ML20112H179

Provides Assessment of Impact of Recently Identified Single Failure Scenario on Previous Plant LOCA Analyses & Addl Assessment of Safety Significance of Discovery

ML20112H179
Person / Time
Site:	Vermont Yankee
Issue date:	06/10/1996
From:	Duffy J VERMONT YANKEE NUCLEAR POWER CORP.
To:	NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM)
References
BVY-96-76, NUDOCS 9606180033
Download: ML20112H179 (6)

Text

. _ ..

?

WERMONT YANKEE

-NUCLEAR POWER CORPORATION 1

- l

..- - Ferry Road, Brattleboro, VT 05301-7002 ENGINEERING OFFICE N 580 MAIN STREET BOLToN. MA 01740 (508) 779-6711 June 10,1996 BVY 96-76 United States Nuclear Regulatory Commission ATTN: Document Control Desk

, Washington,DC 20555

References:

(a) License No. DPR-28 (Docket No. 50-271)

(b) Letter, VYNPC to USNRC, BVY 96-61, dated 5/9/96

Subject:

Impact of New Single Failure Scenario on Previous Vermont Yankee LOCA Analyses The purpose of this letter is to provide an assessment of the impact of a recently identified single failure scenario on previous Vermont Yankee loss-of-coolant-accident (LOCA) '.nalyses and an additional assessment of the safety significance of this discovery.

In LER 96-10 [ Reference (b)], Vermont Yankee reported the discovery of the dependency between the residual heat removal (RHR) minimum flow valve and the cross-powered RHR pumps. This dependency impacted the limiting single failure in the LOCA analysis for Vermont l Yankee. Subsequent to issuing that LER, Mr. Ralph Landry of the NRC Staff visited our engineering office to review the technical work performed in concluding that no significant safety concern existed. This involved a review of both current and previous LOCA analyses supporting Vermont Yankee operation. During the course of the review, a telephone conference was conducted with General Electric (GE) personnel regarding the design basis LOCA analyses for operating cycles prior to Cycle 17 (i.e., prior to adoption of the new RELAP5YA-BWR based LOCA analysis).

As described in the attached letter, the original design of the Vermont Yankee plant bounds the new single failure scenario. The maximum peak clad temperature (PCT) cases for the design basis LOCA analysis (intermediate-to-large sized breaks) prior to Cycle 17 would be unaffected by the availability of the RHR pumps due to the timing of the event. For other PCT cases the previous models were overly conservative. Using more realistic licensing techniques, (e.g.

SAFER /GESTR), one core spray pump plus the automatic depressurization system (ADS) will, most likely, maintain adequate cooling. Further, best estimate calculations using previous v

8 i 9606100033 960610 PDR ADOCK 05000271 g p PDR +

s United States Nuclear Regulatory Commission VERMONT YANxto NUCLEAR POWER CORPORATION June 10,1996 Page 2 of 2 models showed acceptable results with one core spray pump and ADS for any break size and location.

We trust that the information provided satisfies your request; however, should you have any questions, please contact this office.

Sincerely, VERMOr;T YANKEE NUCLEAR POWER CORPORATION

&Htid -

James J. Duffy Licensing Engineer Attachment c: USNRC Region I Administrator USNRC Resident Inspector - VYNPS USNRC Project Manager - VYNPS l

J G GE Nuclear Energy 175 CurtnerAvenue, Mc 172 Sets Jose, CA 95111 June 9,1996 _

To: Michele Sironen Yankee Atomic Electric Company Subject. RHR Minimum Flow Bypass Valve Impact on ECCS Analysis

References:

1. Vermont Yankee LER 96-010,

2. Phone discussions ofMay 30,1996 and June 7,1996.

Dear Michele,

Attached please find GE's assessment regarding the impact on the Vermont Yankee ECCS performance analysis if the RHR minimum flow bypass valves fail to open and, as a result,

' cause damage to the RHR pumps.

Please contact me if you require additional support on this issue.

Best regards, l hYk- l David A. Hamon Phone (408) 925-4593 Fax (408) 925-1412 cc: E.G. Thacker

I Backgruend:

l It was recently discovered at[Vennont Yankee that the minimum flow bypass valves for RHR the (LOCA) Residual Heat Removal signal. The RHR are used to (dunps) pumpsCoolant provide Low Pressure would not Injection open on a I

(&CI) flow to the reactor prpssure vessel (RPV) following a LOCA. As a result of the failure of the mimmum flow bypass valve to open, if the LPCI flow to the vessel is not initiated within two minutes, there is a possibility that the RHR pumps may not function ,

ca operly. .

j Discussion: _

The original BWR/4 design was based on adequate emergency core cooling system (ECCS) performance with S plus one core spray for the entire break spectmm 'lliis was demonstrated using the els, assumptions and acceptance criteria which existed at l the time of the design in mid-1960s. Thus, the recently discovered situation at j Vermont Yankee which couldl lead to loss of all RHR pumps if they do not begin injecting l into the RPV in less than tso minutes is within the original design basis of Vermont J Yankee.

Subsequent to the original BWR/4 design, the ECCS acceptance criterion for peak daMag temperature (PCT) was inade more restrictive, and numerous penalties [i.e.,

counter-current flow limitaticn (CCFL), core spray distnbution, etc.] were imposed on GE's ECCS evaluation models. 10CFR50 Appendix K was also developed which produced changes in both nKxlels and analysis assumptions. These changes lead to calculated PCTs which were very close to the 10CFR50.46 acceptance criterion for Vermont Yankee.

The ECCS analyses documested in References 1 and 2 have a number of events with calculated PCTs that are v ' close to 2200"F. All of the cases with high PCTs are for intermediate-to-large sized (0 8 ft' to DBA) recirculation discharge breaks with failure of the LPCI injection valve in the unbroken recirculation line. None of these cases take any credit for LPCI flow reaching the vessel. Flow from the two L"CI pumps injecting into the broken recirculation loop is assumed to be lost out the break before reachin6 the RPV.

l Flow from the two LPCI pun,ips which could inject into the unbroken recirculation loop 1 never reaches the RPV due tp the assumed failure to open of the LPCI injection valve.

GE's conservative SAFE and REFIAOD evaluation models were used in the Reference 1 and 2 evaluations. Use of more realistic models such as SAFER /GESTR would produce much lower calculated PCTs.

In addition to the single failure considerations described above, the depressurization rate of the intermediate-to-large ! sired breaks was examined. For the 0.8 ft2 break, the calculated RPV pressure at do minutes into the event is already very close to atmospheric pressure, and larger breaks pould depre.ssurize even faster. Consequently, the LPCI i

purnps would be up and runnihg in less than two minutes for all events in Reference I and i 2 that reauked in high calculated PCTs. l l

Following the Three Mile Island (TMI) Accident, GE performed a large number of realistic ECCS performance evaluations to demonstrate the BWR response to a variety of small break accident events. l Specific examples of small break events for BWR/4 plants with only one low pressure EjCC system available are shown in Figure Groups 3.5.2.1-4 (LPCS + ADS) and 3.5.2.1-13 LPCI + ADS) of Reference 3. A similar case is shown in Figure Group 3.5.2.1-17 which(takes credit for one LPCS, one LPCI and ADl of these cases show calculated core average temperatures ofless than 1000*F, but results for the high power bundle were not apecifically evaluated in this study. A comparison of the case with one LPCS plus bne LPCI to the case withjust one LPCS indicates that the core average temperature results are not very sensitive to whether or not the IPCI pump actually operates. l l

In the mid 1980s, realistic ECCS evaluations were performed using SAIT for the BWR Owners' Group (BWROG) k devek>p success criteria for use in probabilistic safety -

analyses (PSAs) [4]. These , analyses demonstrate that one core spray system plus two AUS valves pruvide adequate core cooling (Idgh power bundle PCT < 2200*F) for the entire range of postulated break sizes and locations. Much lower small break temperatures would be obtained by taking credit for the full complement of ADS valves.

I ,

More recently, GE has obtalhed approval for its improved long-term thennal-hydraulic model, SAFER /GESTR. SAFER builds directly on a combination of GE's previously used i

ECCS performance evaluahon model Major improvenmus over the earlier SAFE /REFLOOD evaluation model have been incorporated in the following areas:

I f

. Nodal Representatlan ofthe Reactor Pressure Vessel (RPV):, All major regions of the vessel are represented separa:cly. This allows a more realistic void fraction profile and mass inventory to be calculated and provides better modeling of the counter-current flow limiting (CCFL) phenomerum.

l

. Hydesalic Calculations: The CCFL phenornenon is treated at all restrictions between the major core regions. Th most significant performance improvement occurs at the i bundle entrance where CCFL effects hold liquid in the bundles early in the transient,

&is delaying dryout time IyAlso, improvements in bypass leakage modeling allow more coount from the bypass region to enter the bundles contributing to earlier reflood.

i

. Core Heat Transfer Realistic heat transfer coefficients are used for the various flow regimes: nucleate boiling,it ransition boiling, dispersed droplet fdm boiling, pool fdm 3

boiling, steam cooling und single-phase liquid forced convection. Realistically

. calculated heat transfer,j coupled with more accurate mass balance and mass

distnbution, results in sub, stantial lower PCTs and more accurate prediction of the water level response. I i

I

Although no SAFER /GESTR analysis has been performed for Vermont Yankee, GE l

< anticipates that SAFER would most likely show acceptable Appendir K results for small breaks with only one core spray and ADS available.

Conclusion:

The original design of the Vemx>nt Yankee plant bounds the recently discovered situation which could lead to the loss o{all RHR pumps. In addition, the cases with high calculated PCTs which are reported in References 1 and 2 would be unaffected by the availability of )

the RilR pumps. These cased depressurize to below the shutoff head of the RHR pumps well before two minutes, and these casca also took no credit for RIIR/IECI flow reaching the RPV.

I By using more realistic licensmg analysis techniques, it can be shown that one core spray plus ADS will most likely nph'ntain adequate core cooling for unall breaks that don't depressurize to below the shdtotT head of the RIIR pumps in less than two minutes. In addition, by using best-estimhe analysis assumptions it has been shown that one core spray plus two ADS valves arp adequate for any break size and location.

I

References:

I. NEDO-21697, Losssf-Coolant Accident Analysis Report for Vermont Yankee NuclearPowerStatiof, August 1977. l l

2. NEDO-21662-2, Lospl-of-Coolant Accident Analysis Report for James A.

FitzPatrick Nuckar Power Plant (Lead Plant), Revisic n 2, July 1977.

l

3. NEDO-24708 A, AdditionalInformationfor NRC Staff Generic Report on Boiling ,

Water Reactors, Voluhes 1 and 2, Revision 1, December 1980.

l

4. NEDC-30936P, BWR Owners' Group Technical Specification Imprznnnent i Methodology (with nstration for BWR ECCS Actuation Instrumentation), \

I Part I, November 19 5.

l l l l