ML20236U653

Safety Evaluation Re Effects of Errors in Util Analysis of Small Break Loca.Use of Incorrect Values in Analysis Results in Incomplete Compliance w/10CFR50,App K.Plant Poses No Risk to Public Health Due to Meeting 10CFR50.46 Requirements

ML20236U653
Person / Time
Site:	Davis Besse
Issue date:	11/27/1987
From:	NRC
To:
Shared Package
ML20236U649	List: ML20236U647 ML20236U653
References
TAC-62086, NUDOCS 8712030260
Download: ML20236U653 (9)

Text

/ 'o,, UNITED STATES l" o NUCLEAR REGULATORY COMMISSION g o j WASHINGTON, D. C. 20555

%,...../  !

1 I

REVIEW OF NON-CONSERVATIVE ASSUMPTIONS USED IN SMALL BREAK LOSS-0F-COOLANT-ACCIDENT ANALYSES l DAVIS-BESSE i

1.0 INTRODUCTION

q By a telephone call on July 10, 1986 and a subsequent follow-up letter j dated December 10, 1986 (Ref. 1), Toledo Edison Company reported to Region III that non-conservative assumptions were found in the small break loss-of-coolant-accident (SBLOCA) analyses performed by the Babcock and Wilcox Company (B&W) for Davis-Besse (Ref. 8). The non-conservative .

assumptions used in the SBLOCA analyses were identified as follows: (1)a reactor coolant system (RCS) low pressure trip setpoint of 2065 psia instead of 1900 psia was used, (2) a main feedwater (MFW) coastdown time of 43.5 seconds instead of 7 seconds was assumed, (3) steam generator water was controlled at 32 feet instead of 10 feet by the auxiliary feed-water (AFW) system,and(4)theAFWsystemprovidedwatertobothsteam l

generators (SG's)insteadofonlyonesteamgeneratorfortheworstcase l single failure assumption. Region III requested (Ref. 2) NRR's assistance in evaluating the significance of these non-conservative assumptions on the SBLOCA licensing analyses documented in B&W Topical Report BAW-10075A, Revisions 1 (Ref. 8).

The Reactor Systems Branch (RSB) of NRR has reviewed the licensee's assess-ment (Ref. 4) of the effect of these non-conservative assumptions on the l

SBLOCA licensing analyses for Davis-Besse and has prepared the following evaluation.

8712030260 871127 DR ADOCK0500g6

4

- 2 ;,

2.0 EVALUATION _

The licensee had B&W perform a qualitative evaluation to determine the impact of the non-conservative assumptions on their SBLOCA analytical results. The results of this evaluation are documented in Reference 4.

The licensee submitted, et the staff's request, Reference 4 for review.

The B&W evaluation in Reference 4 was primarily based on the following B&W analyses: 1)BAW-10075A, Revision 1(Ref.8),2)BAW-10154A(Ref.6),

3) calculational results of the high pressure injection line break, core

' ~

flood tank break and 0.5 ft break, and 4) the auxiliary feedwater system-analysis and other related analyses. The effects of the assumptions on-the SBLOCA analysis are evaluated and summarized below.

2.1 Low Pressure Reactor Trip Setpoint A low pressure reactor trip setpoint of 2065 psia instead of 1900 psia ,

(the Technical Specification value less uncertainty) was used for the SBLOCA licensing analyses. For the licensing analyses, it is assumed that the loss of offsite power occurred when the reactor tripped. Before the reactor trip, the reactor coolant pumps and main feedwater pumps are assumed operational. During this pre-trip period, the SG's maintain full heat removal capacity and thus maintain a balance between the reactor heat generation and SG heat removal. Furthermore, the results of recent analyses (Ref 5) performed by B&W indicated that the use of a 2065 psia, instead of 1900 psia, 1cw pressure reactor trip setpoint only delays the l

trip tiine and has an insignificant effect on the SBLOCA analysis results because of the maintainance of full SG heat transfer capacity prior to reacter trip.

Based upon the licensee's evaluation, the recent analytical results and I the staff's assessment of the licenstrg assumptions, the staff concludes thtt this assumption has minimal inptet on the SBLOCA analysis results for Davis-Besse.

2.2 Steam Generator Models The steam generator heat transfer models are affected by the following <

l assumptions: (1)thenumberofSG'savailable;(2)theSGlevel l control setpoint; and (3) MFW coastdown time. The current Davis-Besse SSLOCA analyscs assumed that both SG's received flow from the AFW system. However, the worst single failure event assumed for the -

SBLOCA licensing analyses is loss of one diesel generator. For Davis Besse, this failure results in the AFW system providing flow to only one SG.

The AFW system is actuated in the SBLOCA analyses based upon the loss of offsite power assumed at the time of reactor trip. After accounting for the startup of the AFW system, the analysis assures full AFW flow until the SB level reaches its level control setpoint. Within the Davis-Besse SLLOCA analyses, the SG level control setpoint was assumed

4_

to be 32 feet. However, the actual SG 1evel contori setpoint is only 10 feet at Davis-Besse.

Following the reactor trip and loss of offsite power, the MFW pumps trip and coast down. Water is added to the SG during the MFW coast-down period. This coastdown period was assumed to be 43.5 seconds in the Davis-Besse SBLOCA analyses. The licensee repurted that a  !

i' l 7 second coastdown period is more appropriate for the plant.

l To evaluate the impact of these non-conservative assumptions related to SG modeling for the Duis-Besse SBLOCA analyses, B&W examined -

previous SBLOCA results. These previous Bt.W analyses show that the effect of the SG modeling on the SBLOCA results are functions of break size. For a SBLOCA with a break size larger than 0.02 f t , the blowdown from RCS through the break was sufficient for heat removal and ultimately resulted in a negative tetuperature gradient between the primary and secondary sides of the steam generator. This  :

l negative temperature gradient resulted in a reverse SG beat transfer  !

(i.e.,theSGservesasaheatsource). These analyses further showed that the reverse SG heat transfer had a negligible effect on the consequences of the SBLOCA. Thus, it was concluded that the non-conservative assumptions for the SG modeling would not have a significant impact on brea'k sizes greater than 0.02 fte.  ;

For smaller size SBLOCA's, B&W stated that SG heat transfer affects the transient and is needed to assure the ECCS performance. However, B&W indicated that their plants were designed to assure that, before any ccre uncovering occurs, the RCS water level decrease would result f

in steam in the primary side of the SG tube region with adequate heat removal occurring via condensation. Condensation will occur if AFW injection occurs through the AFW nozzles or if the SG secondary side level exceeds the water level in the primary side of the SG tube. As a result of this condensation, the RCS pressure would decrease to a value near that of the secondary side (1000 psia) and would result in l

l sufficient ECCS injection to assure adequate long term cooling for these smaller sized SBLOCAs. Minimum AFW flow and SG pool levels have been determined by B&W which assure an adequate condensing surface.

For the Davis-Besse plant, these calculations show that a 3 foot level in both SG's (equivalent to 6 foot level in one SG) will result in sufficient heat transfer area for those SBLOCA's (such as a 0.01 ft

, break) where SG heat removal is necessary. Therefore, B&W concluded l

that the 10 foot SG 1evel control setpoint for the AFW system will assure an adequate condensing surface and thereby adequate decay heat removal. .

B&W also indicated that the existing analyses showed that sufficient SG heat transfer is available to assure an acceptable ECCS performance for the AFW system controlled at 10 foot level, with or without MFW.

Therefore, B&W concluded that the use of MFW coast down time of 43.5 seconds instead of 7 seconds does not significantly affect the previous SBLOCA analysis results. i l

1 l

1

Based upon the licensee's evaluation and the staff's assessment of the licensing assumptions, the staff concludes that the non-conservative assumptions for the SG modeling have minimal impact on the results of the Davis-Besse SBLOCA analyses results.

2.3 SBLOCA Analysis - 0.04 ft. Break In accordance with the requirements of TMI Action Item II.K.3.31 of NUREG-0737, the licensee recently submitted new SBLOCA analytical results(Ref.5)forreview. A SBLOCA break size of 0.04 ft.' was i

e analyzed using a modified version of the 1985 B&W ECCS model (Refs. 6 and 7), which satisfied the requirements of TMI Action Item II.K.3.30 and was approved by NRC. The assumptions for the SG availability.

AFW actuation setpoint, MFW coastdown time and low RCS pressure trip l

setpoint used for this analysis are consistent with the Davis-Besse plant configuration and Technical Specification values. Thus, this , ;

analysis foms a basis for direct comparison between the results obtained using the appropriate plant data and new computer model versus the previous plant SBLOCA analyses with the non-conservative assumptions and present models. The results of this recent analysis showed that while there were some differences in system response, the performance requirements of 10 CFR 50.46, e.g., peak cladding temperature shall be less than 2200 F, were satisfied. Therefore, the licensee concluded that the non-conservative assumptions had only a small effect on the previous SELOCA analytical results for Davis-Besse.

l

l i

The analyses of Reference 5 are currently being evaluated by the staff and the review is expected to be completed by December 31, 1987.

Based on the staff's limited examination of these analyses, we find j that the results further support the licensee's assessment that the noncunservative assumptions discussed in this evaluation have only a'small effect on the SBLOCA licensing analyses.

l l

3.0 _CCNCLOSIONS j

The staff has evaluated the effect of the nonconservative assumptions ]

used in the Davis-Besse SBLOCA analyses. Based on its review of the assessment report (Ref. 4) provided by the licensee, the staff concludes that some of the plant data used in the existing SBLOCA licensing analysis are incorrect, and thus the analyses are not in complete compliance with  !

the requirements of 10 CFR 50, Appendix K. However, the staff has also concluded that there is reasonable assurance that the Davis-Besse ECCS ,

satisfies the performance requirements of 10 CFR 50.46. Therefore, continued operation of the plant poses no undue risk to the public health and safety.

Revised SBLOCA analyses (Ref. 5) have been submitted by the licensee in order to satisfy the requirements of TMI Action Item II.K.3.31. The' staff is scheduled to complete its review by December.31, 1987. Assuming that the staff finds the licensee's submittal to be acceptable, the staff 1

will then be able to conclude that Devis-Besse is in full compliance with 10 CFR 50 Appendix K and 10 CFR 50.46.

)

.8 1 i

i

4.0 REFERENCES

1. Letter with an attachn'ent from 1.. Storz (Toledo Edison) tc NRC, dated December 10, 1986.

2. Letter from C. Norelius (Region III/NRC) to G. Holaban (NRC), Request -

l for Technical Assistance - Potential Non-conservathe Assumptions .

1 Used in Davis-Besr,e Small Break LOCA Analysis, dated July 18, 1986. j h

1

3. Letter from J. Williarc, Jr. (Toledo Edison) to J.. Stolz (NRC), dated November 6, 1986.

)

1 .

l

4. Lett,er with Attacnment from J. Williams, Jr. (Toledo Edison) to

]

J. Stolz (NRC), dated November 25, 1986.

9

5. Attachment to Reference 3, BAW-1981,-Small Break Loss-of-Coolant Accident Analysis for the B&W 177-FA Raised-Lorp Plant in Response to HUREG-0737, Item II.K.3.31, dated October 1986. )

6. BAW-10154A, N. Savani, et al., B&W's Small Break LOCA ECCS Evaluation Model, dated July 1985.

i l

.1 j

.i .;

1

7. BAW-10092A, Rev. 3, J. Cudlin, et al., CRAFT 2 Fortram Program for j Digital Simulation of a Multinode Reactor During Loss of Coolant, dated July 1985.

q

8. BAW-10075A, Rev. 1 Multinode Analysis of Small Breaks for BlW's l 177-Fuel-Assembly Nuclear Plants with Raised Loop Arrangement and Intervals Vent Valves, dated March 1976.

1

'l 1

i l

1 l

l^ ..

1 l

i

- - - _ _ - - - - _ _ _ - - . - _