ML20205P632
| ML20205P632 | |
| Person / Time | |
|---|---|
| Site: | Byron, Braidwood, 05000000 |
| Issue date: | 11/01/1988 |
| From: | Chrzanowski R COMMONWEALTH EDISON CO. |
| To: | Murley T Office of Nuclear Reactor Regulation |
| References | |
| 5277K:1-4, NUDOCS 8811080244 | |
| Download: ML20205P632 (7) | |
Text
r _
^
x Commonwealth Edison
) One First Nat onal Pla.ta. Chjcago. lilinois O T Addras R:pty to: Past Othes Box 767 sj Chicago, ilknois 60690 November 1, 1988 i
i t
Mr. Thomas E. Murley, Director i
Nuclear Reactor Regulation i
U.S. Nuclear Regulatory Commission
[
Washington, DC 20555 l
Subjects Byron Units 1 and 2 Braidwood Units 1 and 2 Steam Generator Tube Rupture Analysis NRC Docket No. 's 50-liiL151_ sad 59-45alA11
[
References:
(a) September 1, 1988, letter frma j
R. A. Chrsanowski to T. E. Murley.
j (b) February 28, 1983, lettar f raa J. Gallo to I. Smith, R. Cole, and A. D. Callihan.
I (c) February 1983, Byron Risk Study:
A i
Probabilistic Risk Evaluation Based on the
{
Zion Probabilistic Safety Study, Westinghvuse Electric Corporation.
[
(d) Schneider, A.W. Estimated Trequency of Loss of Offsite Power for Edison Nuclear Units, Commonwealth Edison Company, May 1988.
i l
Dear Mr. Murleys l
f-The results of the Byron and Braidwood Stations Steam Generator Tube Rupture ($GTR) Analysis were provided in reference (a).
In the analysis, the Steam Generator Power Operated Relief Valvss (PORVs) were relied upon to prevent the steam generator f rom potentially overfilling as a result of a SGTR event coincident with the Loss of Offsice Power (LOOP).
Subsequently, f
Commonwealth Edison performed a review of the aralysis to determine any I
necessary control measures for the Steam Generator PORVs. The following i
discussion encompasses the probability that an event will occur which requires j
the Steam Generator PORVs and the availability of the Steam Generator PORVs.
i I
A copy of Mr. M. Hitchler's direct testimony regarding a contention in Byron Station's licensing hearing was provided in reference (b).
This letter contains a quantitative assessment of the frequency of a SGTR both in l
normal operations and under accident conditions, such as a Main Steam Line i
Break (MSLB) and a Loss-of-Coolant Accident (LOCA), and an assessment of their impact on safety was presented. The model for assessing SGTR was presented as being conservative. The results of the assessment concluded that the g
frequency of postulated tube ruptures combined with, or as a consequence of.
j transient conditions and accident conditions (such as MSLBs and LOCAs) are L
extremely low over the 40 years.! plant operation.
7
($
882 ceo244 8813o3 r
p ADOCK 05000454
)
PNU L_
A diccussion wcs alco procentcd for cxcluding SGTR from furth3r consideration in the design basis.
!t was acknowledged in the testimony that, although a specific numerical definition for the cutoff of probability of occurrence of events that must be included in the design basis does not exist, there is some guidance based on other events that have been defined as design basis limits. These include the seismic design basis of 10-3 to 10-4 per year, the LOCA and MSLB potential occurrence of approximately 10-3 per year, and the NRC safety goal limit for severe core damage of 10-4 per year. The testimony further notes that the SGTR event is beyond the preceding design envelope discussed. This does no mean the SGTR will never happen, but rather the frequency of occur ~ence is sutticiently small relative to the operating life of the units to be co'sidered incredible.
Lastly, at was indicated that SGTR events are predicted to result in l
severe core damage at frequoncles of 10-7 per year. This value was also l
documented in reference (c).
This is an extremely low value and well below the NRC safety goal limit and, as such, has a negligible contribution to l
severe core damage.
nince these factors and arguments were entered into the
[
direct testimony in the ASLB hearing leading to the Byron operating license, 3
it is reasonable to maintain that they, by virtue of acceptance, constitute de facto licensing bases for Byron and Braldwood.
Since thu SGTR Analysis submitted in reference (a) addresses a specific scenario of margin to overfill, it was necessary to determine the ef fect of the Steam Generator PORV availability. An evaluation was performed to determine the probability of having two Steam Generator PORVs available (given various combinations of out-of-service times) when required to mitigate the consequences of a SGTR event coincident with a LOOP.
provides details of the evaluation performed.
Considering a scenarlo where one Steam Generator PORY la allowed to be out-of-service for 1 year and a second Steam Generator PORV la allowed to be out-of-service 6 days, the conditional probabilit of having the required Steam Generator PORVs available is approximately 2X10- per ytar.
If this is considered to occur simultaneously with both a SGTR and LOOP, the frequency of havin twovalvesout-of-servicewhenrequiredisontheorderof3X10gmorethan per year.
This is based on the frequency of a steam generator tube rupture being estimated as 1.4X10-2 per year based on a calculation performed by Westinjhouse. The frequency of a LOOP is estimated as 4.0X10-2 per year or 1.1X10-4 per day as documented in reference (d).
The initiation of a SGTR and the LOOP are independent events, i.e.,
one event does not increase the probability that the other event will occur. Therefore, the frequency of the joint events, SGTR and LOOP, were conservatively estimated to occur over a 24 hour2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> period instesd of instantaneously. This value is less than the 10-7 value per reactor year total frequency due to all SCTR scenarios which result in core damage as presented in reference (b) and (c).
Further conservatism is introduced when it is noted that not only does a SGTR event with LOOP and less than two Steam Generator PORVs available for RCS cooldown have to occur, but it must be coincident with yet another (as of yet underfined) improbable failure to result in core damage.
-~
Byrcn cnd Brcidwood Stations hevo sovoral oxioting requirem:nts that help to ensure these valves will be available if required. Currently the Steam Generator PORVs are contained in the Byron and Braldwood Technical Specifications under the containment isolation provisions of Speelfication 3.6.3.
This specification addresses several surveillance requirements to be performed on a specified frequency or after any maintenence actis'ty. These surveillances include testing in accordance with ASME Section XI, In-Service Testing. All testing is performed in accordance with approved station procedures and includes the following items:
stroke time testing on a quarterly basis, Indicator testing at least once per 24 months, steam generator pressurn control loop calibrations, and environmental qualification parirdic parts and lubricant replacements.
The Steam Generator PORVs r.a be operated remotely or locally.
g Remote operation can occur from two locations; either the Remoto Shutdown Fanel or the Main Cc.' trol Room. The FORVs can be opened automatically in response to a high pressure in the. team generators, or they can be open1d manually by the operator using a control switch on the Remote Shutdown Psn91 or in the Main Control Room.
The PORVs ;an also be opened manually by using a locally mounted hydraulle pump to positiou the valves.
For the purpose of demonstrating local operation, actual in-p?>nt performance tests were conducted to derive operator response time. These l
tests included the time required to identify the need to locally operate the valve, dispatch an operator, open the manual isolation valve (if closed) and actuate the valve. This netted a total response time of 9 1/2 minuter.
In the Steam Generator Tube Rupture Analysis, reactor cooldowns are i
not initiated until 19 minutes utter the Reactor Trip in the margin to j
overfill case and 29 minutes aftor the Reactor Trip in the offsite dose case.
i Time after trip la referenced due to this being the point where tht' nned for the Steam Generator ?ORVs would be recognised.
As it can be seen by the above response times, it can be demonstrated that even under conditions of the Steam Generator PORVs being manually isolated the response times assumed in the Steam Generator Tube Rupture Analysis can be met.
The Emergercy Procedures in place at Byron and Braidwood, currently address using the Steam Generator PORVs to provide cooldown on a Steam Generator Tube Rupture with Loss of offsite Power.
Byr:: and Braidwood Stations will be conducting training for licensed and Operating Department personnel on the Steam Generator Tube 1
i Rupture Analysis, to include the use of the Steam Generator PORVs and time f rames involved.
In conclusion, the contribution of a SGTR to core melt frequency is on the order of 10-7 per year. The evaluation from the attachment of the l
probability of occurrence of a SGTR with a LOOP and not having the required Steam Generator PORVs available is approximately 10-8 per year.
The probability of this occurring is w611 below the contribution of SGTR to core melt frequency.
Additional conservatism is also available from the fact that the 10-8 per year value does not include the probability of this ccenario j
actually resulting in core molt. The $ team Cenerator PORVs are currently
^
included in the Technical Specifications.
Perlooie surveillances, maintenance and testing of the valves are performed in accordance with the Technical i
Specifications, In-Service Inspection Prooram and Equipeent Qualification l
Program. This adds l
-4.-
n,
+
cssurc co th t tho volvoo will be cvoiltblo thsn reqairco.
In codition, tho training progrem and Eme gency Procedures will enhance operator awareness of I
ths valves availability status for RCS Cooldown in the unlikely event of a SGTh with LOOP.
i Commonwealth Edison believes that due tol i
r (1) the very small frequency of this event occurring, (2) the negligible impact on core melt probability, (3) the existing contro? s to ensure avellability of the Steam Generator PORVs, (4) the emergasey procedures that are in place, and (b) the training program for the operators that is underway, adequate administrative controls already exist, and no further actions are required.
If any questions arise on t'nis matter, please direct them to this office.
Very truly yours, hYt1JetW)
- e. A.j/Chrsanowski Nuclear Licensing Administrator
/sc1 att.
cet Byron Resident Inspector Braidwood Resident Inspector L.N. 01shan (NRR)
S. Sands (NRR)
Region III office 5277Kil-4
Att:chme:t !
Summ ry cad Evolu: tion of W:stingh:uro SG PORY Out of Service Report References Westinghouse Report entitled! "Steam Generator PORV Out of Service Time", dated October 14, 1988.
Summary 1 The Westinghouse report first reviewed Byron and Braidwood SG PORV operating performance data to determine the random failure rate for the SG PORV.
A SG PORV failure is defined as a complete inability (neither remoto, local, or manually) to open the valve.
A review of SG FORV failures over the period of Byron and breldwood plant operation (Initial startup data was not included) resulted in a calculated SG PORV total random failure rate QV = 7.92 E-3 per valve.
The Westinghouse report then derived the equation for calculating the RCS cooldown f ailure f requency (defined as a steam generator overfill occurring as a result of a delayed / inadequate RCS cooldown) due to an inadequate number of available intact SG PORVs.
The NFS SGTR analysis for Byron and Braie>ood previously determined 2 SG PORVs on intact SGs are required for an adequate RCS cooldown to prevent SG overfill. The RCS cooldown in11ure frequency equation was composed of three terms, representing the Individual frequencies assiclated with the three possible SG PORV OOS scenarios. These are given as follows:
"Q2/3" is the RCS cooldown failure frequency due to only random failures when no valves are OOS.
"Oo = 0.25 X Q2/3 + 0.75 X Q1/2", as the RCS cooldown failure freqvency given 1 vaiv: Is 005.
"Q1 =
- .5 X Q1/2 + 0.5",
is the RCS cooldown failure frequency given 2 valves are 00s.
Where:
"Q2/3 = 3(QV)**2 X (1-QV) + QV**3" = 1.87 E-4 la the frequency of 2 out of 3 SG PORVs rande,1y falling.
"Q1/2 = 2(QV) X (1-QV) + Ova *2"
= 1.58 E-2, is the frequency of 1 out of 2 SG PORVs randomly falling.
The equation for total RCS cooldown failure frequency (due to PORV unavailability) O over a year period (T 365 days), given 1 valvw OOS for 71 days and 2 valves DOS for T2 days, is then O = 17:0,25To20,25T1102/l_+_(0.7370:0.15T110142_+_Qi5T1 T
The Westinghouse report then determined the current frequency of a SGTR event at the Byron and Braidwood plants to be 1.4 E-2 per reactor year based upon the latest Industry data.
The calculated LOOP treguencies were
-2 cbtcIned from cn intercal Commonw3cith Edisc2 rcport, cnd theso volu s eOro conservatively bounded by a 4.0 E-2 per reactor year ftequency. The frcquency of a SGTR coincident within the same day as a 100P (both events are considered in~ dependent), was then calculated to be 1.54 E-6 per reactor year. The frequency of a SGTH, coincident LOOP, and RCS cooldown failure could then be calculated for any variation ot 2 PORV 005 time (T1 and T2), by determining the RCh cooldown frequency failure Q, and multiplying this times the 1.54 E-6 fiTR/ LOOP frequency.
This calculated RCS cooldown failure frequency resulting from inadequate SG PORV capacity, can then be compared to the established core melt frequency attributed to all SGTR events for the Byron and Braidwood plants of 1.0 E-7 per reactor year.
The conservative premise is to determine allowable SG PORV Dut of Service (OOS) times which generate RCS cooldown failure frequencies less than the 1.0 E-7 core melt frequency. Since the SGTR/ LOOP and RCS cooldown failure event of interest only causes SG over*111 to occur, an additional event whose probability by de(inition is less than one is also required. Therefore, the combined frequency of a SGTR/ LOOP /RCS cooldown f ailure e-rent Isading to any core melt damage is conservatively ensured to be less than the established 1.0 E-7 combined core melt frequency due to all SGTR events.
Evaluallout The referenced Westinghouse report derav44 the calculated espected frequency of a Steam Generator Tube Rupture (SGTR) event, with a coincident Loss of Of fsite (LOOP) and RCS cooldown f ailure leading to steam ponerator overfill, due to various SG PORV failure / unavailability scenarios. After reviewing and accepting the Westinghouse report derivation, it was decided to select a range of PORV unavailability frequency limits of 2.0 E-2, 4.0 E-2, and 6.0 E-2 per reactor year, such that the combined SGTA/ LOOP /hCS cooldown failure frequencies would bei 3.08 E-8, 6.16 E-8, and 9.24 E-8 respectively.
Since these values are all below the 1.0 E-7 core melt frequency, this establishes a very conservative analytical method for evaluating a wlde range of SG PORV 00S times.
The plot of the TO and 71 OOS combinations producing failure frequeneles Q equalling the conservative evaluation limits are shown on Table 1.
A review of numerous SG PORV OOS times To and 71, indicates the first valve DOS time To is almost insignificant, while the second valve DOS time is much more of a factor determining the failure frequency Q.
A very wide range of 00S combinations (f rom 52 weeks /6 days to 6 weeks /43 days) on the 1st/2nd PORVs is obtained for a small variation in the failure frequency limit Q.
This indicates that given the very conservative derivation of the SG PORY 005 time ascessment, and the low initiating frequency of a SGTR/ LOOP (1.54 E-6 reactor year), the SG PORY OOS times are not considered to be a significant contributor to the overall core melt frequency due to the SGTR event.
Figure 1 2nd SG PORY OOS Time vs ist SG PORY 00S Time As a Function of Failure Frequency Q 50_
IH Q = 6.0 E-2
?.
Ep 30-Q = 4.0 E-2 g
0 I
20-j o(
0 = 2.0 E-2 510-1 0-i i
i i
i i
i 0
5 10 15 20 25 30 35 40 45 50 55 ist SG PORY 005 Time (Weeks) f