ML20207A625

Safety Evaluation of TR WCAP-14449(P), Application of Best Estimate Large Break LOCA Methodology to Westinghouse PWRs with Upper Plenum Injection. Rept Acceptable

ML20207A625
Person / Time
Issue date:	05/21/1999
From:	NRC (Affiliation Not Assigned)
To:
Shared Package
ML20207A621	List: ML20207A618 ML20207A622 ML20207A625
References
TAC-M94035, NUDOCS 9905270106
Download: ML20207A625 (16)

Text

_ pu .

g k UNITED STATES g' fj NUCLEAR REGULATORY COMMISSION

" I '

WASHINGTON D.C. sees &copi j

..... ENCLOSURE 1 SAFETY EVALUATION BY THE OFFICE OF NUCLEAR REACTOR REGULATION RELATED TO ACCEPTABILITY OF THE TOPICAL REPORT WCAP-14449(P)

" APPLICATION OF BEST ESTIMATE LARGE BREAK LOCA METHODOLOGY TO WESTINGHOUSE PWRS WITH UPPER PLENUM INJECTION" FOR REFERENCING IN PLANT LICENSE APPLICATIONS AND OTHER LICENSING ACTIONS WESTINGHOUSE ELECTRIC COMPANY (TAC NO. M94035) 1 INTRODUCTION In August 1995, the Westinghouse Electric Company OM) submitted WCAP-14449(P),

" Application of Best Estimate Large Break LOCA Methodology to Westinghouse PWRs With ,

Upper Plenum injection," dated August 1995 for Nuclear Regulatory Commission review and '

approval. WCAP 14449(P) describes a realistic emergency core cooling system (ECCS) evaluation model (EM) that W proposes for use in license applications and other licensing actions to demonstrate plant conformance with the requirements of 10 CFR 50.46 for  ;

postulated pressurized water reactor (PWR) large-break (LB) loss-of-coolant accidents  ;

(LOCAs) for Westinghouse 2-loop plant designs with upper plenum injection (UPI).  ;

i The ECCS design of W 2-loop UPI plants differs from that of the W 3- and 4-loop plant ECCS designs in that in the 2-loop UPI design the ECCS low pressure injection (LPI) is delivered to the upper plenum region above the core in the reactor vessel, whereas in a 3- or 4-loop W design the ECCS LPI is delivered to the cold leg piping of the reactor primary system. Certain thermal / hydraulic parameters, such as condensation and entrainment, become more important in regions between the point of injection and the core, because they influence when and how much injected water actually gets to the core. The purpose of the methodology described in WCAP-14449(P) la to account for these and other thermal / hydraulic differences between 2-loop UPI, and 3- and 4-Ioop Westinghouse plant designs.

The computer code used in the 2-loop UPI EM is WCOBRA/ TRAC. WCOBRA/ TRAC is Westinghouse's modified version of the NRC's COBRA / TRAC program. This program includes the capability to model three-dimensional flow behavior in the reactor vessel and all the major components in the primary system. In this EM, W uses the three-dimensional modeling capability only in simulating the reactor vessel; W represents the other primary system components with one-dimensional models.

Throughout this evaluation report the terms " realistic LOCA EM" and "best- estimate LOCA EM" may be used interchangeably as was done in Regulatory Guide (RG) 1.157, "Best Estimate Calculations of Emergency Core Cooling System Performance." The principal difference between this best estimate EM (WCOBRA/ TRAC) and the EM approved for application to Westinghouse 3- and 4-loop designs, described in WCAP-12945, is in the identification of the parameters to be ranged in the " Global Conditions" response surface. The staff concluded that this difference does not change the nature of the methodological processes of the approved model.

9905270106 990521 s PDR TOPRP EPlVWEST _ _ -

-i

( C PDR )

Cko527ofora f

_.~.-..n-~~ , - -~ - - " - - - - ~~~

2 2 STAFF EVALUATION Though the staff has concluded that the overall methodology of WCOBRA/ TRAC has not been __

changed, and limited its review to model adaptations to accomodate the UPI design, the staff review focused on assuring that the same processes were used to develop and justify this application of the model and that the same criteria are met by this application. The staff reviewed the changes described in WCAP-14449(P) to accommodate the 2-loop UPI design to assure continued conformance with the requirements of 10 CFR 50.46. In quantifying the "high level of probability" criterion of 10 CFR 50.46(a)(1)(i), the staff determined that a 95th percentile

' probability based on best approximations of the constituent parameter distributions continues to be appropriately high for this application. Because this application only applies to LBLOCA design-beQ analyses (which assume a single failure), a higher probability of not exceeding the ECCS uhna is not needed to ensure a safe design. Below this point, uncertainties are not a dominant contributor to the overall ECCS reliability, and further conservatism in the code prediction does not contribute to an increase in the public health and safety.

The staff also referred to the guidance provided in RG 1.157 and the methodology described in NUREG/CR-5249 (NUREG-5249), " Quantifying Reactor Safety Margins" (" Application of Code Scaling, Applicability, and Uncertainty (CSAU) Evaluation Methodology to a Large Break, Loss-of-Coolant Accident"), December 1989. The CSAU Methodology is also described in .

NUREG-1230, " Compendium of ECCS Research for Realistic LOCA Analysis", December 1988. l The staff evaluated the WCOBRA/ TRAC EM with the technical assistance of the Idaho National l Engineering Laboratory (INEL). The INEL evaluation and findings are described in detailin tha j INEL technical evaluation report (TER) which is attached as part of this report. l 2.1 Comoarison With the CSAU Methodoloav l Our review was structured to follow the logical progression of the CSAU methodology, in order ,

to easily identify and evaluate any deviations in the EM from the acceptable CSAU approach. j TER Section 3.1, WCAP-14449(P), and a_W letter to the NRC dated August 6,1998, provide i comparisons of the WCOBRA/ TRAC 2-loop UPI EM to the CSAU approach. The CSAU methodology is identified in RG 1.157 as one approach which conforms with the guidance provided by the regulatory guide. NUREG-5249, Section 2, lists 14 steps, divided among  ;

3 elements, which constitute the CSAU methodology. Figure 1, which is taken from  !

NUREG-5249, gives an overview of the CSAU methodology and the steps which comprise it. l This section provides our review of the WCOBRA/ TRAC EM and its comparison to the 14 steps of the CSAU methodology. A simplified description of the WCOBRA/ TRAC 2-loop UPI EM and its implementation is given in Section 2.2. This description is like the WCOBRA/ TRAC EM for 3- and 4-loop plants, except for the identification of different parameters to be ranged in the " Global Conditions" response surface.

2.1.1 Reauirements and Caoabilities NUREG-5249 categorizes the first 6 steps as the " Requirements and Capabilities" element. The applicability of a code to the analysis of a tr3nsient (in this case, LBLOCA)in a nuclear power ,

plant is determined by a comparison of the scenario and plant dictated requirements with the  !

simulation capabilities of the code.

1

i -

ELDeff 1 sEC3JfBO(T5 Merv statet ac cacc m ,

'non" cocE A. ... _ ,

CASABUTES 4

s.o r.ce2 sutMDt CIhmEft notaadwfato.

cant -

4 sare ama e asocsia eas oma WWLCynestA AN EENTFT af@ 8tmer asEmi & CEpWELafEPS E 84'adh4ma (89t?) J 4

m. m c xx e A CemLft.

1

[

DDeff 2 e's'E*O ver.

?

AssEsm ett '

MC RANGDC (lF am PARAaCTERS -

ecoaura,rup.

test e o C.L CLLaTNlpG .

4 6 . 4 coeur e .a a avos acaratups av ~ em,ur,s.

4s a see e a sc7s imme.as=

maato *-- oaAsrs war oa's saa - = vues em .-,, ax.,,,,,

i 1 4

.co,e

• se sus are WN giap,,, "a C turtuh4mf 9 ACCunaCT I

,,...,a.C,.

am a- ",;'&,c.,

=

)

e i I

,. mecaw artet

- mia,e,,, w acaevonoave sanaarves se i

a*C Statt a arwo,n, ,, l mc is

. Ca6takaTups ,

I g g....g.,.g. .g

.-1..=.,a,o. g 1 WWEB .. . ,

( gain gng, ,

nneer 3 .

4

,,,,,59,x,,g g ,,,,, f SENsrTMTV me , ,o,g van ,,,,

N%W ANALTES TO CALCU a1E specru: n. duo i.

>a wwc e**

. I l

NSLO1059 l Figure 1. Code semiing. applicability and uncenainty (CS AU) evaluation methodology.

l

,e,, ~ - ~ . . . + , _ . . .

' ~,7: ==r.o_* l-

3 Step 1. Scenario Specification - This step is needed because the dominant processes and safety parameters can change from one scenario to another. The identification and ranking of these processes and parameters can lead to a simplification making

- analyses manageable. We have considered the LBLOCA in assessing model adequacies and in defining the high probability criterion for this EM, as discussed in Section 2.0. W has specified that the EM will be used for LBLOCA scenarios. TER Section 3.1 states that this is the same as for the 3- and 4-loop EM. The staff agrees and concludes that this W specification is consistent with CSAU Step 1.

Step 2. Nuclear Power Plant Selection - Because the dominant phenomena and their interactions differ to varying degrees with the plant design, the plant class must be selected to assess model adequacies and to specify applicability limits for the EM.

W has specified that this version of its EM is to be applied to Westinghouse 2-loop

' UPI designs. This is discussed in TER Section 3.1.1. This.W specification is consistent with CSAU Step 2.

Step 3. Phenomena identification and Ranking - As indicated in Steps 1 and 2, plant behavior is not equally influenced by all processes and phenomena that occur during a transient. Accordingly, this step allows for the simplification of the analysis to make it manageable. We have reviewed phenomena identification and ranking to ensure that the analysis methodology has identified an appropriate set of important phenomena in its simplified form. As part of the methodology,E has developed a phenomena identification and ranking table (PIRT). The PIRT for the 3- and 4-loop EM identified eight physical models as being significant to the overall uncertainty; critical flow, pump performance, fuel rod parameters, core heat transfer, emergency core cooling bypass, entrainment, accumulator nitrogen, and condensation. These parameters were treated in the.W uncertainty 3- and 4-loop methodology. The PIRT developed by Westinghouse for the 2-loop UPI plants concluded that the important phenomena for the 2-loop design are similar to those for 3 and 4-loop designs; however, the PIRT identified that certain parameters should be given additional emphasis in the upper plenum, while other parameters on the cold leg side of the core warranted lesser emphasis. Sections 3.1.2 and 4 of the TER discuss the W PIRT for the 2-loop UPI design. Differences in phenomena descriptions and rankings between the Westinghouse 2-loop PIRT and the 3- and 4-loop PIRT were found to be appropriate to the differences in design. We, therefore, conclude that thew PIRT is consistent with CSAU Step 3.

Step 4. Frozen Code Version Selection - This step ensures that changes to the code after an evaluation has been completed do not impact the conclusions and that changes occur in an auditable and traceable manner. Our review has considered the frozen code selection to ensure that the objectives of this step are met. E selected the WCOBRA/ TRAC, MOD 7A, Revision 1 code. TER Section 3.1.3 discusses the selection of the frozen code. We have concluded that the.W frozen code selection of WCOBRA/ TRAC, MOD 7A, Revision 1, is consistent with the objectives of CSAU Step 4.

, aun --

• M "'" " Appn.pp h WgW_. _ _ e P

4 Step 5. Provision of Complete Code Documentation - This step provides documentation of the frozen code version that allows evaluation of the code's application to a postulated transient or accident for a specific plant and that provides a traceable record of the code and any modifications made to 1 The importance of this step was demonstrated in this review. WCAP-12945(P-A) contains documentation of WCOBRATTRAC, MOD 7A, Revision 1, and describes the other elements of the as approved methodology for 3- and 4-loop designs. WCAP-14449(P) contains documentation of how the same overall methodology, including WCOBRATTRAC, is applied to 2-loop UPI designs. After issuance of the staff SER covering the 2-loop adaptation of the methodology, Westinghouse will submit an approved version of WCAP-14449(P), which will include the same type of material as WCAP-12945(P- l A), but focusing on issues associated with adapting the methodology to the 2-loop UPI design. The staff will review this report when it is submitted. With the exception of a user manual and a user guide, complete documentation equivalent to ,

that identified in CSAU Step 5 has been provided and reviewed. With regard to the l user manual and user guide, this information is not within the direct scope of this l review, which focuses on the technical adequacy of the methodology.  !

Westinghouse develops this documentation for all its analytical models consistent 1 with their approvals, and retains and maintains these user documents in-house, l

cons: stent with 10 CFR 50.46(a)(1)(i) and Appendix K, Part II. We there. fore conclude that the documentation provided is consistent with the intent of CSAU Step 5. Specific steps necessary for the_W submittal of the approved version of WCAP-14449(P) is provided in Section 3. CSAU Step 5 is also discussed in TER Section 3.1.4.

Step 6. Determination of Code Applicability - In this step, code capabilities for the given scenario and plant design are qualitatively assessed by relating four analytical elements, field (conservation) equations, closure equations, numerics, and structure and nodalization, to the modeling requirements identified as important in the PIRT developed m Step 3. In doing so, this step combines the findings from the previous 5 steps to make a code applicability finding and to identify any modifications needed to make that finding. Our review has considered this step in establishing that WCOBRATTRAC is applicable for use in this methodology. ,W evaluated I WCOBRA/ TRAC for applicability to LBLOCA analyses in a fashion like the CSAU l determination, and additionally evaluated and justified that applicability for use on its i 2-loop UPI designs. We have reviewed the_W applicability evaluation, and considered the intended application in our technical evaluations of the EM. We found the W evaluation acceptab!e because it demonstrated that the WCOBRA/ TRAC EM provides the capability of modeling phenomena identified as i important in simulating the scenarios and plant designs for which the methodology is intended. TER Section 3.1.5 also discusses CSAU Step 6. We conclude that the code applicability determination is consistent with CSAU Step 6. i 2.1.2 Assessment and Ranoina of Parameters NUREG-5249 categorizes Steps 7 through 10 as the " Assessment and Ranging of Parameters" element. This element assesses the code capabilities to calculate processes important

5

- (identified in the PlRT) to the scenario (PWR LBLOCA) against expe'imental data to determine code accuracy, scale-up capability, and ranges of parameter variations needed for sensitivity studies.

~

Step 7.- Establish Assessment Matrix - in this step, with reference to the PlRT table (Step 3),

an assessment matrix of separate and integral effects tests is developed to best address the important phenomena and components. Our review considered the adequacy of the data base for evaluating the code accuracy in calculating the important phenomena identified in the PIRT, the capability of the code to scale up the phenomena to nuclear power plant conditions, and the influence of nodalization on the calculation. Our review of the W PlRT is discussed in TER Sections 3.1.2 and 4. W has established such an assessment matrix for 3- and 4-loop designs, including about 100 tests, which was reviewed and found to cover all the phenomena identified as important in the W PlRT. In WCAP-14449(P) W provided assessments, considering Upper Plenum Test Facility (UPTF) and Cylindrical Core Test Facility (CCTF) data, to evaluate the capability of WCOBRA/ TRAC, MOD 7A, Revision 1 to calculate the phenomena identified in the PIRT as being more important for UPI LBLOCA analyses. We conclude that W has established an assessment matrix consistent with CSAU Step 7.

Stsp 3. Nuclear Power Plant (NPP) Nodalization Definition - NUREG-5249 discusses the trade-offs and comparisons in determining an adequate NPP nodalization. TER Sections 3.2.2 and 6 provida a detailed discussion of the review of the nodalization. W has defined the nodalization to be used for the 2-loop design which is roughly equivalent to the CSAU nodalization, and supports the nodalization by direct comparisons to test data. This was done similar to the way it had been done to support the nodalization for 3- and 4-loop designs. We conclude that W

~

has defined a 2-loop nodalization to be used for its 2-loop UPI designs consistent with CSAU Step 8.

. Step 9. Definition of Code and Experimental Accuracy-In this step, simulations of the experiments from the Step 7 assessment matrix using the NPP nodalization defined in Step 8 are used to determine a minimum value for the code accuracy. In addition, the W methodology determined a code bias and uncertainty for the important phenomena identified in the PIRT by ranging them. TER Sections 3.2.3 and 7, and WCAP-14449(P) provide additional discussion on this subject. We conclude that the W has defined code and experimental accuracy consistent with, but not the same as, that in CSAU Step 9. A more detailed discussion of the overall uncertainty methodology is provided in Section 2.2.

Step 10. Determination of the Effect of Scale - In this step, the potential effects of scale (plant versus experiment) on unce:tainty are assessed. This has been covered in our review, as discussed in TER Sections 3.2.4 and 8. The effect of scale is addressed by W for the 3- and 4-loop version of the modelin essentially the same way as CSAU except that W has also considered full-scale UPTF data in its assessment. For the 2-loop UPI version, W also showed that COBRA / TRAC results gave scaling trends similar to generic trends noted for UPI designs in the 2D/3D program. We conclude that the W has determined the effect of scale in a manner consistent with CSAU Step 10.

2.1.3 Sensitivity and Uncertainty Analysis NUREG-5249 categorizes Steps 11 through 14 as the " Sensitivity and Uncertainty Analysis" element. In this element, the effects of individual contributors to the total uncertainty are

. L ~~ ~

_ _ ~ ~ ~. u:- _ . _

p  ;

L __ _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __

I 6

obtained and the propagation of uncertainty through the transient is accounted for. Section 2.2 provides a simp!!fied description of the WCOBRA/ TRAC EM and its use in accounting for propagated uncertainties for either the 3- and 4-loop version or for the 2-loop UPI version.

Step 11. Determination of the Effect of Reactor input Parameters and State -This step accounts for the uncertainties in plant calculations that may result from uncertainties in the plant operating state at the initiation of the transient. The W methodology for 3- and 4-loop plants includes sensitivity studies to determine the effect of these inputs and boundary conditions as a l

basis for treatment of their values as either nominal, bounding, or explicitly treated in the 3

uncertainty methodology. Key reactor state parameters identified from this step were average  !

reactor coolant system (RCS) temperature, RCS pressure, accumulator parameters (pressure, i temperature, and volume), safety injection temperature, hot rod total peaking factor, and parameters describing the axial power distribution.

For UPl plants, W determined that condensation and interphase drag in the upper plenum and parts of the core were also important parameters to be included in the uncertainty evaluation.

Sensitivity studies showed that 2-loop plant analyses are not sensitive to vessel and downcomer condensation. Sensitivity studies also demonstrate that split breaks are expected to be limiting for UPI plants. The limiting split break is identified deterministically. Based on this, interphase drag and condensation in the upper plenum and parts of the core are ranged in the global conditions response spectrum in place of the break discharge coefficient and vessel and downcomer condensation. This is discussed in greater detailin TER Sections 3.3.1 and 9. We conclude that W has determined the reactor input parameters and state consistent with CSAU Step 11.

Step 12. Performance of NPP Sensitivity Calculations -In this step, sensitivity calculations are performed to determine code output sensitivity (PCT and metal-water reaction) to various plant operating conditions that arise from uncertainties in the reactor state at the initiation of the transient. Sensitivity calculations were performed for representative 3- and 4-loop design plants, typical of the classes of plants to which the methodology described in WCAP-12945 would be applied. These sensitivity studies were used to develop response surfaces for use in the overall uncertainty methodology. These studies and their results were included in our review of WCAP-12945. We concluded that W had performed NPP sensitivity calculations consistent with CSAU Step 12.

For 2-loop designs, sensitivity calculations were performed to develop response surfaces as was done for the 3- and 4-loop designs, using the appropriate set of parameters for 2-loop design., as identified in Step 11. We conclude that W performed 2-loop sensitivity studies consistent with CSAU Step 12.

Step 13. Determination of Combined Bias and Uncertainty - In this step, the individual uncertainties resulting from code modeling of important phenomena, scale effects, and NPP input variations are combined. W has developed response surfaces from the sensitivity studies in Step 12 and has determined uncertainty for initial and boundary conditions. This has not

changed from what was done for 3- and 4-loop plants. We conclude that W has determined combined bias and uncertainty for 2-loop UPI plants consistent with CSAU Step 13.

m _ _ _ _ - - - -

i 7

Step 14. Determination of Total Uncertainty-In this step, the statement of total unce tainty is given as the probability for the limiting value of the primary safety criteria. As noter earlier, the WCOBRA/ TRAC EM uses a 95th percentile estimate to demonstrate compliance with the "high

- - probability" requirement of 10 CFR 50.46. Determination of the total uncertainty is accomplished in the W methodology by a Monte Carlo process with sample response surfaces that closely parallels the CSAU process in establishing a PCT at a high level of probability. This has not changed from what was done for 3. and 4-loop plants. We conclude thatE has determined the total uncertainty consistent with CSAU Step 14.

From the above comparisons, we conclude that the WCOBRA/ TRAC 2-loop UPI EM is essentially the same as the methodology approved for 3- and 4-loop plants and is structured consistent with the CSAU methodological ptocess, although some differences, reflecting the intended use of the methodology to address licensing requirements exist for a vari 3ty of similarly designed plants with varying normal operating conditions.

2.2 Simotifie'd Descriotion of the Westinohouse Realistic LBLOCA EM and Uncertainty Evaluation The WCOBRA/ TRAC EM is used to perform analyses to demonstrate conformance with the l requirements of 10 CFR 50.46. The specific quantified output of this EM are parameter values i for comparison to the performance criteria of 10 CFR 50.46(b)(1-3). The parameter of direct I

interest is the calculated PCT. Local and core-wide oxidation values are also calculated.

Figure 2 depicts this simplified description of the uncertainty evaluation for the WCOBRA/ TRAC EM, which is a pictorial representation of the EM implementation of CSAU elements 2 and 3. As stated in the previous section, we have concluded that the methodological process for 2-loop UPI plants is essentially the same as that approved for 3- and 4-loop plants described in WCAP-12945 (P-A). Figure 2 is the same as that used to represent the 3- and 4-loop methodological process. We have concluded that the technical adjustments made to accommodate the 2-loop UPI design do not represent substantive changes that would change the overall methodological process, and they do not appear in the figure. A more detailed and precise description of the uncertainty evaluation of the WCOBRA/ TRAC EM and its implementation is given in TER Section 2, in Sections 2,6, and 9.2 of the TER for the 3- and 4-loop methodology, and in WCAP-12945-P-A, Volume 5.

The uncertainty evaluation process begins with a calculation of PCT for a " nominal case" l

LBLOCA event for the subject plant, in which most inputs are set at nominal values, a limiting

single failure is assumed, and some inputs are set at bounding values based on sensitivity studies.

l l

The remainder of the analysis determines an adjustment to the nominal PCT to account for analysis uncertainty as propagated through the LOCA transient. This adjustment is quantified by iteration throut.1 two major response surfaces and a distribution function.

The first response surface, covering uncertainties in power-related values, is constructed based on nominal case results and provides a first step adjustment to the nominal PCT for each une..~..nn - . - . . . . . ~ - - - -

• l - _

_ +-..--_3-~ _ _ _ . _ _ _

. ' -  ! ' iI ' .;>i,,  ; , '

T C

P S A..;::............:..:.. ...: ...: . . . . . . ,

S t o

., > > mr &r m, > mm' .r > > I m- > >> m -

E p ,

S C

,p t

o H

O s..:

n

....w ........* .:..

. . . . . , B R i o e c

P t i

d n

f a

r u

o S N C e s

n O

l a o b - p I o s T e l

G...: :.:.............- R -

A C _

N I

P l

a n

M i m

o R N E CT 2 n

T P a ,, y m- ,, .,

i oer .

E As t

b i

uug D r

_ n t si o

i t

iF D

Y i d

n C

P t y

n T

_ o l l

e C a N

t l i n r e

a m c I i t

o n A

i n U

.:.............N T T C

I. ..:

R P E A A e

n C

c i

o f a

_ t r N u u

b i

S U t i

r s

e s

n D op M r e

w R s

e E P o

E

_ T C

B P l

a W i n

m o

N

_ ei* iI: ,:,,1 , . -.,  ! t,. j, t.:.l.; ,[!

l

.m m .

8 iteration. The next step adjustment to the nominal PCT for each iteration, covering uncertainties in plant initial conditions, is a distribution function based on uncertainty distributions determined by comparisons to plant data and which are quantified by WCOBRA/ TRAC analyses.

+ - -The second response surface is a composite of " global" thermal hydraulic effects and " local" (hot spot) effects. Each of these effects is related to the other and has its own associated distributions. At least 9 WCOBRA/ TRAC runs are used to create a response surface for the

" global" effects and several hot spot paths for each " global" path. Included in these analyses are the uncertainty distributions for the important thermal-hyJraulic models identified earlier in the CSAU process. In addition, this response surface accounts for the propagation of uncertainties related to interaction of the fuel rod modeling parameter and heat transfer coefficient uncertainties. This response surface is based on nominal case results and provides a third step adjustment to the nominal PCT for each iteration.

The two major response surfaces, and the distribution function, and their results are treated separately, and each response surface, and the distribution function, contributes its own PCT adjustment for each iteration. This separate treatment concept, proposed based on the relative independence of the important phenomena groupings being ranged, has been called l " superposition." A final adjustment or correction to the PCT is made to account for any I inaccuracies resulting from propagated uncertainties, superposition and selection of limiting break type assumptions is performed at the end of each iteration. The superposition concept was the subject of considerable review attention during the 3- and 4-loop model review, including the analytical adjustment,

', . 2.3 Comoarison with RG 1.157 Guidance i

in its review of WCAP-14449(P), the staff found the.W methodology to be consistent with the CSAU methoc. logy. RG 1.157 references the CSAU approach as a methodology that implements the MG 1.157 guidance. Compliance with RG 1.157 guidance is implicit from consistency with the CSAU process, and from the sameness of the 2-loop UPI methodological process and that of the approved 3- and 4-loop methodology. For these reasons, we conclude that the W 2-loop UPI methodology conforms with RG 1.157 guidance.

2.4 Technicallasues The above sections discuss our review of the overall WCOBRATTRAC EM process. Our review also included a comprehensive in-depth review of technical details of the 2-loop UPI EM sdjustments to the approved methodology, including code models and correlations, plant models, empirical data comparisons, references, uncertainty and propagation of uncertainty, uncertainty distributions, assumptions, and applications. These were also found acceptable.

The TER provides a detailed description of the review in these areas. Certain issues are discussed in a more summary fashion in this section because of their significance either due to l- technical controversy or importance (PCT impact) as identified in the WCOBRA/ TRAC EM PlRT.

rrrr ~ a r-_ r - -- - -- --

g., .: - ..-

9 2.4.1 Compensatina Errors Compensating errors are errors which offset one another in a given analysis such that the analysis produces apparently accurate results, however, there is a potential for these errors to compound one another in a different case such that significantly inaccurate results could be produced. The concem is that compensating errors could go undiscovered in a benchmarking analysis because of the apparently accurate results, and result in a significantly inaccurate analysis in an application case. Usually these errors are discovered by studying the benchmarking results to identify two or more parameters which may behave consistently with the analysis result, but exhibit behaviors which are mutually inconsistent.

TER Section 12 provides a detailed discussion of compensating errors. This section discusses that in WCAP-12945, .W evaluated post-CHF heat transfer, downcomer ECC bypass / condensation, and blowdown / post-blowdown thermal hydraulics /entrainment as the principal potential sources of compensating errors for 3- and 4-loop plants.

WCAP-14449(P) additionally evaluated the core outside the hot assembly, upper core plate countercurrent flow limit, and entrainment from the upper plenum tc hot legs and steam generators as potential sources of compensating errors for 2-loop UPl plants. The TER concludes that Westinghouse has appropriately accounted for compensating errors in the uncertainty methodology. We concur with this finding and the TER statement that no compensating errors were found that would compromise the ability to perform realistic LBLOCA analyses with the methodology. Therefore, we find the Westinghouse treatment of compensating errors acceptable.

2.4.1.1 Post CHF Heat Transfer I

For 3- and 4-loop plants compensating errors were found in post CHF heat transfer in which pre-rewet heat transfer was underpredicted and rewet temperature was overpredicted. This was addressed by using a lower bound T. in the HOTSPOT analysis for the blowdown cooling

. phase. Two-loop UPI hot assembly response was found to be similar to that for 3 and 4 loop plants. It was concluded that the 3 and 4 loop compensating error assessment for the approved 3- and 4-loop plants methodology for post-CHF heat transfer applies to the hot assembly in 2-loop UPI plants. This is further discussed in TER Section 12.1. We agree with the finding that the approved methodology applies to the 2-loop UPI design.

2.4.1.2 Downcomer ECC Bvoass/ Condensation For the accepted 3- and 4-loop methodology, no significant compensating errors were found in '

the area of downcomer ECCS bypass and condensation in the vessel; and it was concluded that ,

i the WCOBRA/ TRAC treatment of downcomer ECCS bypass is conservative. The downcomer I and lower plenum condensation was accounted for in the uncertainty methodology of the l approved methodology by ranging it over the values in applicable test data.

For the 2-loop UPI methodology,.W sensitivity studies indicate that ECCS bypass ends before UPI injection begins, and that subsequent reestablishment of ECCS bypass would not occur. It

+

W e- .se -

. m- ~  :-

p_ .

- . - -~.- -- -

10 was concluded that the 3- and 4-loop compensating error analysis applies to the 2-loop UPI methodology as well.

W sensitivity studies also indicate that vessel condensation has a negligible e#ect on calculated _

~

PCT for UPI plants.

The review also considered a compensating error resultant from WCOBRA/ TRAC underprediction of interfacial heat transfer coefficient and overprediction of interfacial area in one dimensional components. It was found that this has less of an affect on compensating errors for 2-loop UPI designs than for the 3- and 4-loop designs to which the approved model applies and that,' therefore, the conclusions of the approved methodology apply to the 2-loop UPI methodology. These issues are is discussed in greater detail in TER Sections 12.2 and 12.5.1, which conclude that the compensating error analysis for downcomer ECC bypass and condensation in the vessel from the approved 3- and 4-loop methodology also applies to the 2-loop UPI version. We agree with the TER conclusion.

2.4.1.3 Blowdown / Post-Blowdown Thermal-Hydraulics /Entrainment in the approved 3 and 4-loop methodology, though no significant compensating errors were i

found in the treatment of blowdown / post-blowdown thermal-hydraulics /entrainment, l overprediction of entrainment from the core to the upper plenum and the upper plenum te 1:e  ;

l hot legs and steam generators was identified. This was not found to be a significant compensation error because it was considered conservative due to a resulting increase in

- steam binding. For 2-loop UPl plants, the compensating error analysis of the approved 3- and 4-loop methodology was found to be applicable for the blowdown phase and, in the hot assembly, during reflood as well. However, because of the differences between the designs, for those phenomena in the rest of the core, countercurrent flow limit (CCFL) at the core plate, and entrainment from the upper plenum to the hot legs and steam generators, during reflood, the compensating error assessment from the approved 3- and 4-loop methodology was found to be inapplicable, and was addressed separately in the WCAP-14449(P) review. This is discussed in TER Section 12.3 With regard to 2-loop UPI thermal-hydraulics /entrainment during reflood, the 2-loop UPI methodology addresses core behavior outside the hot assembly, upper core plate subcooled CCFL, and entrainment from the upper plenum to the hot legs and steam generators by ranging interfacial condensation and interfacial drag coefficients, as quantified and verified by GE, UPTF, and CCTF data, by scaling studies. TER Section 12.4 discusses these issues and recommends their acceptability. We agree with the TER.

2.4.1.4 Other Compensatina Errors Other compensating errors were considered in the review; (a) broken loop hot leg mass flow, (b) FLECHT-SEASET Test 31701, excessive entrainment in gravity reflood tests, (c) overpredicting heat transfer in INEL film boiling tests, and (d) overpredicting entrainment and underpredicting heat transfer to vapor. These items were addressed either that the 2 loop UPI l design was sufficiently similar, with regard to the given concern, to the 3- and 4-loop design that the approved 3- and 4-loop compensating error analysis applied to the 2-loop UPI version of the l ,

1

g. 49 maaume p

p WJ e, a e _ -- _ -*p g as he@ -

m _.__.__._- .. q

.2 ,

11 methodology, or by showing that the behavior of the 2-loop UPl design was different from the 3-and 4-loop design in such a way that the concern does not apply to the 2-loop UPI version of the methodology. These issues are discuss in detail in TER Section 12.5.

2.4.1.5 Conclusion Reoardino Comnensatina Errors TER Section 12.6 provides the conclusion regarding compensating errors that the 2-loop version of the approved methodology properly accounts for UPf-specific compensating errors and does not have any serious compensating errors that would compromise the ability to perform realistic analyses of UPI LBLOCAs. We agree with this conclusion.

2.4.2 Uncertainty Methodoloav l

As stated in Section 2.2, the exchange of parameters to be ranged in the global conditions response spectrum to accommodate the 2-loop UPI design does not affect the uncertainty methodology as depicted in Figure 2. The parameters that are exchanged do not change the statistical framework of the methodology. The determination of the correction factoris not changed. We conclude that the uncertainty methodology is essentially the same as that approved for 3- and 4-loop plants and is, therefore, acceptable.

2.4.3 Selection of Parameters to be Exehanoed For the 2-loop UPI version of the methodology, interphase drag and condensation in the upper plenum and parts of the core were selected to be ranged in the global conditions response spectrum in place of the break discharge coefficient and vessel and downcomer condensation, based on the PIRT (CSAU Step 4) and sensitivity studies. This is discussed in TER Section 2.2. l We find this consistent with CSAU processes and, therefore, acceptable.  ;

2.4.4 Determination of Ranoes and Biases Ranges and biases were determined based on applicable UPTF, CCTF, and CE data. TER Section 5 discusses this and recommends its acceptability. We agree with the TER.

In a letter dated April 8,1999, to assure that the 2-loop version of the methodology would not be applied for heat generation rates higher than covered by the UPTF and CCTF data, proposed to I limit the application of the UPI methodology to nominal power levels of 1980 Mwt, low power region average linear heat generation rate less than 6.9 kW/ft, and maximum analyzed linear heat generation rates of 17 kW/ft. We find the proposed limits are acceptable because they are consistent with the range of the UPTF and CCTF data.

2.5 Comoliance With 10 CFR 50.46 Reauirements Based on our review, as summarized above, we conclude that the WCOBRA/ TRAC EM described in WCAP-14449(P) is consistent with the guidance provided by CSAU (NUREG/CR-5249) and RG 1.157. Based on this consistency, we conclude that theE methodology complies with the requirements specified in 10 CFR 50.46 describing a " realistic" evaluation model, and that use of this model produces results consistent with

j. ,

m> _ ,_*_ . .-

_ . m_ __

_ _ _. . _. _. _ _Z - _J '

~~

l _ -- _ .. - ~ _...__- _ ._ _

12 10 CFR 54.46(a)(1)(i) that can be compared to 10 CFR 60.46(b) performance requirements.

We therefore conclude that the WCOBRA/ TRAC EM described in WCAP-12945 meets the requirements of 10 CFR 50.46.

3 LICENSING ANALYSIS PROCESS Section 2.2 and TER Section 6.1 describe the WCOBRA/ TRAC EM process for performing plant calculations. By this process, a nominal PCT calculation is determined using sensitivity studies to identify bounding values for certain input assumptions and parameters, including the loss-of- -

offsite power assumption, steam generator tube plugging, and limiting single failure. The inputs for this nominal, or reference, PCT calculation are used as the base calculation inputs to create a power distribution APCT response surface, an initial plant conditions APCT uncertainty distribution, and a global model conditions APCT response surface. All of the above process elements, nominal calculation,2 response surfaces, and uncertainty distribution, are constructed based upon plant specific WCOBRA/ TRAC calculational runs. Monte Carlo iterations through these response surfaces and uncertainty distribution, combined with the base nominal PCT give a spectrum of calculated PCTs, which are adjusted based on a function determined from additional WCOBRA/ TRAC plant specific calculations. This results in a final PCT spectrum from which a 95th percentile PCT is determined.

In WCAP-14449(P),2-loop UPI plant design analyses are presented to demonstrate use of the WCOBRA/ TRAC EM process. WCAP-14449(P) Table 5-10 (may change in approved version of the document) lists some typical input parameters which are used in the analyses. In addition certain plant and analysis specific information is used, such as identification of limiting break type and metal-water reaction calculation method. The information may vary from plant to plant due to design differences, but must be treated qualitatively in the same way as in the submitted and approved version of WCAP-14449(P). The staff finds the 2-loop UPI process for performing plant analyses described in WCAP-14449(P) acceptable.

As we have mentioned above, some of the quantification of information used and produced in the 2-loop UPI analyses may vary due to design differences. However, except for variances explicitly identified as part of the methodology covered in this review, qualitative changes to the methodology such as, changing the nature of a treatment from deterministic to statistical or vice versa, changing the constituent parameters varied in constructing one of the response surfaces, or changing the processes relating the iterative components of the uncertainty analysis, would ,

require NRC review and approval. -

4 STAFF CONCLUSIONS On the basis of its review and our determination that the 2-loop UPi EM described in WCAP-14449(P) meets the requirements of 10 CFR 50.46, the staff concludes that the WCOBRA/ TRAC EM is acceptable for use in 2-loop UPI Westinghouse design PWR licensing applications, including reference in plant technical specifications and core operating limits reports. Our review finds the 2-loop UPI version of the methodology acceptable. However, authorization of the reference to WCAP-14449(P) in licence amendment applications and other licensing actif. s is reserved until completion of the documentation requirements as discussed in Section 2.1.1, dtep 5.- This documentation shall contain:

-~

13

a. All information submitted and docketed byW related to the 2-loop UPI version of WCOBRA/ TRAC EM (including information important to the findings in this SER, and longer term material related to this review but not important to our findings) to provide a traceable record of the WCOBRA/ TRAC EM and this review.

b. WCAP-14449(P) structured consistent with the 2-loop UPI adaptation of the methodology. The WCAP will provide up front a road map describing the overall 2-loop UPI EM approach and comparing it to the acceptable CSAU approach. It should also provide in an orderly and understandable fashion descriptions of the approved individual correlations pertinent to this adaptation and how they were determined to be appropriate, descriptions of the as-approved validations performed on the code, and description of the approved uncertainty evaluation process. It should be indexed such that information is readily located, and contain in appendices historical review information, such as questions and accepted responses, and original report pages that were replaced. This approved version of WCAP-14449(P) should be provided for staff review in both proprietary and non-proprietary versions within 3 months of issuance of our letter of acceptance.

TER Section 15 provides a discussion of the limits of the scope of our review and conditions of acceptance. We concur with these provisions and have confirmed Westinghouse agreement.

The following are applicability limits, based on scope of review:

1. The use of the WCOBRA/ TRAC EM for long term cooling licensing analyses is not covered in this review.

2. Our review of the WCOBRA/ TRAC EM has been limited to licensing application of the methodology for 2-loop UPI Westinghouse designs. Other designs, such as advanced designs (e.g., AP600), 3- and 4-loop plants, and other vendor designs were not covered.

3. Our review did not cover used of the WCOBRA/ TRAC EM for small break LOCA licensing analyses.

4. Section 2.4.4 of this SER discusses that ranges and biases of parameters were based on data, including UPTF and CCTF data. Of particular concem is the ranging of interfacial drag and condensation, which is based on UPTF and CCTF data. In a letter dated April 8,1999, to assure that the 2-loop version of the methodology would not be applied for heat generation rates higher than covered by the UPTF and CTF data,W proposed to limit the application of the UPI methodology to nominal power levels of 1980 Mwt, low power region average linear heat generation rate less than 6.9 kW/ft, and maximum analyzed linear heat generation rates of 17 kW/ft. We find the proposed limits are acceptable because they are consistent with the range of the UPTF and CCTF data, We also find that use of the methodology above these values is outside the scope of our review, and would require furtherjustification and NRC review.

' * * *~' -'"" "-* ""- ' - " " ' ' ~ ~ ~ ' "

-- . . , . +w * = = + ' - . - e

• .as-14 TER Section 15 also specifies usage conditions, which will be implemented by W in accordance with the TER. These include:

a. A recommended justification for any future time step changes (first listed item). W

~~ ~ ~ ~

will perform this justification as recommended, and retain traceable documentation of this action in its in-house plant records.

b. Under the designation " Westinghouse agreed / committed" items are 13 additional items: ,

i. Items 1,3,4.a 4.b,4.c,6, and 9 are plant specific calculational requirements and W will retain traceable documentation of these actions in its in-house plant records.

ii. Items 2,5,7, and 8 are generic limits on usage of the WCOBRA/ TRAC EM, which, if changed require submittal for NRC review.

iii. Items 8 and 9 require inclusion of specific information in the WCOBRA/ TRAC EM and documentation, including WCAP-14449(P) as appropriate. W willimplement these actions as part of providing the approved version and updates to WCOBRATTRAC in accordance with the TER discussion.

TER Section 15 recommends an additional review item in connection with possible changes to the WCOBRATTRAC EM. This item is covered by existing regulatory provisions of 10 CFR 50.46 related to reporting of code errors and modifications.

The Westinghouse agreed-upon items discussed above are carry-over items from the approved l

' 3- and 4-loop version of the methodology which have been determined to apply to the 2-loop UPI version.

i g e.w e. . - .-

+.,w .. .-=w..- . + . - . . - .

h

,