ML20247H328
| ML20247H328 | |
| Person / Time | |
|---|---|
| Site: | Brunswick  |
| Issue date: | 05/23/1989 |
| From: | Cutter A CAROLINA POWER & LIGHT CO. |
| To: | NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM) |
| References | |
| IEIN-85-071, IEIN-85-71, NLS-89-104, NUDOCS 8905310211 | |
| Download: ML20247H328 (11) | |
Text
-... _ - -. _ - - - _ -. _.._ __-_. -
g'#
c.,
c P.O. Box 1551
- Raleigh, N c. 27602 WAY 2 319B')
SERIAL: NLS-89-104 A. B CUTTER Vice President Nuclear Services Department United States Nuclear Regulatory Commission ATTENTION: Document Control Desk Washington, DC 20555 BRUNSWICK STEAM ELECTRIC PIANT, UNIT NO. 2 DOCKET No. 50-324/ LICENSE NO. NPF-62 EXEMPTION REQUEST 10CFR50, APPENDIX J TYPE A TESTING FREQUENCY Centlemen:
INTRODUCTION Carolina Power & Light Company hereby requests a one-time exemption from the schedule requirements of 10CFR50, Appendix J,Section III. A.6(b) for the Brunswick Steam Electric Plant, Unit No. 2.
The referenced section requires that:
If two consecutive periodic Type A tests fail to meet the applicable acceptance criteria in III.A.5(b),
notwithstanding the periodic retest schedule of III.D, a Type A test shall be perfcrmed at each plant shutdown for refueling or approximately every 18 months, whichever occurs first, until two consecutive Type A tests meet the acceptance criteria in III.A.5(b), after which time the retest schedule specified in III.D may be resumed.
Since 1956, the Company has conducted two Type A (containment integrated leakage rate - ILRT) tests at' Brunswick-2. Each of these tests were considered to be failures due, in part, to leakage penalty additions from Type C (containment isolation valves local leakage rate - LLRT) testing. The problems identified by the Type C testing were corrected.
Overall, these last two Type A tests have indicated a high degree of containment integrity. Therefore, CP&L requests a one-time exemption from the schedule requirements of paragraph III. A.6(b) so that the normal retest schedule can be resumed in accordance with Section III.D.
The requested exemption is based on the guidance provided in Information Notice 85-71, " Containment Integrated Leak Rate Tests," dated August 22, 1985. This notice advises licensees that the NRC Staff will consider exemptions to the accelerated Type A testing frequency incurred after the failure of two successive Type A tests when these failures were a 00l'I 8905310211 890523 i:
PDR ADOCK 05000324 P
PNU m
t i
Document Control Desk NLS-89-104 / Page 2 i
i result of leakage penalty additions made due to Type B and C testing.
Although the previous two Brunswick-2 ILRTs did not meet existing Technical Specification ILRT acceptance criteria, the Company believes the failures were due to extenuating circumstances which are not indicative of overall containment integrity. As such, the primary reason for failing these ILRTs is considered to be leakage penalty additions from Type C testing.
The following discussion provides details of the previous Type A tests, the reasons for the respective failures, and the actions taken to correct the situation.
DISCUSSION 1986 ILRT Synopsis The Brunswick-2 containment was subjected to an ILRT during the period of May 3 to May 5, 1986.
This ILRT failed due to a combination of factors. Using the minimum pathway leakage analysis, the "as found" reactor containment integrated leakage rate indicated that the acceptance criteria of 0.375% by weight per day would have been exceeded. This was due to two penetrations that could not be pressurized during their local leak rate tests and required maintenance to be performed.
In addition, an isolated valve problem was corrected during the performance of the ILRT.
The containment isolation valves for residual heat removal (RHR) B loop injection line are E11-F015B and E11-F017B (Figure 1).
The section of pipe between these valves could not be pressurized when the LLRT for the penetration was attempted.
Subsequent investigations found a worn disk on E11-F015B and a six inch section of the stellite material on the saat of E11-F017B missing. These problems were corrected and post-maintenance testing of the valves was completed satisfactorily.
The previous test history for this penetration does not indicate any recurring problems. As such, trase were considered isolated failures, not indicative of a generic problem.
The LLRT for valve RXS-PV 1222C identified a failure of this valve.
RXS-PV 1222C is the containment isolation sampling valve for reactor building component cooling water supplied to the drywell cooling units (Figure 2).
The existing pneumatically operated valve was replaced with a solenoid operated valve.
Post-maintenance testing and the March 1988 LLRT of the valve indicated no leakage through the replacement valve.
This type of pneumatically operated valve has had a history of problems 4
at Brunswick. As such, the Company has upgraded the similar pneumatically operated valves for Brunswick-1 and Brunswick-2.
During performance of the May 1986 ILRT, the identified leakage rate was 0.590% by weight per da:
This was in excess of the ILRT limit of 0.375% by weight per day.
A leak was discovered at the suppression pool to reactor building vacuum breaker valve CAC-X20A (Figure 3).
In an
l Document Control Desk NLS-89-104 / Page 3 1
1 attempt to isolate the leak, the section of pipe between CAC-X20A and CAC-V16 was pressurized via the LLRT test connection.
This action caused tha ILRT measured leakage to drop to 0.237% by weight per day, well within the ILRT limit.
Following the ILRT, CAC-X20A and CAC-V16 were tested using LLRT methods.
This test was performed so that pre and post maintenance leakage rate values could be established and added to the results of the ILRT.
Several tests were performed; however, each test resulted in a zero measured leakage rate.
This indicates that some dirt or debris was trapped in the sealing surface of the CAC-V16 valve during the ILRT.
CAC-X20A was disassembled and a small rust bloom was discovered beneath the rubber sealing surface. With only slight leakage of the CAC-V16 valve during the ILRT, no back pressure built up to seal the CAC-X20A valve. However, when the volume between the CAC-V16 and CAC-X20A valves was pressurized, the CAC-X20A valve was leak tight. The rust bloom on CAC-X20A was sanded, the disk was painted, and a new rubber seal was installed. No problems were encountered with this penetration during the March 1988 ILRT or subsequent LLRTs performed on this penetration.
Carolina Power & Light Company considers the problems with valves CAC-V16 and CAC-X20A which caused the failure of the May 1986 ILRT of Brunswick-2 to be an isolated incident. Once the problem was corrected, the ILRT measured leakage then dropped to 0.237% by weight per day, well within the ILRT limit. This demonstrates a high degree of containment integrity and does not justify the risk associated with the accelerated Type A testing which unduly stresses the primary containment.
1988 ILRT Synopsis An ILRT of the Brunswick-2 reactor containment building was performed during the period of March 26 to March 28, 1988.
This ILRT failed due to a combination of factors as discussed below.
Using the minimum pathway leakage analysis, the "as found" reactor containment integrated leakage rate indicated that the acceptance criteria of 0.375% by weight per day would have been exceeded.
This was due to a penetration that could not be pressurized during the LLRT.
Also, the region between the two o-ring seals of the drywell head flange could not be pressurized due to a faulty o-ring.
The penetration which failed its LLRT was on a feedwater injection line.
The subject valves were B21-F010B and B21-F032B (Figure 4). Valve B21-F032B experienced a valve stem packing leak.
The asbestos type failed packing was replaced by a different type, Chesterton style 5300, in accordance with recommendations outlined in General Electric Service Information Letter No. 399. The valve subsequently passed the LLRT and was returned to service.
Seat leakage was found through valve B21-F010B due to failure of the valve ethylene-propylene soft seat material manufactured by Anchor Darling. Additional review is ongoing to determine if the sof t seat material is adequate for this application. A
1 Docum:nt Control Desk NLS-89-104 / Page 4 replacement valve seat, made by Parker Seals Division, was installed and the valve was returned to service.
During performance of the "as found" LLRT between the two o-ring seals of the drywell head flange connection, pressure could not be maintained.
Leakage of the drywell head seal resulted from deformation / extrusion of the outer o-ring seal in three locations with apparent softening of the seal material at these locations. The inner o-ring seal was found to be intact.
These o-rings are composed of silicon rubber and were purchased from I. B. Moore, a subvendor of United Seal and Rubber Company.
Sections of the failed o-ring, both in the affected areas and intact areas, along with sections of the inner o-ring and a section of a new unused o-ring, were analyzed to determine the involved failure mechanism. An examination of the failed o-ring found that the observed softening of the seal material was unique to the three failure locations. These locations were softer than the unaffected areas of the o-ring by more than 20 hardness values and were distinguished by a clear tacky material, which was determined to be a constituent of both the failed and new unused o-rings.
Based on this determination, it is believed that failure of the subject o-ring in the three localized areas resulted from incomplete formulation / processing of the o-ring during the manufacturing process. Historical data throughout the industry has been good for these o-rings. As such, Carolina Power & Light Company considers this a one time failure which does not indicate a repetitive concern.
The initial March 1988 ILRT of the Brunswick-2 containment did not meet Technical Specification acceptance criteria due to containment temperature and pressure fluctuations. The sequence of events which lead to this failure are as follows. At 0900 on March 26, 1988, 15 minute frequency test data collection was initiated.
Initial indications showed a slowly rising leakage rate of approximately 0.33% by weight per day. However, Operations was experiencing problems in maintaining a steady residual heat removal temperature which caused fluctuations in the reactor vessel level.
This introduced some periodic perturbations in the observed containment mass weight points and the corresponding mass point leakage rate. Additionally, due to the recent completion of the reactor vessel hydrostatic test, the vessel shell temperature was fluctuating in the range of 125 F to 135 F.
Since this was substantially higher than the containment ambient air temperature, a
heat source existed inside containment. Additional influences on the test data were caused by an operational requirement for two loop RHR shutdown cooling when the reactor vessel level dropped below 200 inches and an increase in RHR flow from 5,000 gpm to 7,500 gpm. This caused an additional drop in reactor vessel level resulting in more perturbations of the containment leakage rate.
Leak detection and identification teams were dispatched but no major source of containment leakage was identified. Three minor packing leaks were identified on the RHR containment spray valve Ell-F021A, containment vacuum breaker valve CAC-V17, and the feedwater B loop injection valve B21-F032B.
Documsnt Control Desk NLS-89-104 / Page 5 l
At this time (1230, March 26), no repairs were made.
By 1355, the containment leakabe rate was 0.35% by weight per day and still increasing slowly. However, regression analysis of containment mass weights recorded between the perturbations caused by RHR temperature and reactor vessel level changes indicated a containment leakage rate of approximately 0.31% by weight per day.
At 0745 on March 27, a decision was made to terminate the ILRT. The containment leakage rate had stabilized at approximately 0.39 to 0.40%
by weight per day.
Based on the regression analysis described above, it was believed that the actual containment leakage rate was lower and was probably on the order of 0.31% by weight per day.
However, due to the changes in RHR temperature and reactor vessel level, this could not be positively confirmed.
By 1035 on March 27, reactor vessel level had been raised to 235 inches, single loop RHR shutdown cooling had been established, Operations had committed to maintaining better RHR temperature control and the packing leaks on valves E11-F021A and CAC-V17 had been repaired. Containment ambient air temperature changes had been continuously monitored and were still within the temperature stabilization criteria. Containment pressure was well above the required 49 psig criteria at approximately 50.3 psig.
The ILRT was officially restarted at 1200 on March 27.
The containment leakage rate exhibited a gradual increasing trend, reaching a maximum valve of 0.39% per day at 1930.
Leakage detection and identification was again initiated but no areas of significant leakage were observed.
From 1930 on March 27 to 1200 on March 28, the containment leakage rate showed a continual and gradual decreasing trend. The containment integrated leakage rate test was concluded at 1200 on March 28 with an acceptable measured mass point leakage rate value of 0.307% by weight per day. The leakage rate at the upper 95 percent confidence level was 0.312% by weight per day.
Although the March 1988 ILRT of Brunswick-2 was a failure, Carolina Power & Light Company considers the problem which caused t.he failure to be a direct result of the operational difficulties in maintaining a constant containment temperature and pressure. The worst case leak rate during this initial ILRT attempt was 0.40% by weight per day.
While minor packing leaks were repaired prior to the second ILRT, no significant leaks were identified. This demonstrates a high degree of containment integrity and does not justify the risk associated with the accelerated Type A testing which unduly stresses the primary containment.
10CFR50.12 ANALYSIS Carolina Power & Light Company has reviewed this request and determined that the exemption should be granted pursuant to 10CFR50.12(a)(2)(ii) and (v), in that application of the regulation in this particular ins :.nce is not necessary to achieve the underlying purpose of the rule
~ Documsnt Control Desk NLS-89-104 / Page 6 l
and the exemption provides only temporary relief from the applicable requirement with which CP&L has made a good faith effort to comply.
The purpose of Type A testing is to measure and ensure that the leakage through the primary reactor containment does not exceed the maximum allowable leakage.
It also provides assurance that periodic surveillance, maintenance and repairs are made to systems or components penetrating the containment. The corrective actions, described above, ensure the integrity of the primary containment.
The Company beli(ves that containment integrity is better served through Type B and C testing and the associated maintenance than through accelerated Type A testing which unduly stresses the primary containment.
Granting of this exemption will provide only temporary relief from the schedule requirements of 10CFR50, Appendix J,Section III.A.6(b).
This one-time exemption will enable Brunswick-2 to resume the retest schedule specified in Section III.D of 10CFR50, Appendix J and, therefore, prevent unnecessary pressurization of the containment to peak accident pressure. This exemption will not apply if the next test is deemed a failure.
Such a failure would constitute two consecutive failures and Section III.A.6(b) would again apply.
SUMMARY
Carolina Power & Light Company believes that the previous two ILRTs, while technically considered failures, have demonstrated a level of containment integrity which meets or exceeds the standards required of similar nuclear units. The existing allowable leakage rate (L )
a specified in the Brunswick-2 Technical Specifications is 0.5% by weight per day.
This value is extremely conservative when compared with similar BWRs such as: Hatch, 1.2% by weight per day; Millstone, 1.2% by weight per day; Pilgrim, 1.5% by weight per day; and Duane Arnold, 2% by weight per day.
This overly conservative L, limit has resulted in difficulties in meeting the ILRT limit of (0.5)(.75) - 0.375% by weight per day.
Carolina Power & Light Company is in the process of preparing a license amendment request which justifies raising the L limit t 1.0%
a by weight per day.
The initial 1986 and 1988 ILRT results (Mass Point at 95% upper confidence limits) were 0.590 and 0.400 by weight per day.
As such, the previous two ILRTs would have passed an ILRT limit of 0.75%
by weight per day prior to leakage penalty additions.
In conclusion, Carolina Power & Light Company believes containment integrity is best ensured through strong Type B and C testing programs.
These programs have proved effective in identifying problem areas at Brunswick-2. Accelerated integrated leak rate testing results in elevated outage costs and unnecessarily focuses management attention without resulting in a significant safety benefit.
i J
' Document Control Desk NLS-89-104 / Page 7 ADMINISTRATIVE INFORMATION Carolina Power & Light Company requests approval of this exemption by August 5, 1989 in order to provide sufficient time for planning of the upcoming Brunswick-2 refueling outage, currently scheduled to begin in September 1989.
Should this exemption be approved, the next Type A test at Brunswick-2 will be performed during the Reload 9 outage, currently scheduled to begin March 1991. A license amendment request, adding a footnote to Surveillance Requirement 4.6.1.2.b, will be submitted for the Brunswick-2 Technical Specifications in the near future.
This footnote will state that Brunswick-2 is exempt from the Type A testing required during the upcoming Reload 8 outage.
Please refer any questions regarding this submittal to Mr. William R.
Murray at (919) 836-8661.
Yours very tr y,
/,/
b,
,f fk' A. B. Cutter ABC/ MAT cc:
Mr. S. D. Ebneter Mr. W. H. Ruland Mr. E. G. Tourigny l
L_______-_-____---_-__-_
I. '
h k\\
MINI-STEAM DRYWELL TUNNEL O
/
\\
j
-\\
1
.F0508 F0158 F0178 REACTOR a
PRESSURE YESSEL k
I V
f 1
FIGURE 1 e
w
4 i
iAn
\\ T.\\
\\ \\
DRYWELL
[
h t><lPY12228
- A SAMPLING YALVES REACTOR U
PRESSURE.
f- -
--b<J PY 1222C -
]-
YESSEL I
I 1
N. /
3
{'Y53 A
g
' Y52 '
DRYWELL
/
COOLING
/'
UNITS -
-s
'x
/
/ 's V28 s-FIGURE 2 1
1.
1 4
e
_n- - - -. - - -. - - ~. - - - - --- - - - - - - - - - - -
.+
g, 1;
1.-
s
\\'
\\
N N
.\\
\\.
\\
YS.
- SUPPRESSION -
[
CHAMBER.
s3
.s s
X200
' Y 17 Y16 X204 N-\\
s\\
\\
/;
NORTH RHR 4
FIGURE 3 4
__._._______._________.m.___.____.________.._.__,__._m
1.*
~, 4
). j,.. -
.4 t(
6 j
RCIC/
I DRYWELL RWCU
[(
\\
IO'00 b.
F0528
~
REACTOR
~
PRESSURE YESSEL L
f I '
I PIT j
flGURE 4 i
l.
1
_ _ _ _ - _ _. _. _. _ _. _ _ _ _