ML17347A593

Retran Code:Transient Analysis Model Qualification, Technical Evaluation Rept

ML17347A593
Person / Time
Site:	Saint Lucie, Turkey Point, 05000000
Issue date:	07/31/1987
From:	INTERNATIONAL TECHNICAL SERVICES, INC.
To:	Office of Nuclear Reactor Regulation
Shared Package
ML17342A775	List: ML17342A774 ML17347A593
References
CON-FIN-D-1350 ITS-NRC-87-02, ITS-NRC-87-2, NUDOCS 8707230432
Download: ML17347A593 (19)

Text

ITS/NRC/87-02 July 1987 Technical Evaluation of FPJ Topical Report:

RIKRAN Code: Transient Analysis Model Qualification Intmzmtional Ta2nical Sexvices, Inc.

420 Imcia3ton Avenue New York, New York 10170 8707230432 870713 PDR ADOCK 05000250 P.

PDR I

TABLE OF ~NXS Page 1.0 RMQQ

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

SuI

~

~

1 2.0 In~uau JLDAlCL4ulon e

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

2 3.0 Topical Report Ob)ectives

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

2 4.0 CoII~~ Y~el

'e plier je Qlellng

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

3 4 ~ 1 Code VerslonS e

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

4 4.2 Nodallzatlons.

~

~

4 4 ~ 3 XMltJ.lng e

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

5 5.0 Tra J

m

~

nSlent Zujalysls e

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

5 5.1 Turkey Point Zoss of One Main Feechmter Pump Event...

6 5.2 Turkey Point RCP Coastdown Test.............

6 5.3 Turkey Point Uncontrolled RCCA Withdrawal........

7 5 ~ 4 Turkey Point Small Break DXAe

~

~

~

~

~

~

~

~

~

~

~

~

7 5.5 Turkey Point Stuck Open Steam Generator Relief Valve..

8 5.6 Turkey Point Steam Generator Tube Rupture..

~

~

~

~

~

~

8 5.7 St. ZIIcie Natural Cixculation Cooldown Event......

9 5.8 St. ZIIcie - Main Steam Line Break.............

9 5.9 St. Zucie Zoss of Load 10 5.10 St. ZIIcie 'Generator Trip Test...

10 5.11 St. Zucie Main Steam Xsolation Valve Closure Event...

10 5.12 St. Zucie Inadvertent Opening of the PORVS.......

11 5.13 St. Zucie Loss of Foxced Flcar..............

11 5.14 St. Zucie CE'A Drop....

~

~

11 6.0 Conclus lons and E ecomendatio11s

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

12 7.0 ferences e

~

~

~

~

~

~

~

~

~

~

~

~

~

Re 14

1. 0 Summary The Florida Bmer and Light Ccangany (FPL) submitted a topical report

[1] for the purpose of demonstration of their technical competence to utilize the KK,RAN computer code for performing transient systems analysis for their Turkey Point and St. Zucie Plants.

The applicant has demonstrated the capability to utilize the BEIBAN computer code to perform systems transient calculations on their plants by a series of camparisons between the RERAN calculated results with

FSAR, Westinghouse Generic study results and some plant data.

However, no efforts were made by FPL either to qualify or verify RERAN models for use in transient analysis or to explain in depth the predicted txansient behavior, each of which would have provided a

basis for evaluation of their understanding of the plant behavior and the BETRAN code.

We, therefore, feel that the report does not contain sufficient materials to serve as a document which may be generally referenced to support future FPL licensing submittals.

We rceammend, therefore, that in future submittals, FPL should be, on a transient-by-transient

basis, asked (1) to thoroughly qualify RERAN models including nodalizations and control system modeling for their plants and (2) to provide thorough discussion of the transient reeQts.

2. 0 Introduction The FPL topical report entitled "RLTRAN Code Transient Analysis Model Qualification" was submitted to demonstrate the ted'.cal ccaagetence which FPL has developed for performU@ transient systems analysis with the RLTHAN computer code.

FPL presented 14 transients in the report:

8 of these are for St.

Zucie plants (Combustion Engineering plants) and 6 are for Turkey Point plants (3-loop Westinghouse plants).

Six (6) of these transients are compared to the FSAR analyses, five (5) are benchmarked with the actual plant data taken during the events or tests, and three (3) are compared to the Westinghouse Generic Studies.

A matrix of these transients is shown in Table 1.

We have reviewed the applicant's efforts to demonstrate their ability to utilize the REIGN computer code to perform transient analysis for Turkey Point and St. tucie Plants.

Out evaluation is summarized in this report.

3.0 To ical Re rt Ob ectives The applicant's original stated objective of this topical [1] were:

1.

to present R1;TRAN base model verification results for each of FPL's plants.

It was requested that these models be approved by the NRC for non-ZDCA licensing support analysis.

It was intended to demonstrate the adequacy of these REZBAN base models for non-IDCA licensing analysis.

2.

to demonstrate the proficiency of FPL personnel to perform system safety analyses per NRC Generic Letter 83-11.

Since the time the original topical report was submitted by the

applicant, the applicant informed the NRC of their desire to reduce the scope of review intended by this submittal as expressed in the applicant's letter L-87-91, dated March 2, 1987 [3].

The revised objectives of the topical report are to:

1.

show that the models FPL developed for the Turkey Point and St.

tucie plants adequately simulate the behavior of the plant under steady state and transient conditions, 2.

show that the FPL staff knaws how to apply the models to a

comprehensive and diverse group of transients, so that in future applications of the R1HBAN code in licensing submittals, FPL would have already addressed and had reviewed and approved by the NRC Staff the resolution concerns raised in Generic Letter 83-11.

Both of these points have been topics of lengthy discussion among the

FPL, NRC, and ITS staff.

FPL chose not to provide detailed justification for its nodalization, model selection and control system modeling on a

transient-by-transient basis or to provide detailed analysis at this time.

Instead, in the

future, FPL will either develop a

complete licensing methodology topical or will provide sufficient detail to support each licensing action individually. It is, therefore, our understanding that the level of approval now desired by FPL does not include these points in entirety but rather a portion of each, since otherwise FPL would be required to perform more extensive and thorough analysis than was presented in the topical report.

thus the focus of this review is upon the ability of FPL to apply the models they have developed for the Turkey Point and St. tucie plants to the series of transients to which they have been applied in this submittal.

4.0 ter Model

'n

general, the term "model'ncludes three items:

(i) the plant

nodalization; (ii) the user selection of phenamenological models fram among those programmed in the code (i.e.,

the bubble rise model or the non-equilibrium model);

and (iii) the user generated input which models the plants control systems.

Each of these is transient dependent.

FPL has not discussed its control system modeling and has only occasionally and briefly discussed its selection of phenamenological models.

4.1 Code Versions A variety of unidentified REIGN versions was used by FPL in the analyses pated in the report [2].

R1KEQN02/HOD02 has been reviewed and granted limited and conditional approval by the NRC for future use.

REZRAN02/HOD03 is a "correctecV'ersion of RETRAN02/MOD02 but has not yet been formally approved by the NRC for use in licensing analysis.

Nevertheless, for the purpose of demonstrating technical competence (but not for model qualification), in our opinion it is acceptable to use any of the versions of the code.

FPL has developed a single-loop plant nodalization for natural circulation cooldawn test at St. tucie Unit 1 using REXBAN01.

While this is an acceptable approach in gaining experience in model development and insight into the transient behavior, any future submittals must be performed with an approved version of R1HBAN02.

Mo-loop nodalizations were also developed and used in the analysis for both plants.

Although both of these plant models are briefly described in the report and the nodalization diagrams are presented, this is generally presented without justification since FPL chose not to seek approval for any particular combination of input and RERAN version for any specific application [3].

In future submittals for qualifying licensing

models, sensitivity studies leading to justification of the nodalization

used in the analysis must be presented.

4. 3 Modelirg Following up the election mentioned in the preceding

section, very little justification of phenomenological or contxol system modeling selection or discussion of the limitations of these models was provided in the report.

lherefore, the applicant must justify its selection of phenomenological models and demonstrate that each is being used within its range of validity on a transient-by-transient basis.

In addition, the applicant must also justify its modeling of the plant's control system on a transient-by-transient basis.

5.0 Transi t Anal sis The majority of the transients were compared to results submitted in the Final Safety Analysis Reports (FSAR) for St. tucie and Turkey Point.

As

such, the presentation of results and discussions were vexy brief.

No discumions of the possible origin of differences between FSAR and FPL analyses were provided to the reviewer, which makes the evaluation of the applicant's understarxHng of the computer codes and models and its understanding of the specific transient difficult to evaluate.

The applicant, in response to our questions, stated in several instances that it could not obtain enough details concerning the code used in the FSAR or the Westinghouse studies, the initial conditions or the assunptions in the FSAR analysis to be able to provide model ccangarison or to be able to explain the differences in results.

Kherefore no conclusions can be drawn from these II analyses which can be used by the applicant to justify models.

Furthermore, because of the lack of detail in the applicant's own analysis, generally, these analyses do not significantly contribute to the applicant's demonstration of its analytical capability.

Camparison with Westinghouse Generic Assessment Studies was also lacking details and the plant analyzed was enmesh different in some cases that only a ted could be em@ared.

Therefore, those analyses do not serve to qualify plant specific models, althaugh the trend analysis does support, within the conditions mentioned, the applicant s demonstration of ability to utilize the code.

Nevertheless, portions of the applicant's analyses of a

Small Break

ZDCA, a

Steam Generator Tube Rupture and a

Steam Line Break accident contain good discussions of selected phenamenology going on during those transients, and indicate a

sound understanding of those portions of those transients.

Future submittals containing analyses of entire transients to that degree of discussion together with appropriate cross reference to, and explanation by use of plots of computer output of, various plant parameters would be sufficient to meet the requirements for demonstration of technical camp etence.

5.1 Turke Point lass of One Main Feedwater The agreement between the code calculations was not very good for the loss of one main feedmter pump event.

This particular event was not very helpful in model qualification since the plant monitoring camputer failed to record the exact sequence of events and the cause of the reactor trip, and therefore same of the key plant p-~meters such as core power, hot.

aTKl cold 1eg terIg38rature I SG Iluxture 1evel and feedmter f1car rate were not available for camparison.

Even the measured feedmter flow rate after the reactor trip was judged to be erroneous.

Notwithstanding this difficulty, FPL managed to obtain the global Umxi.

5.2 Turke Point RCP Coastdawn Test

%he results from the Reactor Coolant Pump Coastdawn Test showed that the flaws for the case of one pump coastdawn appear to begin diverging at about 5 seconds before the test was terminated.

FPL indicated that the analysts believe the divergence was unimportant because they presented other analytical results which indicate good results dawn to less than 20 4 flcar.

'Ihus, if the applicant had used an appraved version of the code, it would

have justified its pump model for the region dawn to a flow of roughly 40%

of nominal where the ~ts begin to diverge.-

5.3 Turke Point Uncontrolled RCCA Withdrawal This is a physics daminated transient for which the camparison is between

codes, not to experimental data.

%he camputed results appear to diverge toward the end of the calculations.

The applicant stated that both the REIBAN and FSAR analyses used the slaw wittx~wal case with a reactivity insertion rate of 2.5 x 10

@sec, yet the reactor trip times differ by roughly 5 seconds in a relative1y short transient.

FPL did not ana1yze the difference, but instead simply stated that the reason for difference was because "these calculations were performed with different

codes, by different organizations (leading to diffuses in input) at different times'.4 Tur e Point Smal Break Although only code camparison work was submitted in this section, the thorough discussion of phenamena during major portions of this transient demonstrates a

good understanding of the

code, the phencanena in the transient, and haw to do analyses, etc.

Hawever, the applicant did not accampany those general phenamenological explanations by cross refmmme to and explanation by use of plots of the appropriate plant parameters and therefore there are a number of unsupported stateaents (such as a reference to the existence of countercurrent flow from the top steam generator volumes and the upper head),

and it is therefore difficultto evaluate whether these phenamenological discussions are co~.

In addition, these discussions do not cover the entire trzmient, and therefore the analysis is incamplete.

Figures presented in the topical show that FPL was able to follaw the txerds of the transients and since there was no athmpt to simulate an identical case by using identical initial conditions and assumptions, this is all one can 6)g38cto

5.5 Turke Point Stuck Generator Relief Valve The applicant indicated that.

although most of the initial conditions were obtained fram the Westinghouse Generic Study, no attempt was made to exactly match the Westinghouse results.

Same assumptions were changed to plant specific rather than generic or were made more conservative to produce a more limiting calculation.

Results indicate that the purpose of the analysis was accomplished in that RERAN began camguting lower inlet core coolant temperature at roughly 1300 sec and as much as 30 F by 3500'F.

1'PL resgonse provided plausible explanations of certain of the differences between the FPL analysis and that of Westinghouse, (for example, the report suggests that the slower repressurization rate observed in the RERAN prediction, which roughly starts at 600 sec, is due to a difference in the tea~wtures of the charging fluid (R1HBAN used 40 'F and the flaw rate and tenperature in the Westinghouse analysis were not knawn) )..

Nevertheless, this analysis does not discuss the details of the plant tater behavior, and is therefore incamplete.

5.6 Turke Point Steam Gene tor Tube Tu ture

'Ihis analysis was performed as part of PIS considerations and input assumptions were ta)o n fram the Westinghouse Generic sb.may and the end point was different than one would use for SGIR analyses.

%his cmgmtation was ternLnated at 600 sec when the primary and secondary pressures became stabilized (not ecpilibrated) and therefore the break flaw stabilized at about 100 lb/sec.

%he FPL ~nse included several good discussions of phenamena occurring during the transient which support the conclusion that the applicant has the capability to interpret the results, but as in its presentation of the

SBIDCA, the applicant did not either support its phenamenological discussions by cross reference to appropriate plots or caver the entire transient.

5.7 St. tucie Natural CirnQation Cooldawn Event FPL's topical report stated that "system action and system parameters were modeled to represent actual operating conditions as closely as possible."

However, in the response FPL stated that the "FPL REXBAN model for this event was deliberately constructed to provide over-prediction of voiding in the upper head,, so as to provide conservative limiting reactor cooldown rates" and not to simulate this event closely.

The results do indicate that KHBAN is computing the onset of upper head void formation at roughly 3.75 hours8.680556e-4 days <br />0.0208 hours <br />1.240079e-4 weeks <br />2.85375e-5 months <br />, which is about 0.25 hour2.893519e-4 days <br />0.00694 hours <br />4.133598e-5 weeks <br />9.5125e-6 months <br /> earlier than the plant data indicated (see the figure provided with the FPL response)

FPL also states that in 1970 they performed an mctension study of natural cirniLation within the upper head which lead to their current nodalization.

Presentation of these results should be contained in any future submittal employing and relying on that nodalization.

5.8 St. tucie - Main Steam Line Break In comparing the main steam line break transients analysis with EKTRAN to the vendor's CESEC computations, the two calculations did not shear good agreement in the return to power, which is the primary concern of this transient.

%he FSAR analysis predicted a return to pcmer of about 1% while REZRAN did not.

FPL attributed the difference in return to power to a difference of roughly 0.4%

between the total reactivities in the two calculations. In addition, a slightly higher Doppler reactivity was used in the K%RAN calculation.

It seems to us, hcarever, that the differences may be due to CESEC's treatment of the vessel and core flaws which account for asymmetric effects and which conservatively compute reactivity insertion.

Since FPL did not use a split-core, they may not have simulated the CESEC modeling.

FPL had a

good general discussion of the thermal hydraulic behavior of the privy, but that discussion was not supported by cross reference to appropriate plots, and it is therefore difficultto assess the accuracy of those discussions.

5.9 St. tucie Loss of load Results of the FPL calculation for a loss of load transient did not agree well with the vendor calculated results.

%he sequence of events in the topical report covers only the first 14 seconds of the transient when there was a

rapid heatup of the primary side, after which there was substantial divergence.

In their response, FPL demonstrated that initial pressurizer level had a substantial impact on ensuing pressure response.

However, it appears that it would also have been fruitfulto investigate the impact of assuaed primary to secondary heat transfer coefficients and secondary side modeling since it appears that there are substantial differemes between the CESEC and RETRAN predictions of second-~

side behavior.

5.10 St. tucie Generator Tri Test FPL attribute the difference in the test data and REXBAN results at roughly 10 sec to the one-node steam generator model.

FPL states that they recently conducted sensitivity studies with a multi-node SG model, where the SG dame was explicitly modeled, which showed that the sensitivity to the noding can explain a sudden change of SG pressure of this magnitude (20 to 30 psia).

The resulting differences in plant parameters may be explained by this change and should, therefore, be included in any future FPL submittal in respect of their analysis of this transient.

5.11 St. tucie Main Steam Isolation Valve Closure Event Results of the FPL calculation for this transient show a

reasonable agreement with the test data.

%he FPL response suggests that the slight discrepancy in the pressurizer level'was due to an error introduced in converting the output into the ccaoparable units for plotting purpose.

10

5.12 St. tucie I dvertent 'fthe TORVs

'lhe results show that two calculations agree reasonably well for the first 40 seconds.

Around this time, the REIRAN computed secondary side pressure begins to behave very differently from the FSAR pressure which results in the main steam safety valves cycling differently due to the fact RERAN ccnqMtes the SG pressure to increase much slower than does CESEC.

FPL has attributed the difference to differences in SI flear, break flear and pressurizer models but has submitted no parametric analyses or other supporting analyses.

This response is not helpful since it is merely speculative.

5.13 St. tucie ?ass of Forced Flmr Relatively good agreement was achieved between the REXBAN and FSAR calculated results.

'Ihere is, however, some inconsistency, since the pressurizer pressure was 15 to 25 psi higher in the REIBAN calculation while the core outlet temperature was roughly 5 'F lower in the RERAN ccaqmtation at the peak.

%he FPL response suggests that this is due to slight differences in the physics parameters, which does not explain the inconsistency.

5.14 St. tucie - CEA, Dro The FPL response implies that the source of the difference exhibited by the R1TRAN and FSAR results with respect to the treatment of reactivity is due to the fact that CE assumed 1.2 seconds to insert the CEAs while FPL assumed

3. 0 seconds.

This difference between the basic assungNions makes the value of the caparison questionable.

11

6. 0 Conclusions and Recxemendatio In conclusion, the section on qualification does demonstrate that FPL is able to utilize KKRANto perform systems transient calculations on their plant.

However, there are a naker of outstanding issues which must be discussed and reviewed in greater depth, prior to acceptance of any specific plant analysis by FPL.

The comparison of the single-and two-loop models to operating reactor data provides a limited measure of qualification of their plant models and usage of REZRAN but FPL elected not to seek approval for any specific combination of input or code version for any specific application 't this point.

lherefore, additional comparisons between computed results and test data, together with appropriate nodalization and sensitivity studies should be submitted in future reports using an appmved version of RETBAN if these models are to be used in licensing suhmittals.

%he dition of the code calculated transient results was frequently minimal and was expanded to same degnm in the ~nse

[3].

In the future, reports which are submitted must contain a detailed discussion of the transients submitted.

%his would provide the reviewer with the proper material to make a determination as to whether the analyst has the basic understanding of the code calculations and enable ccaaplete evaluation of the submitted material for the desired purpose.

Future submittals, in order to be sufficient for approval,

should, on a transient-by-transient basis:

(a) describe and justify nodalization; (b) describe and justify control system modeling; (c) describe and justify user selection of phenomenological models and demonstrate that such models are being used within their range of applicability; and 12

(d) present enough plots of physical pamneters throughout the plant and discuss the major changes in slope in each by reference to physical phenomena and by cross reference to and with explanations based on the plots of other variables.

7. 0 References 1.

"PITRAN Code: Transient Analysis Model Qualification," FPL Report

'Qi~6, July 1985 2.

FPL's

Response

to Request for Additional Information, L-87-164, April 10, 1987.

3 ~

FPL ZBtter L 87 91 froHl C 0

'Woody (FPL) to NRCg March 2 I 1987 14

Table l.

FPL Transients Transient Description Plant Natural Circulation Oooldown St. Zucie Plant Data 1-loop REZiUKOl model Main Steam Line Break St. Zucie Zass of one Main Feechater Bmp Event Plant Data Loss of Ioad St. Zucie Generator Trip Test St. Zucie St. Zucie Plant Data Inadvertent PORV Opening St. Zucie Turkey Point Test REBQHOl plant model

Table 1. Gont'd loss of Forced Flow St. tucie RCCA. Withdrawal Turkey Point St. tucie

~mk Open Steam Generator Relief Valve Steam Generator Tube Rupture Turkey Point

• Smeary Report on Reactor Vessel Integrity for Westinghouse Operative Plants WC'-10019, December 1981.

July 13, 1987

~

~

DI "IBUTION

/Do t-Pil'e w/o en'~,

.PDa Reading w/o encl.

D. I)ilier w/encl.

D. NcDonald w/encl.

'h DOCKET NO(S).

50-250 and 50-251

- fir. C.

O. Moody Group Vice President Nuclear Energy Florida Power and Light Company Post Office Box. 14000 Juno Beach, Florida 33408

SUBJECT:

TURKEY POINT UNITS 3 AND 4 The following documents concerning our review of the subject facility are transmitted for your information.

Notice of Receipt of Application, dated

~

Draft/Final Environmental Statement, dated Notice of.Availability of Draft/Final Environmental Statement, dated Safety Evaluation Report, or Supplement No.

. dated Environmental Assessment and Finding of No Significant Impact, dated Q Notice of Consideration of Issuance of Facility Operating License or Amendment to Facility Operating License, dated Qg Bi-Meekly Notice; Applications and Amendments to Operating Licenses Involving No Sigil tl d

C id i,d d~f p

lP]

Exemption, dated Construction Permit No.

CPPR-

, Amendment No.

dated Facility Operating License No.,

Amendment No.

dated Order Extending Construction Completion Date, dated Monthly Operating Report for transmitted by letter dated Q Annual/Semi-Annual Report-transmitted by letter dated

Enclosures:

As stated Division of Reactor Projects-I/II Office of Nuclear Reactor Regulation cc: See next page OFFICE)

SURNAME/

DATEP NRC FORM 318110i801 NRCM 0240

~ ~ ~ ~ ~ ~ ~ ~ ~ ~

~ ~ ~ ~ ~ ~ ~

4

~ ~ ~ ~ ~ ~ ~

4

~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ 0 ~ ~

~

~

~ ~ ~

OFFICIAL R EGO R D COPY

~ ~ ~ ~ ~ ~ ~ ~ ~

~ ~

~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~

~

~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~