Section 2~2.6). .
If the heat' transfer logic returns to n'ucleate boiling during this lockout period, YAEC still calculates the nucleate boiling heat transfer coefficient but multiplies it by a degrading factor'(XMNB).
For. meeting' pump model requirements, YAEC adequately assessed the jet I
pump model 'and the modal is adequate for licensing calculations with the i forward and reverse loss coefficients presented in the licensing model submitta.i. The centrifugal pump model used is the same as RELAP5/ MODI and the earlier RELAP4 pump model. This model has proven acceptable in'many previous analyses. The basis for homologous curves and two phase multiplier were provided in response to Q.V.13 in Reference 7.
Concerning convective heat transfer coefficients during blowdown for BWR fuel rods and channel boxes under spray cooling, YAEC will not apply the specified heat transfer coefficients'but rather use the RELAP5YA best estimate heat, transfer algorithm. The algorithm uses many best estimate correlations covering the range of conditions expected in a LOCA. For licensing ' applications, YAEC intends to use the multisurface radiation, the return-to-nucleate boiling lockout, the return-to-transition boiling lockout, and the rewet and quench heat transfer options.
To support the decision of using the RELA 05YA heat transfer algorithm, YAEC provided TLTA Test 6422/3 heat transfer coefficient test data in response to Q.V.6 of Reference 10. This test was a large break test which was. very similar to TLTA Test 6425/2 and used average bundle power and
. average ECC. injection. Comparisons of available TLTA Test 6422/3 heat transfer coefficient test data and RELAP5YA TLTA Test 6425/2 test ,
predictions, presented in Figures 19, 20, and 21, for several elevations in 64 l
0 0
3
. 0 5
2 5 >
L L t E E eE d
" "9N o" 0 1
7 7 i
a 0 t
A tY o 2
a a5 t P )
a aA." c tl i t
a a E6 o D D R(
- 0 h
- g 5E 1
M A I T
A 0
' s 0 1
g\ A 3I III l
}
A 0
l Ii
'g 5 g i1 :
s ' jA
\
\ .
\ - - - 0 0 0 0 0 0 0 0 0 0 0 1 1 1
0, 0, 0, 0 0 1 0 1 1
- J a O" i n g B yn" g!" 3x em il
lla 0
0 3 . _
s .
0 i
t o
n i' 5 c . -
si
. )
2 nd oe i r t p nl . f l a E
i E y E e g v2
" d" o e/
0 n1 l l 5 "9
7 6 0 a 10 do l
0 e2 4
t t Y ol A oi s s] 0 t6 a
aaS t t P e d
o d
o 2 ss t
R R a a A. " H t t l 1 r r ne oT a aE8 H e e ) s DD R( E n t n
c id rn r
o e e s aa C C 0 p g
- 5
( ma
- - [
1 E ca ot d
- * . *l*
!'U j- I T
Mt e/n3 i2 c2
-l ' i4 A '0 f6 f _
/ l 0 et os _
1 _
ce A r T _ _
s e e fT eA sL nT a
rr ek A 0 t o _
s A i 5 t f _
i a e "0 l
d
\ :I * ! * ! ' ' h0
)}1 i I'9. \ .\
- I l.r j j 1 ii l
d e
1 j -l . l n"
kf - - 0 u9 0
B7 0 Q 0 0 0 0 0 Q 0 0 1 1 2 1
0 Q 0 r e
1 u 0 Q g .
0 j i F
1
. ! a" = .
" y uin '::QsH mm _
llll! l l ,I _
1JIIl 1
0 0
3 . -
s
, n -
i o .
t 0 c
, I 5 i d
2 ne or i p a t L7 E
n a2 v/
e- e5 d2 l2 "0 u2 N'
0 e4 6
0 1
I 0 t at t A "'
Y 2 s a5 se P" nT a A' )
o t L" sd a E" cin ra D R e a 0 b pa mt
- I 5 oa
- A 1 E cd A M t3 I
n/
e2 i2 T c4 i6
' 0 f ft
/ l 0 es oe 1 cT
/ A rA eT
/ fL
= f t
/ A sT n
ar ro A 1 0 tf
\
\ 5 t" a2 8 \ e2 j1 I i s 3
) h1 l
e-1 d" .
n1 y+s\ - ,
- - 0 u0 B1 0 0 0 0 0 0 1 0 0 0 0 1 1
2 1 e 0, 0, 0 1
r u
g 0 0 i 0 1 F 1
cJ Y"t =D% gn2t8U Eza Pax e
1 il1 jli
the test bundle showed that RELAP5YA underpredicted heat transfer coefficients. Also, as noted in the discussion of the "best' estimate" transient prediction of TLTA Test 6425/2-(Section 2.4.2), the RELAP5YA temperature predictions bounded the measured PCT. RELAP5YA heat transfer predictions for TLTA Test 6426/1 showed higher predicted cladding -
temperatures than measured in the test especially in the post-CHF regime using the Condie-Bengston film boiling correlation. Based on these trends, as well as the fact that the heat transfer algorithms perform more realistic response modeling, the RELAP5YA heat transfer algorithm is recommended for LOCA calculations.
I e
D 68 ,
i I
.________________a
r
~s ,
a 1
.l
- 6. REVIEW OF UTILITY QUESTION RESPONSES ',
s .
During the- review-of RELAP5YA as a code for licensing analyses, three.
. sets of questions were submitted.to.YAEC by the NRC.(References 5 through' !
l
,,: '12)~ -The first:two sets of questions (References 5-9 and 10) were.a'Lresult
. )
'- ~
, of .a. general RELAP5YA code review' relating to- PWR and BWR applications, respectively. After a more detailed review of the code', the previous two l 1
- questionsetsand[their. answers,andareviewof.new~1ysubmitteddata,'a. j Lthird. set of questions (References 11 and 12) resulted. The' responses to- I
.all questions'were reviewed and found acceptable and. included in the .j consideration of RELAP5YA for use in licensing calculations and the submitted Vermont Yankee LOCA model as a LOCA licensing analysis model.
l I
i l
+
l l
i
] ;
i
,W-l 69
- __7___,
r 1
' ; F4 i' 5
~71,RECOMMENDATIONSAND' CONCLUSIONS L7 l -.
l ,
Th'e.RELAP5YA' submittal b'y YAEC was reviewed to determine the code's acceptability for uselin BWR.large and'small break loss-of-coolant best
' estimate'and licensing analyses. Based'on this review,'it is. recommended 9 that the' code' be accepted for. best estimate applications; as ' outlined in .
~
-Reference 4 (SECY-83-472), pertaining to the Vermont Yankee Nuclear Power'
, Station. Many of the required Appendix K models have also been reviewed and are recommended for. acceptance. Use of RELAP5YA is based on the.
following comments and restrictions.
Certain model specific areas, enumerated in Section 3.0, are unacceptable for any LOCA analyses since their use would. involve known coding' errors whose update corrections YAEC elected not to incorporate o in RELAP5YA. , These $recific modeling areas are:
- 1. .Cannot use control variables as power input to a heat structure
- 2. Cannot .use reactor. kinetics when no power is generated from-gamma decay heating
- 3. ~Cannot use VALVE component with form loss coefficients
- 4. Cannot use control v..riables with reactor kinetic feedback.
-The eight modifications described in Section 2.0 that apply to BWR LOCA analyses have been reviewed and comments pertaining to each model are described below.
INTERPHASE DRAG - Based on the steady-state and transient assessment calculations, the-new interphase drag models in
- RELAPSYA are judged to be properly implemented, to provide an acceptable simulation of the interphase drag related phenomena, 70
I and to yield conservative results whenever best estimate results
~
could'not be obtained.
' CRITICAL FLOW - The Moody Model, used with the YAEC recommended ,
g donor cell ~ static pressure and enthalpy option (where donor cell
]
velocities are low) and without the Moody theoretical slip model,
-' was implemented properly and. provides conservative break flow ]
predictions. To insure the intended conservatism, all licensing !
l
,- .cal cul at ions w ti h 'nonbreak location limiting critical flow must !
.either: (a) impose the Moody Break Flow model at the limiting location, (b) project the limiting area to the break loc.'. tion .
using Moody or (c) include justification for using Moody t the break location only if the .results are more limiting.
The best estimate cr.itical flow model produced. satisfactory results and prediction of experimental data and is recommended for use in best estimate calculations.
JET PUMP - The jet pump model is implemented properly; however, due to its many input related requirements, special justification must be provided for any deviation of the NSSS model forward and reverse loss coefficients.34 In addition, all jet pump nodalizations of the diffuser and tailpipe components must reflect actual geometry.
NUCLEATE BOILING - The new model, using the Chen and Thom i correlations, is judged acceptable for steady-state and transient applications as. implemented.
CRITICAL HEAT FLUX - The new CHF model, using the Modified Biasi and .the Griffith-Zuber correlations, is judged acceptable as implemented.
71 t
y
l.
I-
.REWET.dND' QUENCH'- RELAP5YA~did a reasonab'le job of predicting, the' bottom reflood; tests at THTF. For the important parameters,
- where differences existed between code calculated results and.
test data', the'. code tended to predict conservative results.
~
RELAP5YA also predicted the cladding temperature response of the- J;
'TLTA~ experiments indicating..the ability to predict . top down cla'd
^
quenching. Therefore, the new reflood models'in RELAPSYA are recommended for use in-LOCA calculations.
MULTIPLE SURFACE RAD'IATION - This model adequately predicts the
-thermal.radiationLresponse and is recommended for LOCA application.as'long as appropriate justification is.also' presented.for the associated input view factor matrix. If the view factors are. generated'with a' code, this. code also must receive ,licen sing 'ap' proval .
-HEAT TRANSFER LOCKOUT OPTIONS - Lockout of return-to-nucleate boiling"(with degrading nucleate boiling heat transfer multiplier.
[XMNB]) and transition-boiling have been' implemented in the code. The selection of a value of XMNB =.0.05 is recommended and any increase lshould be justified in any licensing submittal.
FUEL. BEHAVIOR ~- Since the models were derived almost directly from-T00DEE2-EM, and since the assessment review showed very good comparisons' to T00DEE2-EM calculated results, the fuel ' behavior models are judged-to be acceptable as implemented.
A-large number of' options are available to the user'in RELAP5YA.
Therefore, to ensure that the code is used within its capability and with proper input, it is recommended that any' changes to the submitted licensing model input decks be justified in future licensing'submittals. ,
-(,'
l 72 E - - - - - _ _ - - - - - _ _ _ _ - - - - - - - _ - - _ - - - _
Since no applications or assessments of the SEPARATOR component have been-presented by YAEC.to.the NRC, the SEPARATOR component model in-RELAP5YA should be restricted. The multicomponent separator is recommended and performed satisfactorily.
I l' ) Assessment of RELAP5YA'with integral data from TLTA and THTF produced 1 . good results. -RELAP5YA was capable of simulating the response of TLTA for LOCA tests. System parameters were well predicted and the peak cladding. temperatures were bounded by those calculated by RELAP5fA.
1 The review of the submitted Vermont. Yankee Nuclear Power Station LOCA l licensing analysis model was reviewed and'is recommended for'use in LOCA calculations with the following comments and restrictions:
Submittal calculations we're based on Vermont Yankee Cycles 9, 10, and 11. Future licensing applications should provide justification that the fuel cycle in question is representative of the fuel cycle group planned for Vermont Yankee.
It is recommended that the Vermont Yankee NSSS nodalization be accepted. The model provides sufficient detail to capture LOCA
-phenomena. No attempt was made to verify'the input numbers for the Vermont Yankee model and should be reviewed with future submittals.
It is recommended that the HC model be accepted. The HC model use standard applications of continuity and reciprocity for lumping of fuel rods. I
.The use of NSSS and HC model is contingent on the NRC approval of the FROSSTEY fuel rod behavior code and the BWR application of
" l the HUXY view factor code. (Note: HUXY has already been approved for PWR licensing calculations.)
i l
73
2 l
- 8. REFERENCES L , .
l'. = R. ^ T, Fernandez et al. , RELAP5YA--A Computer Program for Light-Water System Thermal-Hydraulic Analysis,-Volume I: Code Description. I, Yankee Atomic Electric Ccmpany Report YAEC-1300P, October 1982.
/ 2. l!. H. Ransom'et al . , RELAP5/ MODI Code Manual, EG&G Idaho Inc. ,
'NUREG/CR-1826,: EGG-2070, November 1980.
9 E
^
-3. R. T. Fernandez and H. C. de Silva, Jr. , Vermont' Yankee BWR Loss-Of-Coolant' Accident Licensing Analysis Metnod, YAEC-1547, June 1986.
L .4. Letter W. J. Dircks,-NRC, to the Commissioners, NRC, " Emergency Core-Cooling System Analysis Methods", SECY-83-472, November 17,'1983.
- 5. Letter J. , A. Kay, . YAEC, to J. A. Zwolinski, NRC, " Response to NRC
- Questions on RELAP5YA," FVY-85-18, March 1,1985.
.6. . Letter J. A.. Kay~, YAEC, to J. A. Zwolinski, NRC, " Response to NRC Questions on RELAP5YA," FVR 85-48, April 30, 1985.
- 7. Letter G. Papanic, YAEC, to J. A. Zwolinski, NRC, " Response to NRC Questions on RELAP5YA,'.' FVR 85-72, July 1,1985.
-8. . Letter G. Papanic,'YAEC, to'J. - A. Zwolinski, NRC, ." Response to NRC
-Questions on RELAP5YA," FVR 85-87,' August 15, 1985.
- 9. Letter G. Papanic, YAEC, to_ J. A. Zwolinski, . NRC, " Response to-NRC Questions on RELAP5YA," FVR 85-121,. November.1, 1985.
- 10. Letter, R. W. Capstick (YAEC) to D. R. Muller (NRC), " Response ~ to 39 Additional NRC Questions on the RELAP5YA. Computer Code,"
FVY 85-122,. December 31, 1985.
- 11. : Letter,'R.' W. Capstick (YAEC) to V. L. Rooney (NRC), " Response:to Additional NRC' Questions on the RELAP5YA Computer Code," FVY 86-94, October 14, 1986.
- 12. Letter, R. W. Capstick (YAEC) to V. L. Rooney (NRC), " Response to Additional NRC Questions on the RELAPSYA Computer Code," FVY 86-104, November 4, 1986, i13. . R. T. Fernandez et al., RELAPSYA--A Comouter Program for Light-Water -
. System Thermal-Hydraulic Analysis, Volume III: Code Assessment, Yankee Atomic Electric Company Report YAEC-1300P, Volume III, ,
October 1982.
.14. Yankee Atomic Electric Comeany WREM-Based PWR ECCS Evaluation Model
.(Version YAEC-058), YAEC-1160, July 1978.
- 15. - E, Janssen, "Two-Phase Flow and Heat Transfer in Multired Geometries, Final . Report," General Electric Company GEAP-13347, March 1971.
74
,. {
s q i
, 3
- j i
-16. 0;'Nyland et al., Hydrodynamic and Heat Transfer Measurements on a j Full-Scale Simulated 360 Rod Marviken Element with Uniform Heat Flux Distribution, FRIGG-2, R4-447/RTL-10007, Sweden 1986.
l 17.. 'B'.-C..Slifer'and J. E. Hence, Loss-of-Coolant Accident and Emergency '
.CoreL Cooling Models for General Electric Boiling Water Reactors, NEDD-10329,.Arcil 1971.
~
- 18. L. Ericson and 'et' al. , Marviken Full-Scale Critical Flow Tests Interim.
Report; Results from Test 10, MX 3-63, Marviken- Power. Station,- Sweden, November 1978.
- 19. G. E. Wilson, INEL One-Sixth Scale Jet Pump Data Analysis, EGG-CAAD-357, EG&G Idaho, Inc., February 1981.
- 20. J.-C.. Chen,."A Correlation for Boiling Heat Transfer to Saturation Fluids. in Convective Flow," I&EC Process Design and Development, 5,
.1966.
- 21. V. E. Schrock and L. M. Grossman, Forced Convection Boiling Studies, Final Report on Forced Convection Vaporization Project, TID-14632,-
University of California at Berkeley,.1959.
'22. . J. R. S. Thom et al . .. " Boiling in' Subcooled Water During Flow Up Heated Tubes or Annuli," Proceedings Inst. of Mechanical Engineers, 3C180, 1966.
c23. ' A. W. Bennett et al. , Heat Transfer to Steam-Water Mixtures Flowing in -
' Uniformly Heated Tubes in Which the Critical Heat Flux Has Been Exceeded, AERE-R5375, Atomic Energy Research Establishment, Great'
. Britain, 1967.
- 24. R. E. Phillips, R. W..Shumway, K. H. Chu, " Improvements to the Prediction of Boiling Transition During Boiling Water Reactor Transients," 20th ASME/AIChE National Heat Transfer Conference, Milwaukee, Wisconsin, August 1981.
- 25. P. 'Griffith et al. , " Critical Heat Flux During a Loss-of-Coolant Accident," Nuclear Safety, 18, May-June 1977.
- 26. Critical Heat Flux Correlation for CE Fuel Assemblies with Standard Spacers Grids Parts 1 and 2, Non-Uniform Axial Power Distribution, Combustion Engineering Topical Report, CEN PD-207, June 1976. ,
- 27. G. L. Yoder et al.., Dispersed Flow Film Boiling in the Rod Bundle Geometry Steady-State Heat Transfer Data and Correlation Comparisons, I
'- ORNL/5822 (Preliminary Draft).
i 1
1 75
_ _ _ - _ - _ _ ____ ____ k
p . ... .
- 1 I
ol
[
j j
l L. . 28. D; A. Powers and R. O. Meyers, Cladding Swelling an'd Rupture Modes,for L . LOCA Analysis, NUREG-0630, April 1980.
- 29. : S. <P. Schultz'and K. E. St. John, Method for the Analysis of Oxide Fuel
- i. Rod' Steady-State Thermal Effects (FROSSTEY) Code /Model Description Manual , YAEC-1249P, . Apri1.1981.
- 30. R.1W. Oehlberg, W. 'V. Johnston, J. A.
Dearien,
"FRAP Fuel Behavior Codes,". Nuclear Safety, 19,15, September-October 1980. ,
~ 31. Clarification of TMI Action Plan Requirements, NRC-NRR, NUREG-0737.
- 32. O. C'.. Jones and P. Saha, " Volumetric Vapor Generation in Nonequilibrium, Two-Phase Flows," Prepared for Advanced Code Review Group Meeting of th'e Water Safety Research Division, USNRC1Washington, D.C., June 2, 1977.
- 33. -Lee . L. S. , G. L. -Sozzi and S. A. Allison,; BWR Large Break Simulation Tests-BWR Blowdown / Emergency Core Cooling Program, Volumes 1 and 2,
- l. NUREG/CR-2229, April 1982.
- 34. Overhead handouts, Presented at'YAEC/NRC Meeting on VY-BWR LOCA Licensing Analysis-Method, Washington, D.C., July 29, 1986.
- 35. .D. M.:VerPlank', SIMULATE-2: A Nodal Core Analysis Program for l Light-Water Reactors,'YAEC-1392P, June 1985.
- 36. L. H. Steves et-al., HUXY: A Generalized Multi-Rod Heatup Code with 10CFR50,' APPENDIX K Heat-up Option User's Manual, Exxon Nuclear
' Company,.Inc., XN-CC-33(A), Rev. 1, November 14, 1985.
.m f
76
i c ,oase sat u a .uct... .eous.,oe, ca m.is o. i...o.,~w.e.a e= roc e. ... .
L".',"M' al8UOGRAPHIC DATA SHEET EGG-RTH-7506 n .. .ucr.o . o., ... ......
,,,,u... ' " ' " " " ' '
.,,'" TECHNICAL EVALUATION REPORT: REVIEW AND EVALUATION OF THE RELAP5YA COMPUTER CODE AND THE 3' VERMONT YANKEE LOCA LICENSING ANALYSIS MODEL FOR . o.,i...o.,co. u .o aa
~ ~'-
USE IN SMALL AND LARGE BREAK BWR LOCAS l
.....o.... June 19"87
> J. L. Jones ........ ....... i P. D. Wheatley =~
I aa 19"87
{
Ju ne'~
,........o...,,,,o.....o.......oo..n,,,.c, .~.c,,,..o....,.....
EG&G Idaho, Inc. .,,,o...,,,... ,
P.O. Box 1625 D6022 t.
Idaho Falls, ID 83415
,,.o....o.....,.1...................,,,.c, ,,,....o,...o.,
U.S. Nuclear Regulatory Commission Infomal Washington, D.C. 20555 . ,. ..oo co. . . . .. . ,
June 1987 o .o-a ... . .. . .o. u
...u..c,..,,
A review was completed of the RELAP5YA computer code to_ determine its capabilities for performing licensing analyses. The review was 14.mited to Boiling Water Reactor (BWR) reactor applications. In addition, a
' Loss-Of-Coolant Accident (LOCA) licensing analysis model, using the RELAPSYA computer code, was reviewed. This model was reviewed for application to the Vermont Yankee Nuclear Power Station to perform full break spectra LOCA and fuel cycle independent analyses. The review of the RELAPSYA code consisted of an evaluation of all Yankee Atomic Electric Company (YAEC) incorporated modifications to the RELAP5/M001 Cycle 18 ,-
computer code from which the YAEC version of the code originated. -
Qualifying separate and integral effects assessment calculations were reviewed to evaluate the validity and proper implementation of the various added models. The LOCA licensing method was assessed by reviewing two RELAP5YA system input models and evaluating several small and large break qualifying transient calculations. The review of the RELAP5YA code modifications and their assessments, as well as the submitted LOCA input model, is discussed and the results of the review are provided.
. . ooc o.. . . ... . . . . . . . . .o.o. ouc... . . ....,.
1
~
'. s.cv. tv c6..s ,ic.vio.
............,o.....o.o. liEassified l
'On"cTassifiec
,,.....o,..cu e ,aic.